Version History
-1.0.95.0 - November XX, 2014 (release scheduled)
--
-
- Updated to SQLite 3.8.7. -
- Make sure SQL statements generated for DbUpdateCommandTree objects are properly delimited. -
- In the SQLiteDataReader.VerifyType method, remove duplicate "if" statement for the DbType.SByte value and move the remaining "if" to the Int64 affinity. Fix for [c5cc2fb334]. ** Potentially Incompatible Change ** -
1.0.94.0 - September 9, 2014
- Updated to SQLite 3.8.6.
- Updated to Entity Framework 6.1.1.
- Refactor and simplify NuGet packages in order to support per-solution SQLite interop assembly files. ** Potentially Incompatible Change ** Index: Doc/Extra/Provider/welcome.html ================================================================== --- Doc/Extra/Provider/welcome.html +++ Doc/Extra/Provider/welcome.html @@ -158,11 +158,11 @@
- vfs: ^The "vfs" parameter may be used to specify the name of ** a VFS object that provides the operating system interface that should ** be used to access the database file on disk. ^If this option is set to @@ -2923,13 +2891,15 @@ ** equivalent to setting the SQLITE_OPEN_PRIVATECACHE bit. ** ^If sqlite3_open_v2() is used and the "cache" parameter is present in ** a URI filename, its value overrides any behavior requested by setting ** SQLITE_OPEN_PRIVATECACHE or SQLITE_OPEN_SHAREDCACHE flag. ** -**
- psow: ^The psow parameter indicates whether or not the +**
- psow: ^The psow parameter may be "true" (or "on" or "yes" or +** "1") or "false" (or "off" or "no" or "0") to indicate that the ** [powersafe overwrite] property does or does not apply to the -** storage media on which the database file resides. +** storage media on which the database file resides. ^The psow query +** parameter only works for the built-in unix and Windows VFSes. ** **
- nolock: ^The nolock parameter is a boolean query parameter
** which if set disables file locking in rollback journal modes. This
** is useful for accessing a database on a filesystem that does not
** support locking. Caution: Database corruption might result if two
@@ -3221,14 +3191,10 @@
** ^(
- SQLITE_LIMIT_VARIABLE_NUMBER
**- The maximum index number of any [parameter] in an SQL statement.)^ ** ** [[SQLITE_LIMIT_TRIGGER_DEPTH]] ^(
- SQLITE_LIMIT_TRIGGER_DEPTH
**- The maximum depth of recursion for triggers.
)^ -** -** [[SQLITE_LIMIT_WORKER_THREADS]] ^(- SQLITE_LIMIT_WORKER_THREADS
-**- The maximum number of auxiliary worker threads that a single -** [prepared statement] may start.
)^ ** */ #define SQLITE_LIMIT_LENGTH 0 #define SQLITE_LIMIT_SQL_LENGTH 1 #define SQLITE_LIMIT_COLUMN 2 @@ -3238,11 +3204,10 @@ #define SQLITE_LIMIT_FUNCTION_ARG 6 #define SQLITE_LIMIT_ATTACHED 7 #define SQLITE_LIMIT_LIKE_PATTERN_LENGTH 8 #define SQLITE_LIMIT_VARIABLE_NUMBER 9 #define SQLITE_LIMIT_TRIGGER_DEPTH 10 -#define SQLITE_LIMIT_WORKER_THREADS 11 /* ** CAPI3REF: Compiling An SQL Statement ** KEYWORDS: {SQL statement compiler} ** @@ -3512,37 +3477,29 @@ ** is negative, then the length of the string is ** the number of bytes up to the first zero terminator. ** If the fourth parameter to sqlite3_bind_blob() is negative, then ** the behavior is undefined. ** If a non-negative fourth parameter is provided to sqlite3_bind_text() -** or sqlite3_bind_text16() or sqlite3_bind_text64() then -** that parameter must be the byte offset +** or sqlite3_bind_text16() then that parameter must be the byte offset ** where the NUL terminator would occur assuming the string were NUL ** terminated. If any NUL characters occur at byte offsets less than ** the value of the fourth parameter then the resulting string value will ** contain embedded NULs. The result of expressions involving strings ** with embedded NULs is undefined. ** -** ^The fifth argument to the BLOB and string binding interfaces -** is a destructor used to dispose of the BLOB or +** ^The fifth argument to sqlite3_bind_blob(), sqlite3_bind_text(), and +** sqlite3_bind_text16() is a destructor used to dispose of the BLOB or ** string after SQLite has finished with it. ^The destructor is called -** to dispose of the BLOB or string even if the call to bind API fails. +** to dispose of the BLOB or string even if the call to sqlite3_bind_blob(), +** sqlite3_bind_text(), or sqlite3_bind_text16() fails. ** ^If the fifth argument is ** the special value [SQLITE_STATIC], then SQLite assumes that the ** information is in static, unmanaged space and does not need to be freed. ** ^If the fifth argument has the value [SQLITE_TRANSIENT], then ** SQLite makes its own private copy of the data immediately, before ** the sqlite3_bind_*() routine returns. ** -** ^The sixth argument to sqlite3_bind_text64() must be one of -** [SQLITE_UTF8], [SQLITE_UTF16], [SQLITE_UTF16BE], or [SQLITE_UTF16LE] -** to specify the encoding of the text in the third parameter. If -** the sixth argument to sqlite3_bind_text64() is not one of the -** allowed values shown above, or if the text encoding is different -** from the encoding specified by the sixth parameter, then the behavior -** is undefined. -** ** ^The sqlite3_bind_zeroblob() routine binds a BLOB of length N that ** is filled with zeroes. ^A zeroblob uses a fixed amount of memory ** (just an integer to hold its size) while it is being processed. ** Zeroblobs are intended to serve as placeholders for BLOBs whose ** content is later written using @@ -3559,30 +3516,23 @@ ** ^Bindings are not cleared by the [sqlite3_reset()] routine. ** ^Unbound parameters are interpreted as NULL. ** ** ^The sqlite3_bind_* routines return [SQLITE_OK] on success or an ** [error code] if anything goes wrong. -** ^[SQLITE_TOOBIG] might be returned if the size of a string or BLOB -** exceeds limits imposed by [sqlite3_limit]([SQLITE_LIMIT_LENGTH]) or -** [SQLITE_MAX_LENGTH]. ** ^[SQLITE_RANGE] is returned if the parameter ** index is out of range. ^[SQLITE_NOMEM] is returned if malloc() fails. ** ** See also: [sqlite3_bind_parameter_count()], ** [sqlite3_bind_parameter_name()], and [sqlite3_bind_parameter_index()]. */ SQLITE_API int sqlite3_bind_blob(sqlite3_stmt*, int, const void*, int n, void(*)(void*)); -SQLITE_API int sqlite3_bind_blob64(sqlite3_stmt*, int, const void*, sqlite3_uint64, - void(*)(void*)); SQLITE_API int sqlite3_bind_double(sqlite3_stmt*, int, double); SQLITE_API int sqlite3_bind_int(sqlite3_stmt*, int, int); SQLITE_API int sqlite3_bind_int64(sqlite3_stmt*, int, sqlite3_int64); SQLITE_API int sqlite3_bind_null(sqlite3_stmt*, int); -SQLITE_API int sqlite3_bind_text(sqlite3_stmt*,int,const char*,int,void(*)(void*)); +SQLITE_API int sqlite3_bind_text(sqlite3_stmt*, int, const char*, int n, void(*)(void*)); SQLITE_API int sqlite3_bind_text16(sqlite3_stmt*, int, const void*, int, void(*)(void*)); -SQLITE_API int sqlite3_bind_text64(sqlite3_stmt*, int, const char*, sqlite3_uint64, - void(*)(void*), unsigned char encoding); SQLITE_API int sqlite3_bind_value(sqlite3_stmt*, int, const sqlite3_value*); SQLITE_API int sqlite3_bind_zeroblob(sqlite3_stmt*, int, int n); /* ** CAPI3REF: Number Of SQL Parameters @@ -4327,11 +4277,11 @@ ** These routines work only with [protected sqlite3_value] objects. ** Any attempt to use these routines on an [unprotected sqlite3_value] ** object results in undefined behavior. ** ** ^These routines work just like the corresponding [column access functions] -** except that these routines take a single [protected sqlite3_value] object +** except that these routines take a single [protected sqlite3_value] object ** pointer instead of a [sqlite3_stmt*] pointer and an integer column number. ** ** ^The sqlite3_value_text16() interface extracts a UTF-16 string ** in the native byte-order of the host machine. ^The ** sqlite3_value_text16be() and sqlite3_value_text16le() interfaces @@ -4574,14 +4524,10 @@ ** ^The sqlite3_result_text(), sqlite3_result_text16(), ** sqlite3_result_text16le(), and sqlite3_result_text16be() interfaces ** set the return value of the application-defined function to be ** a text string which is represented as UTF-8, UTF-16 native byte order, ** UTF-16 little endian, or UTF-16 big endian, respectively. -** ^The sqlite3_result_text64() interface sets the return value of an -** application-defined function to be a text string in an encoding -** specified by the fifth (and last) parameter, which must be one -** of [SQLITE_UTF8], [SQLITE_UTF16], [SQLITE_UTF16BE], or [SQLITE_UTF16LE]. ** ^SQLite takes the text result from the application from ** the 2nd parameter of the sqlite3_result_text* interfaces. ** ^If the 3rd parameter to the sqlite3_result_text* interfaces ** is negative, then SQLite takes result text from the 2nd parameter ** through the first zero character. @@ -4621,11 +4567,10 @@ ** If these routines are called from within the different thread ** than the one containing the application-defined function that received ** the [sqlite3_context] pointer, the results are undefined. */ SQLITE_API void sqlite3_result_blob(sqlite3_context*, const void*, int, void(*)(void*)); -SQLITE_API void sqlite3_result_blob64(sqlite3_context*,const void*,sqlite3_uint64,void(*)(void*)); SQLITE_API void sqlite3_result_double(sqlite3_context*, double); SQLITE_API void sqlite3_result_error(sqlite3_context*, const char*, int); SQLITE_API void sqlite3_result_error16(sqlite3_context*, const void*, int); SQLITE_API void sqlite3_result_error_toobig(sqlite3_context*); SQLITE_API void sqlite3_result_error_nomem(sqlite3_context*); @@ -4632,12 +4577,10 @@ SQLITE_API void sqlite3_result_error_code(sqlite3_context*, int); SQLITE_API void sqlite3_result_int(sqlite3_context*, int); SQLITE_API void sqlite3_result_int64(sqlite3_context*, sqlite3_int64); SQLITE_API void sqlite3_result_null(sqlite3_context*); SQLITE_API void sqlite3_result_text(sqlite3_context*, const char*, int, void(*)(void*)); -SQLITE_API void sqlite3_result_text64(sqlite3_context*, const char*,sqlite3_uint64, - void(*)(void*), unsigned char encoding); SQLITE_API void sqlite3_result_text16(sqlite3_context*, const void*, int, void(*)(void*)); SQLITE_API void sqlite3_result_text16le(sqlite3_context*, const void*, int,void(*)(void*)); SQLITE_API void sqlite3_result_text16be(sqlite3_context*, const void*, int,void(*)(void*)); SQLITE_API void sqlite3_result_value(sqlite3_context*, sqlite3_value*); SQLITE_API void sqlite3_result_zeroblob(sqlite3_context*, int n); @@ -6330,17 +6273,16 @@ #define SQLITE_TESTCTRL_RESERVE 14 #define SQLITE_TESTCTRL_OPTIMIZATIONS 15 #define SQLITE_TESTCTRL_ISKEYWORD 16 #define SQLITE_TESTCTRL_SCRATCHMALLOC 17 #define SQLITE_TESTCTRL_LOCALTIME_FAULT 18 -#define SQLITE_TESTCTRL_EXPLAIN_STMT 19 /* NOT USED */ +#define SQLITE_TESTCTRL_EXPLAIN_STMT 19 #define SQLITE_TESTCTRL_NEVER_CORRUPT 20 #define SQLITE_TESTCTRL_VDBE_COVERAGE 21 #define SQLITE_TESTCTRL_BYTEORDER 22 #define SQLITE_TESTCTRL_ISINIT 23 -#define SQLITE_TESTCTRL_SORTER_MMAP 24 -#define SQLITE_TESTCTRL_LAST 24 +#define SQLITE_TESTCTRL_LAST 23 /* ** CAPI3REF: SQLite Runtime Status ** ** ^This interface is used to retrieve runtime status information @@ -6527,25 +6469,25 @@ ** memory already being in use. ** Only the high-water value is meaningful; ** the current value is always zero.)^ ** ** [[SQLITE_DBSTATUS_CACHE_USED]] ^(- SQLITE_DBSTATUS_CACHE_USED
-**- This parameter returns the approximate number of bytes of heap +**
- This parameter returns the approximate number of of bytes of heap ** memory used by all pager caches associated with the database connection.)^ ** ^The highwater mark associated with SQLITE_DBSTATUS_CACHE_USED is always 0. ** ** [[SQLITE_DBSTATUS_SCHEMA_USED]] ^(
- SQLITE_DBSTATUS_SCHEMA_USED
-**- This parameter returns the approximate number of bytes of heap +**
- This parameter returns the approximate number of of bytes of heap ** memory used to store the schema for all databases associated ** with the connection - main, temp, and any [ATTACH]-ed databases.)^ ** ^The full amount of memory used by the schemas is reported, even if the ** schema memory is shared with other database connections due to ** [shared cache mode] being enabled. ** ^The highwater mark associated with SQLITE_DBSTATUS_SCHEMA_USED is always 0. ** ** [[SQLITE_DBSTATUS_STMT_USED]] ^(
- SQLITE_DBSTATUS_STMT_USED
-**- This parameter returns the approximate number of bytes of heap +**
- This parameter returns the approximate number of of bytes of heap ** and lookaside memory used by all prepared statements associated with ** the database connection.)^ ** ^The highwater mark associated with SQLITE_DBSTATUS_STMT_USED is always 0. **
** @@ -7898,10 +7840,19 @@ #pragma warn -aus /* Assigned value is never used */ #pragma warn -csu /* Comparing signed and unsigned */ #pragma warn -spa /* Suspicious pointer arithmetic */ #endif +/* Needed for various definitions... */ +#ifndef _GNU_SOURCE +# define _GNU_SOURCE +#endif + +#if defined(__OpenBSD__) && !defined(_BSD_SOURCE) +# define _BSD_SOURCE +#endif + /* ** Include standard header files as necessary */ #ifdef HAVE_STDINT_H #include@@ -7938,22 +7889,10 @@ #else /* Generates a warning - but it always works */ # define SQLITE_INT_TO_PTR(X) ((void*)(X)) # define SQLITE_PTR_TO_INT(X) ((int)(X)) #endif -/* -** A macro to hint to the compiler that a function should not be -** inlined. -*/ -#if defined(__GNUC__) -# define SQLITE_NOINLINE __attribute__((noinline)) -#elif defined(_MSC_VER) -# define SQLITE_NOINLINE __declspec(noinline) -#else -# define SQLITE_NOINLINE -#endif - /* ** The SQLITE_THREADSAFE macro must be defined as 0, 1, or 2. ** 0 means mutexes are permanently disable and the library is never ** threadsafe. 1 means the library is serialized which is the highest ** level of threadsafety. 2 means the library is multithreaded - multiple @@ -8136,11 +8075,11 @@ # define ALWAYS(X) (X) # define NEVER(X) (X) #endif /* -** Return true (non-zero) if the input is an integer that is too large +** Return true (non-zero) if the input is a integer that is too large ** to fit in 32-bits. This macro is used inside of various testcase() ** macros to verify that we have tested SQLite for large-file support. */ #define IS_BIG_INT(X) (((X)&~(i64)0xffffffff)!=0) @@ -8215,19 +8154,19 @@ ** be opaque because it is used by macros. */ struct HashElem { HashElem *next, *prev; /* Next and previous elements in the table */ void *data; /* Data associated with this element */ - const char *pKey; /* Key associated with this element */ + const char *pKey; int nKey; /* Key associated with this element */ }; /* ** Access routines. To delete, insert a NULL pointer. */ SQLITE_PRIVATE void sqlite3HashInit(Hash*); -SQLITE_PRIVATE void *sqlite3HashInsert(Hash*, const char *pKey, void *pData); -SQLITE_PRIVATE void *sqlite3HashFind(const Hash*, const char *pKey); +SQLITE_PRIVATE void *sqlite3HashInsert(Hash*, const char *pKey, int nKey, void *pData); +SQLITE_PRIVATE void *sqlite3HashFind(const Hash*, const char *pKey, int nKey); SQLITE_PRIVATE void sqlite3HashClear(Hash*); /* ** Macros for looping over all elements of a hash table. The idiom is ** like this: @@ -8481,31 +8420,10 @@ */ #ifndef SQLITE_TEMP_STORE # define SQLITE_TEMP_STORE 1 # define SQLITE_TEMP_STORE_xc 1 /* Exclude from ctime.c */ #endif - -/* -** If no value has been provided for SQLITE_MAX_WORKER_THREADS, or if -** SQLITE_TEMP_STORE is set to 3 (never use temporary files), set it -** to zero. -*/ -#if SQLITE_TEMP_STORE==3 || SQLITE_THREADSAFE==0 -# undef SQLITE_MAX_WORKER_THREADS -# define SQLITE_MAX_WORKER_THREADS 0 -#endif -#ifndef SQLITE_MAX_WORKER_THREADS -# define SQLITE_MAX_WORKER_THREADS 8 -#endif -#ifndef SQLITE_DEFAULT_WORKER_THREADS -# define SQLITE_DEFAULT_WORKER_THREADS 0 -#endif -#if SQLITE_DEFAULT_WORKER_THREADS>SQLITE_MAX_WORKER_THREADS -# undef SQLITE_MAX_WORKER_THREADS -# define SQLITE_MAX_WORKER_THREADS SQLITE_DEFAULT_WORKER_THREADS -#endif - /* ** GCC does not define the offsetof() macro so we'll have to do it ** ourselves. */ @@ -8517,15 +8435,10 @@ ** Macros to compute minimum and maximum of two numbers. */ #define MIN(A,B) ((A)<(B)?(A):(B)) #define MAX(A,B) ((A)>(B)?(A):(B)) -/* -** Swap two objects of type TYPE. -*/ -#define SWAP(TYPE,A,B) {TYPE t=A; A=B; B=t;} - /* ** Check to see if this machine uses EBCDIC. (Yes, believe it or ** not, there are still machines out there that use EBCDIC.) */ #if 'A' == '\301' @@ -8692,11 +8605,11 @@ ** Assert that the pointer X is aligned to an 8-byte boundary. This ** macro is used only within assert() to verify that the code gets ** all alignment restrictions correct. ** ** Except, if SQLITE_4_BYTE_ALIGNED_MALLOC is defined, then the -** underlying malloc() implementation might return us 4-byte aligned +** underlying malloc() implemention might return us 4-byte aligned ** pointers. In that case, only verify 4-byte alignment. */ #ifdef SQLITE_4_BYTE_ALIGNED_MALLOC # define EIGHT_BYTE_ALIGNMENT(X) ((((char*)(X) - (char*)0)&3)==0) #else @@ -8759,20 +8672,10 @@ # define SQLITE_ENABLE_STAT3_OR_STAT4 1 #elif SQLITE_ENABLE_STAT3_OR_STAT4 # undef SQLITE_ENABLE_STAT3_OR_STAT4 #endif -/* -** SELECTTRACE_ENABLED will be either 1 or 0 depending on whether or not -** the Select query generator tracing logic is turned on. -*/ -#if defined(SQLITE_DEBUG) || defined(SQLITE_ENABLE_SELECTTRACE) -# define SELECTTRACE_ENABLED 1 -#else -# define SELECTTRACE_ENABLED 0 -#endif - /* ** An instance of the following structure is used to store the busy-handler ** callback for a given sqlite handle. ** ** The sqlite.busyHandler member of the sqlite struct contains the busy @@ -8901,18 +8804,16 @@ typedef struct Parse Parse; typedef struct PrintfArguments PrintfArguments; typedef struct RowSet RowSet; typedef struct Savepoint Savepoint; typedef struct Select Select; -typedef struct SQLiteThread SQLiteThread; typedef struct SelectDest SelectDest; typedef struct SrcList SrcList; typedef struct StrAccum StrAccum; typedef struct Table Table; typedef struct TableLock TableLock; typedef struct Token Token; -typedef struct TreeView TreeView; typedef struct Trigger Trigger; typedef struct TriggerPrg TriggerPrg; typedef struct TriggerStep TriggerStep; typedef struct UnpackedRecord UnpackedRecord; typedef struct VTable VTable; @@ -9097,12 +8998,11 @@ UnpackedRecord *pUnKey, i64 intKey, int bias, int *pRes ); -SQLITE_PRIVATE int sqlite3BtreeCursorHasMoved(BtCursor*); -SQLITE_PRIVATE int sqlite3BtreeCursorRestore(BtCursor*, int*); +SQLITE_PRIVATE int sqlite3BtreeCursorHasMoved(BtCursor*, int*); SQLITE_PRIVATE int sqlite3BtreeDelete(BtCursor*); SQLITE_PRIVATE int sqlite3BtreeInsert(BtCursor*, const void *pKey, i64 nKey, const void *pData, int nData, int nZero, int bias, int seekResult); SQLITE_PRIVATE int sqlite3BtreeFirst(BtCursor*, int *pRes); @@ -9389,55 +9289,55 @@ #define OP_ResultRow 35 /* synopsis: output=r[P1@P2] */ #define OP_CollSeq 36 #define OP_AddImm 37 /* synopsis: r[P1]=r[P1]+P2 */ #define OP_MustBeInt 38 #define OP_RealAffinity 39 -#define OP_Cast 40 /* synopsis: affinity(r[P1]) */ -#define OP_Permutation 41 -#define OP_Compare 42 /* synopsis: r[P1@P3] <-> r[P2@P3] */ -#define OP_Jump 43 -#define OP_Once 44 -#define OP_If 45 -#define OP_IfNot 46 -#define OP_Column 47 /* synopsis: r[P3]=PX */ -#define OP_Affinity 48 /* synopsis: affinity(r[P1@P2]) */ -#define OP_MakeRecord 49 /* synopsis: r[P3]=mkrec(r[P1@P2]) */ -#define OP_Count 50 /* synopsis: r[P2]=count() */ -#define OP_ReadCookie 51 -#define OP_SetCookie 52 -#define OP_ReopenIdx 53 /* synopsis: root=P2 iDb=P3 */ -#define OP_OpenRead 54 /* synopsis: root=P2 iDb=P3 */ -#define OP_OpenWrite 55 /* synopsis: root=P2 iDb=P3 */ -#define OP_OpenAutoindex 56 /* synopsis: nColumn=P2 */ -#define OP_OpenEphemeral 57 /* synopsis: nColumn=P2 */ -#define OP_SorterOpen 58 -#define OP_SequenceTest 59 /* synopsis: if( cursor[P1].ctr++ ) pc = P2 */ -#define OP_OpenPseudo 60 /* synopsis: P3 columns in r[P2] */ -#define OP_Close 61 -#define OP_SeekLT 62 /* synopsis: key=r[P3@P4] */ -#define OP_SeekLE 63 /* synopsis: key=r[P3@P4] */ -#define OP_SeekGE 64 /* synopsis: key=r[P3@P4] */ -#define OP_SeekGT 65 /* synopsis: key=r[P3@P4] */ -#define OP_Seek 66 /* synopsis: intkey=r[P2] */ -#define OP_NoConflict 67 /* synopsis: key=r[P3@P4] */ -#define OP_NotFound 68 /* synopsis: key=r[P3@P4] */ -#define OP_Found 69 /* synopsis: key=r[P3@P4] */ -#define OP_NotExists 70 /* synopsis: intkey=r[P3] */ +#define OP_Permutation 40 +#define OP_Compare 41 /* synopsis: r[P1@P3] <-> r[P2@P3] */ +#define OP_Jump 42 +#define OP_Once 43 +#define OP_If 44 +#define OP_IfNot 45 +#define OP_Column 46 /* synopsis: r[P3]=PX */ +#define OP_Affinity 47 /* synopsis: affinity(r[P1@P2]) */ +#define OP_MakeRecord 48 /* synopsis: r[P3]=mkrec(r[P1@P2]) */ +#define OP_Count 49 /* synopsis: r[P2]=count() */ +#define OP_ReadCookie 50 +#define OP_SetCookie 51 +#define OP_ReopenIdx 52 /* synopsis: root=P2 iDb=P3 */ +#define OP_OpenRead 53 /* synopsis: root=P2 iDb=P3 */ +#define OP_OpenWrite 54 /* synopsis: root=P2 iDb=P3 */ +#define OP_OpenAutoindex 55 /* synopsis: nColumn=P2 */ +#define OP_OpenEphemeral 56 /* synopsis: nColumn=P2 */ +#define OP_SorterOpen 57 +#define OP_OpenPseudo 58 /* synopsis: P3 columns in r[P2] */ +#define OP_Close 59 +#define OP_SeekLT 60 /* synopsis: key=r[P3@P4] */ +#define OP_SeekLE 61 /* synopsis: key=r[P3@P4] */ +#define OP_SeekGE 62 /* synopsis: key=r[P3@P4] */ +#define OP_SeekGT 63 /* synopsis: key=r[P3@P4] */ +#define OP_Seek 64 /* synopsis: intkey=r[P2] */ +#define OP_NoConflict 65 /* synopsis: key=r[P3@P4] */ +#define OP_NotFound 66 /* synopsis: key=r[P3@P4] */ +#define OP_Found 67 /* synopsis: key=r[P3@P4] */ +#define OP_NotExists 68 /* synopsis: intkey=r[P3] */ +#define OP_Sequence 69 /* synopsis: r[P2]=cursor[P1].ctr++ */ +#define OP_NewRowid 70 /* synopsis: r[P2]=rowid */ #define OP_Or 71 /* same as TK_OR, synopsis: r[P3]=(r[P1] || r[P2]) */ #define OP_And 72 /* same as TK_AND, synopsis: r[P3]=(r[P1] && r[P2]) */ -#define OP_Sequence 73 /* synopsis: r[P2]=cursor[P1].ctr++ */ -#define OP_NewRowid 74 /* synopsis: r[P2]=rowid */ -#define OP_Insert 75 /* synopsis: intkey=r[P3] data=r[P2] */ +#define OP_Insert 73 /* synopsis: intkey=r[P3] data=r[P2] */ +#define OP_InsertInt 74 /* synopsis: intkey=P3 data=r[P2] */ +#define OP_Delete 75 #define OP_IsNull 76 /* same as TK_ISNULL, synopsis: if r[P1]==NULL goto P2 */ #define OP_NotNull 77 /* same as TK_NOTNULL, synopsis: if r[P1]!=NULL goto P2 */ #define OP_Ne 78 /* same as TK_NE, synopsis: if r[P1]!=r[P3] goto P2 */ #define OP_Eq 79 /* same as TK_EQ, synopsis: if r[P1]==r[P3] goto P2 */ #define OP_Gt 80 /* same as TK_GT, synopsis: if r[P1]>r[P3] goto P2 */ #define OP_Le 81 /* same as TK_LE, synopsis: if r[P1]<=r[P3] goto P2 */ #define OP_Lt 82 /* same as TK_LT, synopsis: if r[P1] iLevel && i */ - -/* A running thread */ -struct SQLiteThread { - pthread_t tid; /* Thread ID */ - int done; /* Set to true when thread finishes */ - void *pOut; /* Result returned by the thread */ - void *(*xTask)(void*); /* The thread routine */ - void *pIn; /* Argument to the thread */ -}; - -/* Create a new thread */ -SQLITE_PRIVATE int sqlite3ThreadCreate( - SQLiteThread **ppThread, /* OUT: Write the thread object here */ - void *(*xTask)(void*), /* Routine to run in a separate thread */ - void *pIn /* Argument passed into xTask() */ -){ - SQLiteThread *p; - int rc; - - assert( ppThread!=0 ); - assert( xTask!=0 ); - /* This routine is never used in single-threaded mode */ - assert( sqlite3GlobalConfig.bCoreMutex!=0 ); - - *ppThread = 0; - p = sqlite3Malloc(sizeof(*p)); - if( p==0 ) return SQLITE_NOMEM; - memset(p, 0, sizeof(*p)); - p->xTask = xTask; - p->pIn = pIn; - if( sqlite3FaultSim(200) ){ - rc = 1; - }else{ - rc = pthread_create(&p->tid, 0, xTask, pIn); - } - if( rc ){ - p->done = 1; - p->pOut = xTask(pIn); - } - *ppThread = p; - return SQLITE_OK; -} - -/* Get the results of the thread */ -SQLITE_PRIVATE int sqlite3ThreadJoin(SQLiteThread *p, void **ppOut){ - int rc; - - assert( ppOut!=0 ); - if( NEVER(p==0) ) return SQLITE_NOMEM; - if( p->done ){ - *ppOut = p->pOut; - rc = SQLITE_OK; - }else{ - rc = pthread_join(p->tid, ppOut) ? SQLITE_ERROR : SQLITE_OK; - } - sqlite3_free(p); - return rc; -} - -#endif /* SQLITE_OS_UNIX && defined(SQLITE_MUTEX_PTHREADS) */ -/******************************** End Unix Pthreads *************************/ - - -/********************************* Win32 Threads ****************************/ -#if SQLITE_OS_WIN && !SQLITE_OS_WINRT && SQLITE_THREADSAFE>0 - -#define SQLITE_THREADS_IMPLEMENTED 1 /* Prevent the single-thread code below */ -#include - -/* A running thread */ -struct SQLiteThread { - uintptr_t tid; /* The thread handle */ - unsigned id; /* The thread identifier */ - void *(*xTask)(void*); /* The routine to run as a thread */ - void *pIn; /* Argument to xTask */ - void *pResult; /* Result of xTask */ -}; - -/* Thread procedure Win32 compatibility shim */ -static unsigned __stdcall sqlite3ThreadProc( - void *pArg /* IN: Pointer to the SQLiteThread structure */ -){ - SQLiteThread *p = (SQLiteThread *)pArg; - - assert( p!=0 ); -#if 0 - /* - ** This assert appears to trigger spuriously on certain - ** versions of Windows, possibly due to _beginthreadex() - ** and/or CreateThread() not fully setting their thread - ** ID parameter before starting the thread. - */ - assert( p->id==GetCurrentThreadId() ); -#endif - assert( p->xTask!=0 ); - p->pResult = p->xTask(p->pIn); - - _endthreadex(0); - return 0; /* NOT REACHED */ -} - -/* Create a new thread */ -SQLITE_PRIVATE int sqlite3ThreadCreate( - SQLiteThread **ppThread, /* OUT: Write the thread object here */ - void *(*xTask)(void*), /* Routine to run in a separate thread */ - void *pIn /* Argument passed into xTask() */ -){ - SQLiteThread *p; - - assert( ppThread!=0 ); - assert( xTask!=0 ); - *ppThread = 0; - p = sqlite3Malloc(sizeof(*p)); - if( p==0 ) return SQLITE_NOMEM; - if( sqlite3GlobalConfig.bCoreMutex==0 ){ - memset(p, 0, sizeof(*p)); - }else{ - p->xTask = xTask; - p->pIn = pIn; - p->tid = _beginthreadex(0, 0, sqlite3ThreadProc, p, 0, &p->id); - if( p->tid==0 ){ - memset(p, 0, sizeof(*p)); - } - } - if( p->xTask==0 ){ - p->id = GetCurrentThreadId(); - p->pResult = xTask(pIn); - } - *ppThread = p; - return SQLITE_OK; -} - -SQLITE_PRIVATE DWORD sqlite3Win32Wait(HANDLE hObject); /* os_win.c */ - -/* Get the results of the thread */ -SQLITE_PRIVATE int sqlite3ThreadJoin(SQLiteThread *p, void **ppOut){ - DWORD rc; - BOOL bRc; - - assert( ppOut!=0 ); - if( NEVER(p==0) ) return SQLITE_NOMEM; - if( p->xTask==0 ){ - assert( p->id==GetCurrentThreadId() ); - rc = WAIT_OBJECT_0; - assert( p->tid==0 ); - }else{ - assert( p->id!=0 && p->id!=GetCurrentThreadId() ); - rc = sqlite3Win32Wait((HANDLE)p->tid); - assert( rc!=WAIT_IO_COMPLETION ); - bRc = CloseHandle((HANDLE)p->tid); - assert( bRc ); - } - if( rc==WAIT_OBJECT_0 ) *ppOut = p->pResult; - sqlite3_free(p); - return (rc==WAIT_OBJECT_0) ? SQLITE_OK : SQLITE_ERROR; -} - -#endif /* SQLITE_OS_WIN && !SQLITE_OS_WINRT */ -/******************************** End Win32 Threads *************************/ - - -/********************************* Single-Threaded **************************/ -#ifndef SQLITE_THREADS_IMPLEMENTED -/* -** This implementation does not actually create a new thread. It does the -** work of the thread in the main thread, when either the thread is created -** or when it is joined -*/ - -/* A running thread */ -struct SQLiteThread { - void *(*xTask)(void*); /* The routine to run as a thread */ - void *pIn; /* Argument to xTask */ - void *pResult; /* Result of xTask */ -}; - -/* Create a new thread */ -SQLITE_PRIVATE int sqlite3ThreadCreate( - SQLiteThread **ppThread, /* OUT: Write the thread object here */ - void *(*xTask)(void*), /* Routine to run in a separate thread */ - void *pIn /* Argument passed into xTask() */ -){ - SQLiteThread *p; - - assert( ppThread!=0 ); - assert( xTask!=0 ); - *ppThread = 0; - p = sqlite3Malloc(sizeof(*p)); - if( p==0 ) return SQLITE_NOMEM; - if( (SQLITE_PTR_TO_INT(p)/17)&1 ){ - p->xTask = xTask; - p->pIn = pIn; - }else{ - p->xTask = 0; - p->pResult = xTask(pIn); - } - *ppThread = p; - return SQLITE_OK; -} - -/* Get the results of the thread */ -SQLITE_PRIVATE int sqlite3ThreadJoin(SQLiteThread *p, void **ppOut){ - - assert( ppOut!=0 ); - if( NEVER(p==0) ) return SQLITE_NOMEM; - if( p->xTask ){ - *ppOut = p->xTask(p->pIn); - }else{ - *ppOut = p->pResult; - } - sqlite3_free(p); - -#if defined(SQLITE_TEST) - { - void *pTstAlloc = sqlite3Malloc(10); - if (!pTstAlloc) return SQLITE_NOMEM; - sqlite3_free(pTstAlloc); - } -#endif - - return SQLITE_OK; -} - -#endif /* !defined(SQLITE_THREADS_IMPLEMENTED) */ -/****************************** End Single-Threaded *************************/ -#endif /* SQLITE_MAX_WORKER_THREADS>0 */ - -/************** End of threads.c *********************************************/ /************** Begin file utf.c *********************************************/ /* ** 2004 April 13 ** ** The author disclaims copyright to this source code. In place of @@ -22456,11 +21903,11 @@ /* ** This routine transforms the internal text encoding used by pMem to ** desiredEnc. It is an error if the string is already of the desired ** encoding, or if *pMem does not contain a string value. */ -SQLITE_PRIVATE SQLITE_NOINLINE int sqlite3VdbeMemTranslate(Mem *pMem, u8 desiredEnc){ +SQLITE_PRIVATE int sqlite3VdbeMemTranslate(Mem *pMem, u8 desiredEnc){ int len; /* Maximum length of output string in bytes */ unsigned char *zOut; /* Output buffer */ unsigned char *zIn; /* Input iterator */ unsigned char *zTerm; /* End of input */ unsigned char *z; /* Output iterator */ @@ -22571,17 +22018,16 @@ pMem->n = (int)(z - zOut); } *z = 0; assert( (pMem->n+(desiredEnc==SQLITE_UTF8?1:2))<=len ); - c = pMem->flags; sqlite3VdbeMemRelease(pMem); - pMem->flags = MEM_Str|MEM_Term|(c&MEM_AffMask); + pMem->flags &= ~(MEM_Static|MEM_Dyn|MEM_Ephem); pMem->enc = desiredEnc; + pMem->flags |= (MEM_Term); pMem->z = (char*)zOut; pMem->zMalloc = pMem->z; - pMem->szMalloc = sqlite3DbMallocSize(pMem->db, pMem->z); translate_out: #if defined(TRANSLATE_TRACE) && defined(SQLITE_DEBUG) { char zBuf[100]; @@ -22900,19 +22346,10 @@ if( z==0 ) return 0; while( *z2 ){ z2++; } return 0x3fffffff & (int)(z2 - z); } -/* -** Set the current error code to err_code and clear any prior error message. -*/ -SQLITE_PRIVATE void sqlite3Error(sqlite3 *db, int err_code){ - assert( db!=0 ); - db->errCode = err_code; - if( db->pErr ) sqlite3ValueSetNull(db->pErr); -} - /* ** Set the most recent error code and error string for the sqlite ** handle "db". The error code is set to "err_code". ** ** If it is not NULL, string zFormat specifies the format of the @@ -22930,22 +22367,22 @@ ** ** To clear the most recent error for sqlite handle "db", sqlite3Error ** should be called with err_code set to SQLITE_OK and zFormat set ** to NULL. */ -SQLITE_PRIVATE void sqlite3ErrorWithMsg(sqlite3 *db, int err_code, const char *zFormat, ...){ +SQLITE_PRIVATE void sqlite3Error(sqlite3 *db, int err_code, const char *zFormat, ...){ assert( db!=0 ); db->errCode = err_code; - if( zFormat==0 ){ - sqlite3Error(db, err_code); - }else if( db->pErr || (db->pErr = sqlite3ValueNew(db))!=0 ){ + if( zFormat && (db->pErr || (db->pErr = sqlite3ValueNew(db))!=0) ){ char *z; va_list ap; va_start(ap, zFormat); z = sqlite3VMPrintf(db, zFormat, ap); va_end(ap); sqlite3ValueSetStr(db->pErr, -1, z, SQLITE_UTF8, SQLITE_DYNAMIC); + }else if( db->pErr ){ + sqlite3ValueSetNull(db->pErr); } } /* ** Add an error message to pParse->zErrMsg and increment pParse->nErr. @@ -22955,16 +22392,16 @@ ** %z A string that should be freed after use ** %d Insert an integer ** %T Insert a token ** %S Insert the first element of a SrcList ** -** This function should be used to report any error that occurs while +** This function should be used to report any error that occurs whilst ** compiling an SQL statement (i.e. within sqlite3_prepare()). The ** last thing the sqlite3_prepare() function does is copy the error ** stored by this function into the database handle using sqlite3Error(). -** Functions sqlite3Error() or sqlite3ErrorWithMsg() should be used -** during statement execution (sqlite3_step() etc.). +** Function sqlite3Error() should be used during statement execution +** (sqlite3_step() etc.). */ SQLITE_PRIVATE void sqlite3ErrorMsg(Parse *pParse, const char *zFormat, ...){ char *zMsg; va_list ap; sqlite3 *db = pParse->db; @@ -22993,11 +22430,11 @@ ** The return value is -1 if no dequoting occurs or the length of the ** dequoted string, exclusive of the zero terminator, if dequoting does ** occur. ** ** 2002-Feb-14: This routine is extended to remove MS-Access style -** brackets from around identifiers. For example: "[a-b-c]" becomes +** brackets from around identifers. For example: "[a-b-c]" becomes ** "a-b-c". */ SQLITE_PRIVATE int sqlite3Dequote(char *z){ char quote; int i, j; @@ -23497,11 +22934,11 @@ ** A variable-length integer consists of the lower 7 bits of each byte ** for all bytes that have the 8th bit set and one byte with the 8th ** bit clear. Except, if we get to the 9th byte, it stores the full ** 8 bits and is the last byte. */ -static int SQLITE_NOINLINE putVarint64(unsigned char *p, u64 v){ +SQLITE_PRIVATE int sqlite3PutVarint(unsigned char *p, u64 v){ int i, j, n; u8 buf[10]; if( v & (((u64)0xff000000)<<32) ){ p[8] = (u8)v; v >>= 8; @@ -23521,21 +22958,32 @@ for(i=0, j=n-1; j>=0; j--, i++){ p[i] = buf[j]; } return n; } -SQLITE_PRIVATE int sqlite3PutVarint(unsigned char *p, u64 v){ - if( v<=0x7f ){ - p[0] = v&0x7f; + +/* +** This routine is a faster version of sqlite3PutVarint() that only +** works for 32-bit positive integers and which is optimized for +** the common case of small integers. A MACRO version, putVarint32, +** is provided which inlines the single-byte case. All code should use +** the MACRO version as this function assumes the single-byte case has +** already been handled. +*/ +SQLITE_PRIVATE int sqlite3PutVarint32(unsigned char *p, u32 v){ +#ifndef putVarint32 + if( (v & ~0x7f)==0 ){ + p[0] = v; return 1; } - if( v<=0x3fff ){ - p[0] = ((v>>7)&0x7f)|0x80; - p[1] = v&0x7f; +#endif + if( (v & ~0x3fff)==0 ){ + p[0] = (u8)((v>>7) | 0x80); + p[1] = (u8)(v & 0x7f); return 2; } - return putVarint64(p,v); + return sqlite3PutVarint(p, v); } /* ** Bitmasks used by sqlite3GetVarint(). These precomputed constants ** are defined here rather than simply putting the constant expressions @@ -24207,15 +23655,16 @@ } /* ** The hashing function. */ -static unsigned int strHash(const char *z){ +static unsigned int strHash(const char *z, int nKey){ unsigned int h = 0; - unsigned char c; - while( (c = (unsigned char)*z++)!=0 ){ - h = (h<<3) ^ h ^ sqlite3UpperToLower[c]; + assert( nKey>=0 ); + while( nKey > 0 ){ + h = (h<<3) ^ h ^ sqlite3UpperToLower[(unsigned char)*z++]; + nKey--; } return h; } @@ -24283,45 +23732,40 @@ sqlite3_free(pH->ht); pH->ht = new_ht; pH->htsize = new_size = sqlite3MallocSize(new_ht)/sizeof(struct _ht); memset(new_ht, 0, new_size*sizeof(struct _ht)); for(elem=pH->first, pH->first=0; elem; elem = next_elem){ - unsigned int h = strHash(elem->pKey) % new_size; + unsigned int h = strHash(elem->pKey, elem->nKey) % new_size; next_elem = elem->next; insertElement(pH, &new_ht[h], elem); } return 1; } /* This function (for internal use only) locates an element in an -** hash table that matches the given key. The hash for this key is -** also computed and returned in the *pH parameter. +** hash table that matches the given key. The hash for this key has +** already been computed and is passed as the 4th parameter. */ -static HashElem *findElementWithHash( +static HashElem *findElementGivenHash( const Hash *pH, /* The pH to be searched */ const char *pKey, /* The key we are searching for */ - unsigned int *pHash /* Write the hash value here */ + int nKey, /* Bytes in key (not counting zero terminator) */ + unsigned int h /* The hash for this key. */ ){ HashElem *elem; /* Used to loop thru the element list */ int count; /* Number of elements left to test */ - unsigned int h; /* The computed hash */ if( pH->ht ){ - struct _ht *pEntry; - h = strHash(pKey) % pH->htsize; - pEntry = &pH->ht[h]; + struct _ht *pEntry = &pH->ht[h]; elem = pEntry->chain; count = pEntry->count; }else{ - h = 0; elem = pH->first; count = pH->count; } - *pHash = h; - while( count-- ){ - assert( elem!=0 ); - if( sqlite3StrICmp(elem->pKey,pKey)==0 ){ + while( count-- && ALWAYS(elem) ){ + if( elem->nKey==nKey && sqlite3StrNICmp(elem->pKey,pKey,nKey)==0 ){ return elem; } elem = elem->next; } return 0; @@ -24360,24 +23804,30 @@ sqlite3HashClear(pH); } } /* Attempt to locate an element of the hash table pH with a key -** that matches pKey. Return the data for this element if it is +** that matches pKey,nKey. Return the data for this element if it is ** found, or NULL if there is no match. */ -SQLITE_PRIVATE void *sqlite3HashFind(const Hash *pH, const char *pKey){ +SQLITE_PRIVATE void *sqlite3HashFind(const Hash *pH, const char *pKey, int nKey){ HashElem *elem; /* The element that matches key */ unsigned int h; /* A hash on key */ assert( pH!=0 ); assert( pKey!=0 ); - elem = findElementWithHash(pH, pKey, &h); + assert( nKey>=0 ); + if( pH->ht ){ + h = strHash(pKey, nKey) % pH->htsize; + }else{ + h = 0; + } + elem = findElementGivenHash(pH, pKey, nKey, h); return elem ? elem->data : 0; } -/* Insert an element into the hash table pH. The key is pKey +/* Insert an element into the hash table pH. The key is pKey,nKey ** and the data is "data". ** ** If no element exists with a matching key, then a new ** element is created and NULL is returned. ** @@ -24387,41 +23837,53 @@ ** the new data is returned and the hash table is unchanged. ** ** If the "data" parameter to this function is NULL, then the ** element corresponding to "key" is removed from the hash table. */ -SQLITE_PRIVATE void *sqlite3HashInsert(Hash *pH, const char *pKey, void *data){ +SQLITE_PRIVATE void *sqlite3HashInsert(Hash *pH, const char *pKey, int nKey, void *data){ unsigned int h; /* the hash of the key modulo hash table size */ HashElem *elem; /* Used to loop thru the element list */ HashElem *new_elem; /* New element added to the pH */ assert( pH!=0 ); assert( pKey!=0 ); - elem = findElementWithHash(pH,pKey,&h); + assert( nKey>=0 ); + if( pH->htsize ){ + h = strHash(pKey, nKey) % pH->htsize; + }else{ + h = 0; + } + elem = findElementGivenHash(pH,pKey,nKey,h); if( elem ){ void *old_data = elem->data; if( data==0 ){ removeElementGivenHash(pH,elem,h); }else{ elem->data = data; elem->pKey = pKey; + assert(nKey==elem->nKey); } return old_data; } if( data==0 ) return 0; new_elem = (HashElem*)sqlite3Malloc( sizeof(HashElem) ); if( new_elem==0 ) return data; new_elem->pKey = pKey; + new_elem->nKey = nKey; new_elem->data = data; pH->count++; if( pH->count>=10 && pH->count > 2*pH->htsize ){ if( rehash(pH, pH->count*2) ){ assert( pH->htsize>0 ); - h = strHash(pKey) % pH->htsize; + h = strHash(pKey, nKey) % pH->htsize; } } - insertElement(pH, pH->ht ? &pH->ht[h] : 0, new_elem); + if( pH->ht ){ + insertElement(pH, &pH->ht[h], new_elem); + }else{ + insertElement(pH, 0, new_elem); + } return 0; } /************** End of hash.c ************************************************/ /************** Begin file opcodes.c *****************************************/ @@ -24472,55 +23934,55 @@ /* 35 */ "ResultRow" OpHelp("output=r[P1@P2]"), /* 36 */ "CollSeq" OpHelp(""), /* 37 */ "AddImm" OpHelp("r[P1]=r[P1]+P2"), /* 38 */ "MustBeInt" OpHelp(""), /* 39 */ "RealAffinity" OpHelp(""), - /* 40 */ "Cast" OpHelp("affinity(r[P1])"), - /* 41 */ "Permutation" OpHelp(""), - /* 42 */ "Compare" OpHelp("r[P1@P3] <-> r[P2@P3]"), - /* 43 */ "Jump" OpHelp(""), - /* 44 */ "Once" OpHelp(""), - /* 45 */ "If" OpHelp(""), - /* 46 */ "IfNot" OpHelp(""), - /* 47 */ "Column" OpHelp("r[P3]=PX"), - /* 48 */ "Affinity" OpHelp("affinity(r[P1@P2])"), - /* 49 */ "MakeRecord" OpHelp("r[P3]=mkrec(r[P1@P2])"), - /* 50 */ "Count" OpHelp("r[P2]=count()"), - /* 51 */ "ReadCookie" OpHelp(""), - /* 52 */ "SetCookie" OpHelp(""), - /* 53 */ "ReopenIdx" OpHelp("root=P2 iDb=P3"), - /* 54 */ "OpenRead" OpHelp("root=P2 iDb=P3"), - /* 55 */ "OpenWrite" OpHelp("root=P2 iDb=P3"), - /* 56 */ "OpenAutoindex" OpHelp("nColumn=P2"), - /* 57 */ "OpenEphemeral" OpHelp("nColumn=P2"), - /* 58 */ "SorterOpen" OpHelp(""), - /* 59 */ "SequenceTest" OpHelp("if( cursor[P1].ctr++ ) pc = P2"), - /* 60 */ "OpenPseudo" OpHelp("P3 columns in r[P2]"), - /* 61 */ "Close" OpHelp(""), - /* 62 */ "SeekLT" OpHelp("key=r[P3@P4]"), - /* 63 */ "SeekLE" OpHelp("key=r[P3@P4]"), - /* 64 */ "SeekGE" OpHelp("key=r[P3@P4]"), - /* 65 */ "SeekGT" OpHelp("key=r[P3@P4]"), - /* 66 */ "Seek" OpHelp("intkey=r[P2]"), - /* 67 */ "NoConflict" OpHelp("key=r[P3@P4]"), - /* 68 */ "NotFound" OpHelp("key=r[P3@P4]"), - /* 69 */ "Found" OpHelp("key=r[P3@P4]"), - /* 70 */ "NotExists" OpHelp("intkey=r[P3]"), + /* 40 */ "Permutation" OpHelp(""), + /* 41 */ "Compare" OpHelp("r[P1@P3] <-> r[P2@P3]"), + /* 42 */ "Jump" OpHelp(""), + /* 43 */ "Once" OpHelp(""), + /* 44 */ "If" OpHelp(""), + /* 45 */ "IfNot" OpHelp(""), + /* 46 */ "Column" OpHelp("r[P3]=PX"), + /* 47 */ "Affinity" OpHelp("affinity(r[P1@P2])"), + /* 48 */ "MakeRecord" OpHelp("r[P3]=mkrec(r[P1@P2])"), + /* 49 */ "Count" OpHelp("r[P2]=count()"), + /* 50 */ "ReadCookie" OpHelp(""), + /* 51 */ "SetCookie" OpHelp(""), + /* 52 */ "ReopenIdx" OpHelp("root=P2 iDb=P3"), + /* 53 */ "OpenRead" OpHelp("root=P2 iDb=P3"), + /* 54 */ "OpenWrite" OpHelp("root=P2 iDb=P3"), + /* 55 */ "OpenAutoindex" OpHelp("nColumn=P2"), + /* 56 */ "OpenEphemeral" OpHelp("nColumn=P2"), + /* 57 */ "SorterOpen" OpHelp(""), + /* 58 */ "OpenPseudo" OpHelp("P3 columns in r[P2]"), + /* 59 */ "Close" OpHelp(""), + /* 60 */ "SeekLT" OpHelp("key=r[P3@P4]"), + /* 61 */ "SeekLE" OpHelp("key=r[P3@P4]"), + /* 62 */ "SeekGE" OpHelp("key=r[P3@P4]"), + /* 63 */ "SeekGT" OpHelp("key=r[P3@P4]"), + /* 64 */ "Seek" OpHelp("intkey=r[P2]"), + /* 65 */ "NoConflict" OpHelp("key=r[P3@P4]"), + /* 66 */ "NotFound" OpHelp("key=r[P3@P4]"), + /* 67 */ "Found" OpHelp("key=r[P3@P4]"), + /* 68 */ "NotExists" OpHelp("intkey=r[P3]"), + /* 69 */ "Sequence" OpHelp("r[P2]=cursor[P1].ctr++"), + /* 70 */ "NewRowid" OpHelp("r[P2]=rowid"), /* 71 */ "Or" OpHelp("r[P3]=(r[P1] || r[P2])"), /* 72 */ "And" OpHelp("r[P3]=(r[P1] && r[P2])"), - /* 73 */ "Sequence" OpHelp("r[P2]=cursor[P1].ctr++"), - /* 74 */ "NewRowid" OpHelp("r[P2]=rowid"), - /* 75 */ "Insert" OpHelp("intkey=r[P3] data=r[P2]"), + /* 73 */ "Insert" OpHelp("intkey=r[P3] data=r[P2]"), + /* 74 */ "InsertInt" OpHelp("intkey=P3 data=r[P2]"), + /* 75 */ "Delete" OpHelp(""), /* 76 */ "IsNull" OpHelp("if r[P1]==NULL goto P2"), /* 77 */ "NotNull" OpHelp("if r[P1]!=NULL goto P2"), /* 78 */ "Ne" OpHelp("if r[P1]!=r[P3] goto P2"), /* 79 */ "Eq" OpHelp("if r[P1]==r[P3] goto P2"), /* 80 */ "Gt" OpHelp("if r[P1]>r[P3] goto P2"), /* 81 */ "Le" OpHelp("if r[P1]<=r[P3] goto P2"), /* 82 */ "Lt" OpHelp("if r[P1] -conch". -** If the conch file does not exist, or its contents do not match the +** If the conch file does not exist, or it's contents do not match the ** host ID and/or proxy path, then the lock is escalated to an exclusive ** lock and the conch file contents is updated with the host ID and proxy ** path and the lock is downgraded to a shared lock again. If the conch ** is held by another process (with a shared lock), the exclusive lock ** will fail and SQLITE_BUSY is returned. @@ -31193,11 +30629,11 @@ ** ** As mentioned above, when compiled with SQLITE_PREFER_PROXY_LOCKING, ** setting the environment variable SQLITE_FORCE_PROXY_LOCKING to 1 will ** force proxy locking to be used for every database file opened, and 0 ** will force automatic proxy locking to be disabled for all database -** files (explicitly calling the SQLITE_SET_LOCKPROXYFILE pragma or +** files (explicity calling the SQLITE_SET_LOCKPROXYFILE pragma or ** sqlite_file_control API is not affected by SQLITE_FORCE_PROXY_LOCKING). */ /* ** Proxy locking is only available on MacOSX @@ -32955,13 +32391,13 @@ ** ** In order to facilitate testing on a WinNT system, the test fixture ** can manually set this value to 1 to emulate Win98 behavior. */ #ifdef SQLITE_TEST -SQLITE_API LONG SQLITE_WIN32_VOLATILE sqlite3_os_type = 0; +SQLITE_API LONG volatile sqlite3_os_type = 0; #else -static LONG SQLITE_WIN32_VOLATILE sqlite3_os_type = 0; +static LONG volatile sqlite3_os_type = 0; #endif #ifndef SYSCALL # define SYSCALL sqlite3_syscall_ptr #endif @@ -33488,11 +32924,15 @@ #endif #define osWaitForSingleObject ((DWORD(WINAPI*)(HANDLE, \ DWORD))aSyscall[63].pCurrent) +#if SQLITE_OS_WINRT { "WaitForSingleObjectEx", (SYSCALL)WaitForSingleObjectEx, 0 }, +#else + { "WaitForSingleObjectEx", (SYSCALL)0, 0 }, +#endif #define osWaitForSingleObjectEx ((DWORD(WINAPI*)(HANDLE,DWORD, \ BOOL))aSyscall[64].pCurrent) #if SQLITE_OS_WINRT @@ -33596,12 +33036,12 @@ #define osInterlockedCompareExchange InterlockedCompareExchange #else { "InterlockedCompareExchange", (SYSCALL)InterlockedCompareExchange, 0 }, -#define osInterlockedCompareExchange ((LONG(WINAPI*)(LONG \ - SQLITE_WIN32_VOLATILE*, LONG,LONG))aSyscall[76].pCurrent) +#define osInterlockedCompareExchange ((LONG(WINAPI*)(LONG volatile*, \ + LONG,LONG))aSyscall[76].pCurrent) #endif /* defined(InterlockedCompareExchange) */ }; /* End of the overrideable system calls */ /* @@ -33831,19 +33271,10 @@ #else osSleep(milliseconds); #endif } -#if SQLITE_MAX_WORKER_THREADS>0 && !SQLITE_OS_WINRT && SQLITE_THREADSAFE>0 -SQLITE_PRIVATE DWORD sqlite3Win32Wait(HANDLE hObject){ - DWORD rc; - while( (rc = osWaitForSingleObjectEx(hObject, INFINITE, - TRUE))==WAIT_IO_COMPLETION ){} - return rc; -} -#endif - /* ** Return true (non-zero) if we are running under WinNT, Win2K, WinXP, ** or WinCE. Return false (zero) for Win95, Win98, or WinME. ** ** Here is an interesting observation: Win95, Win98, and WinME lack @@ -33867,40 +33298,31 @@ /* ** This function determines if the machine is running a version of Windows ** based on the NT kernel. */ SQLITE_API int sqlite3_win32_is_nt(void){ -#if SQLITE_OS_WINRT - /* - ** NOTE: The WinRT sub-platform is always assumed to be based on the NT - ** kernel. - */ - return 1; -#elif defined(SQLITE_WIN32_GETVERSIONEX) && SQLITE_WIN32_GETVERSIONEX +#if defined(SQLITE_WIN32_GETVERSIONEX) && SQLITE_WIN32_GETVERSIONEX if( osInterlockedCompareExchange(&sqlite3_os_type, 0, 0)==0 ){ -#if defined(SQLITE_WIN32_HAS_ANSI) +#if !SQLITE_OS_WINRT && defined(SQLITE_WIN32_HAS_WIDE) && \ + defined(NTDDI_VERSION) && NTDDI_VERSION >= NTDDI_WIN8 + OSVERSIONINFOW sInfo; + sInfo.dwOSVersionInfoSize = sizeof(sInfo); + osGetVersionExW(&sInfo); + osInterlockedCompareExchange(&sqlite3_os_type, + (sInfo.dwPlatformId == VER_PLATFORM_WIN32_NT) ? 2 : 1, 0); +#elif defined(SQLITE_WIN32_HAS_ANSI) OSVERSIONINFOA sInfo; sInfo.dwOSVersionInfoSize = sizeof(sInfo); osGetVersionExA(&sInfo); - osInterlockedCompareExchange(&sqlite3_os_type, - (sInfo.dwPlatformId == VER_PLATFORM_WIN32_NT) ? 2 : 1, 0); -#elif defined(SQLITE_WIN32_HAS_WIDE) - OSVERSIONINFOW sInfo; - sInfo.dwOSVersionInfoSize = sizeof(sInfo); - osGetVersionExW(&sInfo); osInterlockedCompareExchange(&sqlite3_os_type, (sInfo.dwPlatformId == VER_PLATFORM_WIN32_NT) ? 2 : 1, 0); #endif } return osInterlockedCompareExchange(&sqlite3_os_type, 2, 2)==2; #elif SQLITE_TEST return osInterlockedCompareExchange(&sqlite3_os_type, 2, 2)==2; #else - /* - ** NOTE: All sub-platforms where the GetVersionEx[AW] functions are - ** deprecated are always assumed to be based on the NT kernel. - */ return 1; #endif } #ifdef SQLITE_WIN32_MALLOC @@ -35673,11 +35095,11 @@ pFile->h, pFile->locktype, sqlite3ErrName(rc))); return rc; } /* -** If *pArg is initially negative then this is a query. Set *pArg to +** If *pArg is inititially negative then this is a query. Set *pArg to ** 1 or 0 depending on whether or not bit mask of pFile->ctrlFlags is set. ** ** If *pArg is 0 or 1, then clear or set the mask bit of pFile->ctrlFlags. */ static void winModeBit(winFile *pFile, unsigned char mask, int *pArg){ @@ -36687,11 +36109,11 @@ if( p ){ pFd->nFetchOut--; }else{ /* FIXME: If Windows truly always prevents truncating or deleting a ** file while a mapping is held, then the following winUnmapfile() call - ** is unnecessary can be omitted - potentially improving + ** is unnecessary can can be omitted - potentially improving ** performance. */ winUnmapfile(pFd); } assert( pFd->nFetchOut>=0 ); @@ -38545,98 +37967,105 @@ # define expensive_assert(X) #endif /********************************** Linked List Management ********************/ -/* Allowed values for second argument to pcacheManageDirtyList() */ -#define PCACHE_DIRTYLIST_REMOVE 1 /* Remove pPage from dirty list */ -#define PCACHE_DIRTYLIST_ADD 2 /* Add pPage to the dirty list */ -#define PCACHE_DIRTYLIST_FRONT 3 /* Move pPage to the front of the list */ +#if !defined(NDEBUG) && defined(SQLITE_ENABLE_EXPENSIVE_ASSERT) +/* +** Check that the pCache->pSynced variable is set correctly. If it +** is not, either fail an assert or return zero. Otherwise, return +** non-zero. This is only used in debugging builds, as follows: +** +** expensive_assert( pcacheCheckSynced(pCache) ); +*/ +static int pcacheCheckSynced(PCache *pCache){ + PgHdr *p; + for(p=pCache->pDirtyTail; p!=pCache->pSynced; p=p->pDirtyPrev){ + assert( p->nRef || (p->flags&PGHDR_NEED_SYNC) ); + } + return (p==0 || p->nRef || (p->flags&PGHDR_NEED_SYNC)==0); +} +#endif /* !NDEBUG && SQLITE_ENABLE_EXPENSIVE_ASSERT */ /* -** Manage pPage's participation on the dirty list. Bits of the addRemove -** argument determines what operation to do. The 0x01 bit means first -** remove pPage from the dirty list. The 0x02 means add pPage back to -** the dirty list. Doing both moves pPage to the front of the dirty list. +** Remove page pPage from the list of dirty pages. */ -static void pcacheManageDirtyList(PgHdr *pPage, u8 addRemove){ +static void pcacheRemoveFromDirtyList(PgHdr *pPage){ + PCache *p = pPage->pCache; + + assert( pPage->pDirtyNext || pPage==p->pDirtyTail ); + assert( pPage->pDirtyPrev || pPage==p->pDirty ); + + /* Update the PCache1.pSynced variable if necessary. */ + if( p->pSynced==pPage ){ + PgHdr *pSynced = pPage->pDirtyPrev; + while( pSynced && (pSynced->flags&PGHDR_NEED_SYNC) ){ + pSynced = pSynced->pDirtyPrev; + } + p->pSynced = pSynced; + } + + if( pPage->pDirtyNext ){ + pPage->pDirtyNext->pDirtyPrev = pPage->pDirtyPrev; + }else{ + assert( pPage==p->pDirtyTail ); + p->pDirtyTail = pPage->pDirtyPrev; + } + if( pPage->pDirtyPrev ){ + pPage->pDirtyPrev->pDirtyNext = pPage->pDirtyNext; + }else{ + assert( pPage==p->pDirty ); + p->pDirty = pPage->pDirtyNext; + if( p->pDirty==0 && p->bPurgeable ){ + assert( p->eCreate==1 ); + p->eCreate = 2; + } + } + pPage->pDirtyNext = 0; + pPage->pDirtyPrev = 0; + + expensive_assert( pcacheCheckSynced(p) ); +} + +/* +** Add page pPage to the head of the dirty list (PCache1.pDirty is set to +** pPage). +*/ +static void pcacheAddToDirtyList(PgHdr *pPage){ PCache *p = pPage->pCache; - if( addRemove & PCACHE_DIRTYLIST_REMOVE ){ - assert( pPage->pDirtyNext || pPage==p->pDirtyTail ); - assert( pPage->pDirtyPrev || pPage==p->pDirty ); - - /* Update the PCache1.pSynced variable if necessary. */ - if( p->pSynced==pPage ){ - PgHdr *pSynced = pPage->pDirtyPrev; - while( pSynced && (pSynced->flags&PGHDR_NEED_SYNC) ){ - pSynced = pSynced->pDirtyPrev; - } - p->pSynced = pSynced; - } - - if( pPage->pDirtyNext ){ - pPage->pDirtyNext->pDirtyPrev = pPage->pDirtyPrev; - }else{ - assert( pPage==p->pDirtyTail ); - p->pDirtyTail = pPage->pDirtyPrev; - } - if( pPage->pDirtyPrev ){ - pPage->pDirtyPrev->pDirtyNext = pPage->pDirtyNext; - }else{ - assert( pPage==p->pDirty ); - p->pDirty = pPage->pDirtyNext; - if( p->pDirty==0 && p->bPurgeable ){ - assert( p->eCreate==1 ); - p->eCreate = 2; - } - } - pPage->pDirtyNext = 0; - pPage->pDirtyPrev = 0; - } - if( addRemove & PCACHE_DIRTYLIST_ADD ){ - assert( pPage->pDirtyNext==0 && pPage->pDirtyPrev==0 && p->pDirty!=pPage ); - - pPage->pDirtyNext = p->pDirty; - if( pPage->pDirtyNext ){ - assert( pPage->pDirtyNext->pDirtyPrev==0 ); - pPage->pDirtyNext->pDirtyPrev = pPage; - }else{ - p->pDirtyTail = pPage; - if( p->bPurgeable ){ - assert( p->eCreate==2 ); - p->eCreate = 1; - } - } - p->pDirty = pPage; - if( !p->pSynced && 0==(pPage->flags&PGHDR_NEED_SYNC) ){ - p->pSynced = pPage; - } - } + assert( pPage->pDirtyNext==0 && pPage->pDirtyPrev==0 && p->pDirty!=pPage ); + + pPage->pDirtyNext = p->pDirty; + if( pPage->pDirtyNext ){ + assert( pPage->pDirtyNext->pDirtyPrev==0 ); + pPage->pDirtyNext->pDirtyPrev = pPage; + }else if( p->bPurgeable ){ + assert( p->eCreate==2 ); + p->eCreate = 1; + } + p->pDirty = pPage; + if( !p->pDirtyTail ){ + p->pDirtyTail = pPage; + } + if( !p->pSynced && 0==(pPage->flags&PGHDR_NEED_SYNC) ){ + p->pSynced = pPage; + } + expensive_assert( pcacheCheckSynced(p) ); } /* ** Wrapper around the pluggable caches xUnpin method. If the cache is ** being used for an in-memory database, this function is a no-op. */ static void pcacheUnpin(PgHdr *p){ - if( p->pCache->bPurgeable ){ + PCache *pCache = p->pCache; + if( pCache->bPurgeable ){ if( p->pgno==1 ){ - p->pCache->pPage1 = 0; - } - sqlite3GlobalConfig.pcache2.xUnpin(p->pCache->pCache, p->pPage, 0); - } -} - -/* -** Compute the number of pages of cache requested. -*/ -static int numberOfCachePages(PCache *p){ - if( p->szCache>=0 ){ - return p->szCache; - }else{ - return (int)((-1024*(i64)p->szCache)/(p->szPage+p->szExtra)); + pCache->pPage1 = 0; + } + sqlite3GlobalConfig.pcache2.xUnpin(pCache->pCache, p->pPage, 0); } } /*************************************************** General Interfaces ****** ** @@ -38668,224 +38097,179 @@ ** Create a new PCache object. Storage space to hold the object ** has already been allocated and is passed in as the p pointer. ** The caller discovers how much space needs to be allocated by ** calling sqlite3PcacheSize(). */ -SQLITE_PRIVATE int sqlite3PcacheOpen( +SQLITE_PRIVATE void sqlite3PcacheOpen( int szPage, /* Size of every page */ int szExtra, /* Extra space associated with each page */ int bPurgeable, /* True if pages are on backing store */ int (*xStress)(void*,PgHdr*),/* Call to try to make pages clean */ void *pStress, /* Argument to xStress */ PCache *p /* Preallocated space for the PCache */ ){ memset(p, 0, sizeof(PCache)); - p->szPage = 1; + p->szPage = szPage; p->szExtra = szExtra; p->bPurgeable = bPurgeable; p->eCreate = 2; p->xStress = xStress; p->pStress = pStress; p->szCache = 100; - return sqlite3PcacheSetPageSize(p, szPage); } /* ** Change the page size for PCache object. The caller must ensure that there ** are no outstanding page references when this function is called. */ -SQLITE_PRIVATE int sqlite3PcacheSetPageSize(PCache *pCache, int szPage){ +SQLITE_PRIVATE void sqlite3PcacheSetPageSize(PCache *pCache, int szPage){ assert( pCache->nRef==0 && pCache->pDirty==0 ); - if( pCache->szPage ){ - sqlite3_pcache *pNew; - pNew = sqlite3GlobalConfig.pcache2.xCreate( - szPage, pCache->szExtra + sizeof(PgHdr), pCache->bPurgeable - ); - if( pNew==0 ) return SQLITE_NOMEM; - sqlite3GlobalConfig.pcache2.xCachesize(pNew, numberOfCachePages(pCache)); - if( pCache->pCache ){ - sqlite3GlobalConfig.pcache2.xDestroy(pCache->pCache); - } - pCache->pCache = pNew; + if( pCache->pCache ){ + sqlite3GlobalConfig.pcache2.xDestroy(pCache->pCache); + pCache->pCache = 0; pCache->pPage1 = 0; - pCache->szPage = szPage; } - return SQLITE_OK; + pCache->szPage = szPage; +} + +/* +** Compute the number of pages of cache requested. +*/ +static int numberOfCachePages(PCache *p){ + if( p->szCache>=0 ){ + return p->szCache; + }else{ + return (int)((-1024*(i64)p->szCache)/(p->szPage+p->szExtra)); + } } /* ** Try to obtain a page from the cache. -** -** This routine returns a pointer to an sqlite3_pcache_page object if -** such an object is already in cache, or if a new one is created. -** This routine returns a NULL pointer if the object was not in cache -** and could not be created. -** -** The createFlags should be 0 to check for existing pages and should -** be 3 (not 1, but 3) to try to create a new page. -** -** If the createFlag is 0, then NULL is always returned if the page -** is not already in the cache. If createFlag is 1, then a new page -** is created only if that can be done without spilling dirty pages -** and without exceeding the cache size limit. -** -** The caller needs to invoke sqlite3PcacheFetchFinish() to properly -** initialize the sqlite3_pcache_page object and convert it into a -** PgHdr object. The sqlite3PcacheFetch() and sqlite3PcacheFetchFinish() -** routines are split this way for performance reasons. When separated -** they can both (usually) operate without having to push values to -** the stack on entry and pop them back off on exit, which saves a -** lot of pushing and popping. */ -SQLITE_PRIVATE sqlite3_pcache_page *sqlite3PcacheFetch( +SQLITE_PRIVATE int sqlite3PcacheFetch( PCache *pCache, /* Obtain the page from this cache */ Pgno pgno, /* Page number to obtain */ - int createFlag /* If true, create page if it does not exist already */ + int createFlag, /* If true, create page if it does not exist already */ + PgHdr **ppPage /* Write the page here */ ){ + sqlite3_pcache_page *pPage; + PgHdr *pPgHdr = 0; int eCreate; assert( pCache!=0 ); - assert( pCache->pCache!=0 ); - assert( createFlag==3 || createFlag==0 ); + assert( createFlag==1 || createFlag==0 ); assert( pgno>0 ); + + /* If the pluggable cache (sqlite3_pcache*) has not been allocated, + ** allocate it now. + */ + if( !pCache->pCache ){ + sqlite3_pcache *p; + if( !createFlag ){ + *ppPage = 0; + return SQLITE_OK; + } + p = sqlite3GlobalConfig.pcache2.xCreate( + pCache->szPage, pCache->szExtra + sizeof(PgHdr), pCache->bPurgeable + ); + if( !p ){ + return SQLITE_NOMEM; + } + sqlite3GlobalConfig.pcache2.xCachesize(p, numberOfCachePages(pCache)); + pCache->pCache = p; + } /* eCreate defines what to do if the page does not exist. ** 0 Do not allocate a new page. (createFlag==0) ** 1 Allocate a new page if doing so is inexpensive. ** (createFlag==1 AND bPurgeable AND pDirty) ** 2 Allocate a new page even it doing so is difficult. ** (createFlag==1 AND !(bPurgeable AND pDirty) */ - eCreate = createFlag & pCache->eCreate; - assert( eCreate==0 || eCreate==1 || eCreate==2 ); - assert( createFlag==0 || pCache->eCreate==eCreate ); - assert( createFlag==0 || eCreate==1+(!pCache->bPurgeable||!pCache->pDirty) ); - return sqlite3GlobalConfig.pcache2.xFetch(pCache->pCache, pgno, eCreate); -} - -/* -** If the sqlite3PcacheFetch() routine is unable to allocate a new -** page because new clean pages are available for reuse and the cache -** size limit has been reached, then this routine can be invoked to -** try harder to allocate a page. This routine might invoke the stress -** callback to spill dirty pages to the journal. It will then try to -** allocate the new page and will only fail to allocate a new page on -** an OOM error. -** -** This routine should be invoked only after sqlite3PcacheFetch() fails. -*/ -SQLITE_PRIVATE int sqlite3PcacheFetchStress( - PCache *pCache, /* Obtain the page from this cache */ - Pgno pgno, /* Page number to obtain */ - sqlite3_pcache_page **ppPage /* Write result here */ -){ - PgHdr *pPg; - if( pCache->eCreate==2 ) return 0; - - - /* Find a dirty page to write-out and recycle. First try to find a - ** page that does not require a journal-sync (one with PGHDR_NEED_SYNC - ** cleared), but if that is not possible settle for any other - ** unreferenced dirty page. - */ - for(pPg=pCache->pSynced; - pPg && (pPg->nRef || (pPg->flags&PGHDR_NEED_SYNC)); - pPg=pPg->pDirtyPrev - ); - pCache->pSynced = pPg; - if( !pPg ){ - for(pPg=pCache->pDirtyTail; pPg && pPg->nRef; pPg=pPg->pDirtyPrev); - } - if( pPg ){ - int rc; + eCreate = createFlag==0 ? 0 : pCache->eCreate; + assert( (createFlag*(1+(!pCache->bPurgeable||!pCache->pDirty)))==eCreate ); + pPage = sqlite3GlobalConfig.pcache2.xFetch(pCache->pCache, pgno, eCreate); + if( !pPage && eCreate==1 ){ + PgHdr *pPg; + + /* Find a dirty page to write-out and recycle. First try to find a + ** page that does not require a journal-sync (one with PGHDR_NEED_SYNC + ** cleared), but if that is not possible settle for any other + ** unreferenced dirty page. + */ + expensive_assert( pcacheCheckSynced(pCache) ); + for(pPg=pCache->pSynced; + pPg && (pPg->nRef || (pPg->flags&PGHDR_NEED_SYNC)); + pPg=pPg->pDirtyPrev + ); + pCache->pSynced = pPg; + if( !pPg ){ + for(pPg=pCache->pDirtyTail; pPg && pPg->nRef; pPg=pPg->pDirtyPrev); + } + if( pPg ){ + int rc; #ifdef SQLITE_LOG_CACHE_SPILL - sqlite3_log(SQLITE_FULL, - "spill page %d making room for %d - cache used: %d/%d", - pPg->pgno, pgno, - sqlite3GlobalConfig.pcache.xPagecount(pCache->pCache), - numberOfCachePages(pCache)); -#endif - rc = pCache->xStress(pCache->pStress, pPg); - if( rc!=SQLITE_OK && rc!=SQLITE_BUSY ){ - return rc; - } - } - *ppPage = sqlite3GlobalConfig.pcache2.xFetch(pCache->pCache, pgno, 2); - return *ppPage==0 ? SQLITE_NOMEM : SQLITE_OK; -} - -/* -** This is a helper routine for sqlite3PcacheFetchFinish() -** -** In the uncommon case where the page being fetched has not been -** initialized, this routine is invoked to do the initialization. -** This routine is broken out into a separate function since it -** requires extra stack manipulation that can be avoided in the common -** case. -*/ -static SQLITE_NOINLINE PgHdr *pcacheFetchFinishWithInit( - PCache *pCache, /* Obtain the page from this cache */ - Pgno pgno, /* Page number obtained */ - sqlite3_pcache_page *pPage /* Page obtained by prior PcacheFetch() call */ -){ - PgHdr *pPgHdr; - assert( pPage!=0 ); - pPgHdr = (PgHdr*)pPage->pExtra; - assert( pPgHdr->pPage==0 ); - memset(pPgHdr, 0, sizeof(PgHdr)); - pPgHdr->pPage = pPage; - pPgHdr->pData = pPage->pBuf; - pPgHdr->pExtra = (void *)&pPgHdr[1]; - memset(pPgHdr->pExtra, 0, pCache->szExtra); - pPgHdr->pCache = pCache; - pPgHdr->pgno = pgno; - return sqlite3PcacheFetchFinish(pCache,pgno,pPage); -} - -/* -** This routine converts the sqlite3_pcache_page object returned by -** sqlite3PcacheFetch() into an initialized PgHdr object. This routine -** must be called after sqlite3PcacheFetch() in order to get a usable -** result. -*/ -SQLITE_PRIVATE PgHdr *sqlite3PcacheFetchFinish( - PCache *pCache, /* Obtain the page from this cache */ - Pgno pgno, /* Page number obtained */ - sqlite3_pcache_page *pPage /* Page obtained by prior PcacheFetch() call */ -){ - PgHdr *pPgHdr; - - if( pPage==0 ) return 0; - pPgHdr = (PgHdr *)pPage->pExtra; - - if( !pPgHdr->pPage ){ - return pcacheFetchFinishWithInit(pCache, pgno, pPage); - } - if( 0==pPgHdr->nRef ){ - pCache->nRef++; - } - pPgHdr->nRef++; - if( pgno==1 ){ - pCache->pPage1 = pPgHdr; - } - return pPgHdr; + sqlite3_log(SQLITE_FULL, + "spill page %d making room for %d - cache used: %d/%d", + pPg->pgno, pgno, + sqlite3GlobalConfig.pcache.xPagecount(pCache->pCache), + numberOfCachePages(pCache)); +#endif + rc = pCache->xStress(pCache->pStress, pPg); + if( rc!=SQLITE_OK && rc!=SQLITE_BUSY ){ + return rc; + } + } + + pPage = sqlite3GlobalConfig.pcache2.xFetch(pCache->pCache, pgno, 2); + } + + if( pPage ){ + pPgHdr = (PgHdr *)pPage->pExtra; + + if( !pPgHdr->pPage ){ + memset(pPgHdr, 0, sizeof(PgHdr)); + pPgHdr->pPage = pPage; + pPgHdr->pData = pPage->pBuf; + pPgHdr->pExtra = (void *)&pPgHdr[1]; + memset(pPgHdr->pExtra, 0, pCache->szExtra); + pPgHdr->pCache = pCache; + pPgHdr->pgno = pgno; + } + assert( pPgHdr->pCache==pCache ); + assert( pPgHdr->pgno==pgno ); + assert( pPgHdr->pData==pPage->pBuf ); + assert( pPgHdr->pExtra==(void *)&pPgHdr[1] ); + + if( 0==pPgHdr->nRef ){ + pCache->nRef++; + } + pPgHdr->nRef++; + if( pgno==1 ){ + pCache->pPage1 = pPgHdr; + } + } + *ppPage = pPgHdr; + return (pPgHdr==0 && eCreate) ? SQLITE_NOMEM : SQLITE_OK; } /* ** Decrement the reference count on a page. If the page is clean and the -** reference count drops to 0, then it is made eligible for recycling. +** reference count drops to 0, then it is made elible for recycling. */ -SQLITE_PRIVATE void SQLITE_NOINLINE sqlite3PcacheRelease(PgHdr *p){ +SQLITE_PRIVATE void sqlite3PcacheRelease(PgHdr *p){ assert( p->nRef>0 ); p->nRef--; if( p->nRef==0 ){ - p->pCache->nRef--; + PCache *pCache = p->pCache; + pCache->nRef--; if( (p->flags&PGHDR_DIRTY)==0 ){ pcacheUnpin(p); - }else if( p->pDirtyPrev!=0 ){ + }else{ /* Move the page to the head of the dirty list. */ - pcacheManageDirtyList(p, PCACHE_DIRTYLIST_FRONT); + pcacheRemoveFromDirtyList(p); + pcacheAddToDirtyList(p); } } } /* @@ -38900,19 +38284,21 @@ ** Drop a page from the cache. There must be exactly one reference to the ** page. This function deletes that reference, so after it returns the ** page pointed to by p is invalid. */ SQLITE_PRIVATE void sqlite3PcacheDrop(PgHdr *p){ + PCache *pCache; assert( p->nRef==1 ); if( p->flags&PGHDR_DIRTY ){ - pcacheManageDirtyList(p, PCACHE_DIRTYLIST_REMOVE); + pcacheRemoveFromDirtyList(p); } - p->pCache->nRef--; + pCache = p->pCache; + pCache->nRef--; if( p->pgno==1 ){ - p->pCache->pPage1 = 0; + pCache->pPage1 = 0; } - sqlite3GlobalConfig.pcache2.xUnpin(p->pCache->pCache, p->pPage, 1); + sqlite3GlobalConfig.pcache2.xUnpin(pCache->pCache, p->pPage, 1); } /* ** Make sure the page is marked as dirty. If it isn't dirty already, ** make it so. @@ -38920,21 +38306,21 @@ SQLITE_PRIVATE void sqlite3PcacheMakeDirty(PgHdr *p){ p->flags &= ~PGHDR_DONT_WRITE; assert( p->nRef>0 ); if( 0==(p->flags & PGHDR_DIRTY) ){ p->flags |= PGHDR_DIRTY; - pcacheManageDirtyList(p, PCACHE_DIRTYLIST_ADD); + pcacheAddToDirtyList( p); } } /* ** Make sure the page is marked as clean. If it isn't clean already, ** make it so. */ SQLITE_PRIVATE void sqlite3PcacheMakeClean(PgHdr *p){ if( (p->flags & PGHDR_DIRTY) ){ - pcacheManageDirtyList(p, PCACHE_DIRTYLIST_REMOVE); + pcacheRemoveFromDirtyList(p); p->flags &= ~(PGHDR_DIRTY|PGHDR_NEED_SYNC); if( p->nRef==0 ){ pcacheUnpin(p); } } @@ -38969,11 +38355,12 @@ assert( p->nRef>0 ); assert( newPgno>0 ); sqlite3GlobalConfig.pcache2.xRekey(pCache->pCache, p->pPage, p->pgno,newPgno); p->pgno = newPgno; if( (p->flags&PGHDR_DIRTY) && (p->flags&PGHDR_NEED_SYNC) ){ - pcacheManageDirtyList(p, PCACHE_DIRTYLIST_FRONT); + pcacheRemoveFromDirtyList(p); + pcacheAddToDirtyList(p); } } /* ** Drop every cache entry whose page number is greater than "pgno". The @@ -39010,12 +38397,13 @@ /* ** Close a cache. */ SQLITE_PRIVATE void sqlite3PcacheClose(PCache *pCache){ - assert( pCache->pCache!=0 ); - sqlite3GlobalConfig.pcache2.xDestroy(pCache->pCache); + if( pCache->pCache ){ + sqlite3GlobalConfig.pcache2.xDestroy(pCache->pCache); + } } /* ** Discard the contents of the cache. */ @@ -39120,12 +38508,15 @@ /* ** Return the total number of pages in the cache. */ SQLITE_PRIVATE int sqlite3PcachePagecount(PCache *pCache){ - assert( pCache->pCache!=0 ); - return sqlite3GlobalConfig.pcache2.xPagecount(pCache->pCache); + int nPage = 0; + if( pCache->pCache ){ + nPage = sqlite3GlobalConfig.pcache2.xPagecount(pCache->pCache); + } + return nPage; } #ifdef SQLITE_TEST /* ** Get the suggested cache-size value. @@ -39137,22 +38528,24 @@ /* ** Set the suggested cache-size value. */ SQLITE_PRIVATE void sqlite3PcacheSetCachesize(PCache *pCache, int mxPage){ - assert( pCache->pCache!=0 ); pCache->szCache = mxPage; - sqlite3GlobalConfig.pcache2.xCachesize(pCache->pCache, - numberOfCachePages(pCache)); + if( pCache->pCache ){ + sqlite3GlobalConfig.pcache2.xCachesize(pCache->pCache, + numberOfCachePages(pCache)); + } } /* ** Free up as much memory as possible from the page cache. */ SQLITE_PRIVATE void sqlite3PcacheShrink(PCache *pCache){ - assert( pCache->pCache!=0 ); - sqlite3GlobalConfig.pcache2.xShrink(pCache->pCache); + if( pCache->pCache ){ + sqlite3GlobalConfig.pcache2.xShrink(pCache->pCache); + } } #if defined(SQLITE_CHECK_PAGES) || defined(SQLITE_DEBUG) /* ** For all dirty pages currently in the cache, invoke the specified @@ -39182,11 +38575,11 @@ ************************************************************************* ** ** This file implements the default page cache implementation (the ** sqlite3_pcache interface). It also contains part of the implementation ** of the SQLITE_CONFIG_PAGECACHE and sqlite3_release_memory() features. -** If the default page cache implementation is overridden, then neither of +** If the default page cache implementation is overriden, then neither of ** these two features are available. */ typedef struct PCache1 PCache1; @@ -39193,11 +38586,11 @@ typedef struct PgHdr1 PgHdr1; typedef struct PgFreeslot PgFreeslot; typedef struct PGroup PGroup; /* Each page cache (or PCache) belongs to a PGroup. A PGroup is a set -** of one or more PCaches that are able to recycle each other's unpinned +** of one or more PCaches that are able to recycle each others unpinned ** pages when they are under memory pressure. A PGroup is an instance of ** the following object. ** ** This page cache implementation works in one of two modes: ** @@ -39551,11 +38944,11 @@ ** This function is used to resize the hash table used by the cache passed ** as the first argument. ** ** The PCache mutex must be held when this function is called. */ -static void pcache1ResizeHash(PCache1 *p){ +static int pcache1ResizeHash(PCache1 *p){ PgHdr1 **apNew; unsigned int nNew; unsigned int i; assert( sqlite3_mutex_held(p->pGroup->mutex) ); @@ -39583,10 +38976,12 @@ } sqlite3_free(p->apHash); p->apHash = apNew; p->nHash = nNew; } + + return (p->apHash ? SQLITE_OK : SQLITE_NOMEM); } /* ** This function is used internally to remove the page pPage from the ** PGroup LRU list, if is part of it. If pPage is not part of the PGroup @@ -39717,13 +39112,10 @@ UNUSED_PARAMETER(NotUsed); assert( pcache1.isInit!=0 ); memset(&pcache1, 0, sizeof(pcache1)); } -/* forward declaration */ -static void pcache1Destroy(sqlite3_pcache *p); - /* ** Implementation of the sqlite3_pcache.xCreate method. ** ** Allocate a new cache. */ @@ -39764,21 +39156,16 @@ } pCache->pGroup = pGroup; pCache->szPage = szPage; pCache->szExtra = szExtra; pCache->bPurgeable = (bPurgeable ? 1 : 0); - pcache1EnterMutex(pGroup); - pcache1ResizeHash(pCache); if( bPurgeable ){ pCache->nMin = 10; + pcache1EnterMutex(pGroup); pGroup->nMinPage += pCache->nMin; pGroup->mxPinned = pGroup->nMaxPage + 10 - pGroup->nMinPage; - } - pcache1LeaveMutex(pGroup); - if( pCache->nHash==0 ){ - pcache1Destroy((sqlite3_pcache*)pCache); - pCache = 0; + pcache1LeaveMutex(pGroup); } } return (sqlite3_pcache *)pCache; } @@ -39830,99 +39217,10 @@ n = pCache->nPage; pcache1LeaveMutex(pCache->pGroup); return n; } - -/* -** Implement steps 3, 4, and 5 of the pcache1Fetch() algorithm described -** in the header of the pcache1Fetch() procedure. -** -** This steps are broken out into a separate procedure because they are -** usually not needed, and by avoiding the stack initialization required -** for these steps, the main pcache1Fetch() procedure can run faster. -*/ -static SQLITE_NOINLINE PgHdr1 *pcache1FetchStage2( - PCache1 *pCache, - unsigned int iKey, - int createFlag -){ - unsigned int nPinned; - PGroup *pGroup = pCache->pGroup; - PgHdr1 *pPage = 0; - - /* Step 3: Abort if createFlag is 1 but the cache is nearly full */ - assert( pCache->nPage >= pCache->nRecyclable ); - nPinned = pCache->nPage - pCache->nRecyclable; - assert( pGroup->mxPinned == pGroup->nMaxPage + 10 - pGroup->nMinPage ); - assert( pCache->n90pct == pCache->nMax*9/10 ); - if( createFlag==1 && ( - nPinned>=pGroup->mxPinned - || nPinned>=pCache->n90pct - || pcache1UnderMemoryPressure(pCache) - )){ - return 0; - } - - if( pCache->nPage>=pCache->nHash ) pcache1ResizeHash(pCache); - assert( pCache->nHash>0 && pCache->apHash ); - - /* Step 4. Try to recycle a page. */ - if( pCache->bPurgeable && pGroup->pLruTail && ( - (pCache->nPage+1>=pCache->nMax) - || pGroup->nCurrentPage>=pGroup->nMaxPage - || pcache1UnderMemoryPressure(pCache) - )){ - PCache1 *pOther; - pPage = pGroup->pLruTail; - assert( pPage->isPinned==0 ); - pcache1RemoveFromHash(pPage); - pcache1PinPage(pPage); - pOther = pPage->pCache; - - /* We want to verify that szPage and szExtra are the same for pOther - ** and pCache. Assert that we can verify this by comparing sums. */ - assert( (pCache->szPage & (pCache->szPage-1))==0 && pCache->szPage>=512 ); - assert( pCache->szExtra<512 ); - assert( (pOther->szPage & (pOther->szPage-1))==0 && pOther->szPage>=512 ); - assert( pOther->szExtra<512 ); - - if( pOther->szPage+pOther->szExtra != pCache->szPage+pCache->szExtra ){ - pcache1FreePage(pPage); - pPage = 0; - }else{ - pGroup->nCurrentPage -= (pOther->bPurgeable - pCache->bPurgeable); - } - } - - /* Step 5. If a usable page buffer has still not been found, - ** attempt to allocate a new one. - */ - if( !pPage ){ - if( createFlag==1 ) sqlite3BeginBenignMalloc(); - pPage = pcache1AllocPage(pCache); - if( createFlag==1 ) sqlite3EndBenignMalloc(); - } - - if( pPage ){ - unsigned int h = iKey % pCache->nHash; - pCache->nPage++; - pPage->iKey = iKey; - pPage->pNext = pCache->apHash[h]; - pPage->pCache = pCache; - pPage->pLruPrev = 0; - pPage->pLruNext = 0; - pPage->isPinned = 1; - *(void **)pPage->page.pExtra = 0; - pCache->apHash[h] = pPage; - if( iKey>pCache->iMaxKey ){ - pCache->iMaxKey = iKey; - } - } - return pPage; -} - /* ** Implementation of the sqlite3_pcache.xFetch method. ** ** Fetch a page by key value. ** @@ -39978,34 +39276,121 @@ static sqlite3_pcache_page *pcache1Fetch( sqlite3_pcache *p, unsigned int iKey, int createFlag ){ + unsigned int nPinned; PCache1 *pCache = (PCache1 *)p; + PGroup *pGroup; PgHdr1 *pPage = 0; assert( offsetof(PgHdr1,page)==0 ); assert( pCache->bPurgeable || createFlag!=1 ); assert( pCache->bPurgeable || pCache->nMin==0 ); assert( pCache->bPurgeable==0 || pCache->nMin==10 ); assert( pCache->nMin==0 || pCache->bPurgeable ); - assert( pCache->nHash>0 ); - pcache1EnterMutex(pCache->pGroup); + pcache1EnterMutex(pGroup = pCache->pGroup); /* Step 1: Search the hash table for an existing entry. */ - pPage = pCache->apHash[iKey % pCache->nHash]; - while( pPage && pPage->iKey!=iKey ){ pPage = pPage->pNext; } + if( pCache->nHash>0 ){ + unsigned int h = iKey % pCache->nHash; + for(pPage=pCache->apHash[h]; pPage&&pPage->iKey!=iKey; pPage=pPage->pNext); + } /* Step 2: Abort if no existing page is found and createFlag is 0 */ if( pPage ){ if( !pPage->isPinned ) pcache1PinPage(pPage); - }else if( createFlag ){ - /* Steps 3, 4, and 5 implemented by this subroutine */ - pPage = pcache1FetchStage2(pCache, iKey, createFlag); + goto fetch_out; } - assert( pPage==0 || pCache->iMaxKey>=iKey ); - pcache1LeaveMutex(pCache->pGroup); + if( createFlag==0 ){ + goto fetch_out; + } + + /* The pGroup local variable will normally be initialized by the + ** pcache1EnterMutex() macro above. But if SQLITE_MUTEX_OMIT is defined, + ** then pcache1EnterMutex() is a no-op, so we have to initialize the + ** local variable here. Delaying the initialization of pGroup is an + ** optimization: The common case is to exit the module before reaching + ** this point. + */ +#ifdef SQLITE_MUTEX_OMIT + pGroup = pCache->pGroup; +#endif + + /* Step 3: Abort if createFlag is 1 but the cache is nearly full */ + assert( pCache->nPage >= pCache->nRecyclable ); + nPinned = pCache->nPage - pCache->nRecyclable; + assert( pGroup->mxPinned == pGroup->nMaxPage + 10 - pGroup->nMinPage ); + assert( pCache->n90pct == pCache->nMax*9/10 ); + if( createFlag==1 && ( + nPinned>=pGroup->mxPinned + || nPinned>=pCache->n90pct + || pcache1UnderMemoryPressure(pCache) + )){ + goto fetch_out; + } + + if( pCache->nPage>=pCache->nHash && pcache1ResizeHash(pCache) ){ + goto fetch_out; + } + assert( pCache->nHash>0 && pCache->apHash ); + + /* Step 4. Try to recycle a page. */ + if( pCache->bPurgeable && pGroup->pLruTail && ( + (pCache->nPage+1>=pCache->nMax) + || pGroup->nCurrentPage>=pGroup->nMaxPage + || pcache1UnderMemoryPressure(pCache) + )){ + PCache1 *pOther; + pPage = pGroup->pLruTail; + assert( pPage->isPinned==0 ); + pcache1RemoveFromHash(pPage); + pcache1PinPage(pPage); + pOther = pPage->pCache; + + /* We want to verify that szPage and szExtra are the same for pOther + ** and pCache. Assert that we can verify this by comparing sums. */ + assert( (pCache->szPage & (pCache->szPage-1))==0 && pCache->szPage>=512 ); + assert( pCache->szExtra<512 ); + assert( (pOther->szPage & (pOther->szPage-1))==0 && pOther->szPage>=512 ); + assert( pOther->szExtra<512 ); + + if( pOther->szPage+pOther->szExtra != pCache->szPage+pCache->szExtra ){ + pcache1FreePage(pPage); + pPage = 0; + }else{ + pGroup->nCurrentPage -= (pOther->bPurgeable - pCache->bPurgeable); + } + } + + /* Step 5. If a usable page buffer has still not been found, + ** attempt to allocate a new one. + */ + if( !pPage ){ + if( createFlag==1 ) sqlite3BeginBenignMalloc(); + pPage = pcache1AllocPage(pCache); + if( createFlag==1 ) sqlite3EndBenignMalloc(); + } + + if( pPage ){ + unsigned int h = iKey % pCache->nHash; + pCache->nPage++; + pPage->iKey = iKey; + pPage->pNext = pCache->apHash[h]; + pPage->pCache = pCache; + pPage->pLruPrev = 0; + pPage->pLruNext = 0; + pPage->isPinned = 1; + *(void **)pPage->page.pExtra = 0; + pCache->apHash[h] = pPage; + } + +fetch_out: + if( pPage && iKey>pCache->iMaxKey ){ + pCache->iMaxKey = iKey; + } + pcache1LeaveMutex(pGroup); return (sqlite3_pcache_page*)pPage; } /* @@ -40260,11 +39645,11 @@ ** a non-zero batch number, it will see all prior INSERTs. ** ** No INSERTs may occurs after a SMALLEST. An assertion will fail if ** that is attempted. ** -** The cost of an INSERT is roughly constant. (Sometimes new memory +** The cost of an INSERT is roughly constant. (Sometime new memory ** has to be allocated on an INSERT.) The cost of a TEST with a new ** batch number is O(NlogN) where N is the number of elements in the RowSet. ** The cost of a TEST using the same batch number is O(logN). The cost ** of the first SMALLEST is O(NlogN). Second and subsequent SMALLEST ** primitives are constant time. The cost of DESTROY is O(N). @@ -40652,12 +40037,12 @@ /* ** Check to see if element iRowid was inserted into the rowset as ** part of any insert batch prior to iBatch. Return 1 or 0. ** -** If this is the first test of a new batch and if there exist entries -** on pRowSet->pEntry, then sort those entries into the forest at +** If this is the first test of a new batch and if there exist entires +** on pRowSet->pEntry, then sort those entires into the forest at ** pRowSet->pForest so that they can be tested. */ SQLITE_PRIVATE int sqlite3RowSetTest(RowSet *pRowSet, int iBatch, sqlite3_int64 iRowid){ struct RowSetEntry *p, *pTree; @@ -40935,16 +40320,16 @@ ** are synced prior to the master journal being deleted. ** ** Definition: Two databases (or the same database at two points it time) ** are said to be "logically equivalent" if they give the same answer to ** all queries. Note in particular the content of freelist leaf -** pages can be changed arbitrarily without affecting the logical equivalence +** pages can be changed arbitarily without effecting the logical equivalence ** of the database. ** ** (7) At any time, if any subset, including the empty set and the total set, ** of the unsynced changes to a rollback journal are removed and the -** journal is rolled back, the resulting database file will be logically +** journal is rolled back, the resulting database file will be logical ** equivalent to the database file at the beginning of the transaction. ** ** (8) When a transaction is rolled back, the xTruncate method of the VFS ** is called to restore the database file to the same size it was at ** the beginning of the transaction. (In some VFSes, the xTruncate @@ -41237,11 +40622,11 @@ ** be a few redundant xLock() calls or a lock may be held for longer than ** required, but nothing really goes wrong. ** ** The exception is when the database file is unlocked as the pager moves ** from ERROR to OPEN state. At this point there may be a hot-journal file -** in the file-system that needs to be rolled back (as part of an OPEN->SHARED +** in the file-system that needs to be rolled back (as part of a OPEN->SHARED ** transition, by the same pager or any other). If the call to xUnlock() ** fails at this point and the pager is left holding an EXCLUSIVE lock, this ** can confuse the call to xCheckReservedLock() call made later as part ** of hot-journal detection. ** @@ -41320,11 +40705,11 @@ #define SPILLFLAG_OFF 0x01 /* Never spill cache. Set via pragma */ #define SPILLFLAG_ROLLBACK 0x02 /* Current rolling back, so do not spill */ #define SPILLFLAG_NOSYNC 0x04 /* Spill is ok, but do not sync */ /* -** An open page cache is an instance of struct Pager. A description of +** A open page cache is an instance of struct Pager. A description of ** some of the more important member variables follows: ** ** eState ** ** The current 'state' of the pager object. See the comment and state @@ -41492,11 +40877,11 @@ u8 readOnly; /* True for a read-only database */ u8 memDb; /* True to inhibit all file I/O */ /************************************************************************** ** The following block contains those class members that change during - ** routine operation. Class members not in this block are either fixed + ** routine opertion. Class members not in this block are either fixed ** when the pager is first created or else only change when there is a ** significant mode change (such as changing the page_size, locking_mode, ** or the journal_mode). From another view, these class members describe ** the "state" of the pager, while other class members describe the ** "configuration" of the pager. @@ -42535,10 +41920,25 @@ ){ rc = sqlite3OsTruncate(pPager->jfd, pPager->journalOff); } return rc; } + +/* +** Find a page in the hash table given its page number. Return +** a pointer to the page or NULL if the requested page is not +** already in memory. +*/ +static PgHdr *pager_lookup(Pager *pPager, Pgno pgno){ + PgHdr *p = 0; /* Return value */ + + /* It is not possible for a call to PcacheFetch() with createFlag==0 to + ** fail, since no attempt to allocate dynamic memory will be made. + */ + (void)sqlite3PcacheFetch(pPager->pPCache, pgno, 0, &p); + return p; +} /* ** Discard the entire contents of the in-memory page-cache. */ static void pager_reset(Pager *pPager){ @@ -42828,11 +42228,11 @@ } #ifdef SQLITE_CHECK_PAGES sqlite3PcacheIterateDirty(pPager->pPCache, pager_set_pagehash); if( pPager->dbSize==0 && sqlite3PcacheRefCount(pPager->pPCache)>0 ){ - PgHdr *p = sqlite3PagerLookup(pPager, 1); + PgHdr *p = pager_lookup(pPager, 1); if( p ){ p->pageHash = 0; sqlite3PagerUnrefNotNull(p); } } @@ -43107,11 +42507,11 @@ ** Do not attempt to write if database file has never been opened. */ if( pagerUseWal(pPager) ){ pPg = 0; }else{ - pPg = sqlite3PagerLookup(pPager, pgno); + pPg = pager_lookup(pPager, pgno); } assert( pPg || !MEMDB ); assert( pPager->eState!=PAGER_OPEN || pPg==0 ); PAGERTRACE(("PLAYBACK %d page %d hash(%08x) %s\n", PAGERID(pPager), pgno, pager_datahash(pPager->pageSize, (u8*)aData), @@ -43287,11 +42687,11 @@ ** journal files extracted from regular rollback-journals. */ rc = sqlite3OsFileSize(pMaster, &nMasterJournal); if( rc!=SQLITE_OK ) goto delmaster_out; nMasterPtr = pVfs->mxPathname+1; - zMasterJournal = sqlite3Malloc(nMasterJournal + nMasterPtr + 1); + zMasterJournal = sqlite3Malloc((int)nMasterJournal + nMasterPtr + 1); if( !zMasterJournal ){ rc = SQLITE_NOMEM; goto delmaster_out; } zMasterPtr = &zMasterJournal[nMasterJournal+1]; @@ -43356,11 +42756,11 @@ ** DBMOD or OPEN state, this function is a no-op. Otherwise, the size ** of the file is changed to nPage pages (nPage*pPager->pageSize bytes). ** If the file on disk is currently larger than nPage pages, then use the VFS ** xTruncate() method to truncate it. ** -** Or, it might be the case that the file on disk is smaller than +** Or, it might might be the case that the file on disk is smaller than ** nPage pages. Some operating system implementations can get confused if ** you try to truncate a file to some size that is larger than it ** currently is, so detect this case and write a single zero byte to ** the end of the new file instead. ** @@ -43415,11 +42815,11 @@ } /* ** Set the value of the Pager.sectorSize variable for the given ** pager based on the value returned by the xSectorSize method -** of the open database file. The sector size will be used +** of the open database file. The sector size will be used used ** to determine the size and alignment of journal header and ** master journal pointers within created journal files. ** ** For temporary files the effective sector size is always 512 bytes. ** @@ -44477,17 +43877,15 @@ if( !pNew ) rc = SQLITE_NOMEM; } if( rc==SQLITE_OK ){ pager_reset(pPager); + pPager->dbSize = (Pgno)((nByte+pageSize-1)/pageSize); + pPager->pageSize = pageSize; sqlite3PageFree(pPager->pTmpSpace); pPager->pTmpSpace = pNew; - rc = sqlite3PcacheSetPageSize(pPager->pPCache, pageSize); - } - if( rc==SQLITE_OK ){ - pPager->dbSize = (Pgno)((nByte+pageSize-1)/pageSize); - pPager->pageSize = pageSize; + sqlite3PcacheSetPageSize(pPager->pPCache, pageSize); } } *pPageSize = pPager->pageSize; if( rc==SQLITE_OK ){ @@ -44617,11 +44015,11 @@ */ static int pager_wait_on_lock(Pager *pPager, int locktype){ int rc; /* Return code */ /* Check that this is either a no-op (because the requested lock is - ** already held), or one of the transitions that the busy-handler + ** already held, or one of the transistions that the busy-handler ** may be invoked during, according to the comment above ** sqlite3PagerSetBusyhandler(). */ assert( (pPager->eLock>=locktype) || (pPager->eLock==NO_LOCK && locktype==SHARED_LOCK) @@ -45245,12 +44643,12 @@ ** The doNotSpill ROLLBACK and OFF bits inhibits all cache spilling ** regardless of whether or not a sync is required. This is set during ** a rollback or by user request, respectively. ** ** Spilling is also prohibited when in an error state since that could - ** lead to database corruption. In the current implementation it - ** is impossible for sqlite3PcacheFetch() to be called with createFlag==3 + ** lead to database corruption. In the current implementaton it + ** is impossible for sqlite3PcacheFetch() to be called with createFlag==1 ** while in the error state, hence it is impossible for this routine to ** be called in the error state. Nevertheless, we include a NEVER() ** test for the error state as a safeguard against future changes. */ if( NEVER(pPager->errCode) ) return SQLITE_OK; @@ -45582,27 +44980,26 @@ assert( pPager->memDb==0 ); rc = sqlite3PagerSetPagesize(pPager, &szPageDflt, -1); testcase( rc!=SQLITE_OK ); } - /* Initialize the PCache object. */ - if( rc==SQLITE_OK ){ - assert( nExtra<1000 ); - nExtra = ROUND8(nExtra); - rc = sqlite3PcacheOpen(szPageDflt, nExtra, !memDb, - !memDb?pagerStress:0, (void *)pPager, pPager->pPCache); - } - - /* If an error occurred above, free the Pager structure and close the file. + /* If an error occurred in either of the blocks above, free the + ** Pager structure and close the file. */ if( rc!=SQLITE_OK ){ + assert( !pPager->pTmpSpace ); sqlite3OsClose(pPager->fd); - sqlite3PageFree(pPager->pTmpSpace); sqlite3_free(pPager); return rc; } + /* Initialize the PCache object. */ + assert( nExtra<1000 ); + nExtra = ROUND8(nExtra); + sqlite3PcacheOpen(szPageDflt, nExtra, !memDb, + !memDb?pagerStress:0, (void *)pPager, pPager->pPCache); + PAGERTRACE(("OPEN %d %s\n", FILEHANDLEID(pPager->fd), pPager->zFilename)); IOTRACE(("OPEN %p %s\n", pPager, pPager->zFilename)) pPager->useJournal = (u8)useJournal; /* pPager->stmtOpen = 0; */ @@ -45785,11 +45182,11 @@ sqlite3OsClose(pPager->jfd); } *pExists = (first!=0); }else if( rc==SQLITE_CANTOPEN ){ /* If we cannot open the rollback journal file in order to see if - ** it has a zero header, that might be due to an I/O error, or + ** its has a zero header, that might be due to an I/O error, or ** it might be due to the race condition described above and in ** ticket #3883. Either way, assume that the journal is hot. ** This might be a false positive. But if it is, then the ** automatic journal playback and recovery mechanism will deal ** with it under an EXCLUSIVE lock where we do not need to @@ -46147,10 +45544,11 @@ /* If the pager is in the error state, return an error immediately. ** Otherwise, request the page from the PCache layer. */ if( pPager->errCode!=SQLITE_OK ){ rc = pPager->errCode; }else{ + if( bMmapOk && pagerUseWal(pPager) ){ rc = sqlite3WalFindFrame(pPager->pWal, pgno, &iFrame); if( rc!=SQLITE_OK ) goto pager_acquire_err; } @@ -46161,11 +45559,11 @@ (i64)(pgno-1) * pPager->pageSize, pPager->pageSize, &pData ); if( rc==SQLITE_OK && pData ){ if( pPager->eState>PAGER_READER ){ - pPg = sqlite3PagerLookup(pPager, pgno); + (void)sqlite3PcacheFetch(pPager->pPCache, pgno, 0, &pPg); } if( pPg==0 ){ rc = pagerAcquireMapPage(pPager, pgno, pData, &pPg); }else{ sqlite3OsUnfetch(pPager->fd, (i64)(pgno-1)*pPager->pageSize, pData); @@ -46179,20 +45577,11 @@ if( rc!=SQLITE_OK ){ goto pager_acquire_err; } } - { - sqlite3_pcache_page *pBase; - pBase = sqlite3PcacheFetch(pPager->pPCache, pgno, 3); - if( pBase==0 ){ - rc = sqlite3PcacheFetchStress(pPager->pPCache, pgno, &pBase); - if( rc!=SQLITE_OK ) goto pager_acquire_err; - } - pPg = *ppPage = sqlite3PcacheFetchFinish(pPager->pPCache, pgno, pBase); - if( pPg==0 ) rc = SQLITE_NOMEM; - } + rc = sqlite3PcacheFetch(pPager->pPCache, pgno, 1, ppPage); } if( rc!=SQLITE_OK ){ /* Either the call to sqlite3PcacheFetch() returned an error or the ** pager was already in the error-state when this function was called. @@ -46285,16 +45674,17 @@ ** in the page if the page is not already in cache. This routine ** returns NULL if the page is not in cache or if a disk I/O error ** has ever happened. */ SQLITE_PRIVATE DbPage *sqlite3PagerLookup(Pager *pPager, Pgno pgno){ - sqlite3_pcache_page *pPage; + PgHdr *pPg = 0; assert( pPager!=0 ); assert( pgno!=0 ); assert( pPager->pPCache!=0 ); - pPage = sqlite3PcacheFetch(pPager->pPCache, pgno, 0); - return sqlite3PcacheFetchFinish(pPager->pPCache, pgno, pPage); + assert( pPager->eState>=PAGER_READER && pPager->eState!=PAGER_ERROR ); + sqlite3PcacheFetch(pPager->pPCache, pgno, 0, &pPg); + return pPg; } /* ** Release a page reference. ** @@ -46626,101 +46016,10 @@ pPager->dbSize = pPg->pgno; } return rc; } -/* -** This is a variant of sqlite3PagerWrite() that runs when the sector size -** is larger than the page size. SQLite makes the (reasonable) assumption that -** all bytes of a sector are written together by hardware. Hence, all bytes of -** a sector need to be journalled in case of a power loss in the middle of -** a write. -** -** Usually, the sector size is less than or equal to the page size, in which -** case pages can be individually written. This routine only runs in the exceptional -** case where the page size is smaller than the sector size. -*/ -static SQLITE_NOINLINE int pagerWriteLargeSector(PgHdr *pPg){ - int rc = SQLITE_OK; /* Return code */ - Pgno nPageCount; /* Total number of pages in database file */ - Pgno pg1; /* First page of the sector pPg is located on. */ - int nPage = 0; /* Number of pages starting at pg1 to journal */ - int ii; /* Loop counter */ - int needSync = 0; /* True if any page has PGHDR_NEED_SYNC */ - Pager *pPager = pPg->pPager; /* The pager that owns pPg */ - Pgno nPagePerSector = (pPager->sectorSize/pPager->pageSize); - - /* Set the doNotSpill NOSYNC bit to 1. This is because we cannot allow - ** a journal header to be written between the pages journaled by - ** this function. - */ - assert( !MEMDB ); - assert( (pPager->doNotSpill & SPILLFLAG_NOSYNC)==0 ); - pPager->doNotSpill |= SPILLFLAG_NOSYNC; - - /* This trick assumes that both the page-size and sector-size are - ** an integer power of 2. It sets variable pg1 to the identifier - ** of the first page of the sector pPg is located on. - */ - pg1 = ((pPg->pgno-1) & ~(nPagePerSector-1)) + 1; - - nPageCount = pPager->dbSize; - if( pPg->pgno>nPageCount ){ - nPage = (pPg->pgno - pg1)+1; - }else if( (pg1+nPagePerSector-1)>nPageCount ){ - nPage = nPageCount+1-pg1; - }else{ - nPage = nPagePerSector; - } - assert(nPage>0); - assert(pg1<=pPg->pgno); - assert((pg1+nPage)>pPg->pgno); - - for(ii=0; ii pgno || !sqlite3BitvecTest(pPager->pInJournal, pg) ){ - if( pg!=PAGER_MJ_PGNO(pPager) ){ - rc = sqlite3PagerGet(pPager, pg, &pPage); - if( rc==SQLITE_OK ){ - rc = pager_write(pPage); - if( pPage->flags&PGHDR_NEED_SYNC ){ - needSync = 1; - } - sqlite3PagerUnrefNotNull(pPage); - } - } - }else if( (pPage = sqlite3PagerLookup(pPager, pg))!=0 ){ - if( pPage->flags&PGHDR_NEED_SYNC ){ - needSync = 1; - } - sqlite3PagerUnrefNotNull(pPage); - } - } - - /* If the PGHDR_NEED_SYNC flag is set for any of the nPage pages - ** starting at pg1, then it needs to be set for all of them. Because - ** writing to any of these nPage pages may damage the others, the - ** journal file must contain sync()ed copies of all of them - ** before any of them can be written out to the database file. - */ - if( rc==SQLITE_OK && needSync ){ - assert( !MEMDB ); - for(ii=0; ii pgno || !sqlite3BitvecTest(pPager->pInJournal, pg) ){ + if( pg!=PAGER_MJ_PGNO(pPager) ){ + rc = sqlite3PagerGet(pPager, pg, &pPage); + if( rc==SQLITE_OK ){ + rc = pager_write(pPage); + if( pPage->flags&PGHDR_NEED_SYNC ){ + needSync = 1; + } + sqlite3PagerUnrefNotNull(pPage); + } + } + }else if( (pPage = pager_lookup(pPager, pg))!=0 ){ + if( pPage->flags&PGHDR_NEED_SYNC ){ + needSync = 1; + } + sqlite3PagerUnrefNotNull(pPage); + } + } + + /* If the PGHDR_NEED_SYNC flag is set for any of the nPage pages + ** starting at pg1, then it needs to be set for all of them. Because + ** writing to any of these nPage pages may damage the others, the + ** journal file must contain sync()ed copies of all of them + ** before any of them can be written out to the database file. + */ + if( rc==SQLITE_OK && needSync ){ + assert( !MEMDB ); + for(ii=0; ii childPtrSize==4 ); - return (u16)(pIter - pCell); - } - nSize = *pIter; - if( nSize>=0x80 ){ - pEnd = &pIter[9]; - nSize &= 0x7f; - do{ - nSize = (nSize<<7) | (*++pIter & 0x7f); - }while( *(pIter)>=0x80 && pIter intKey ){ + u8 *pEnd; + if( pPage->hasData ){ + pIter += getVarint32(pIter, nSize); + }else{ + nSize = 0; + } + /* pIter now points at the 64-bit integer key value, a variable length ** integer. The following block moves pIter to point at the first byte ** past the end of the key value. */ pEnd = &pIter[9]; while( (*pIter++)&0x80 && pIter maxLocal ); testcase( nSize==pPage->maxLocal+1 ); - if( nSize<=pPage->maxLocal ){ - nSize += (u32)(pIter - pCell); - if( nSize<4 ) nSize = 4; - }else{ + if( nSize>pPage->maxLocal ){ int minLocal = pPage->minLocal; nSize = minLocal + (nSize - minLocal) % (pPage->pBt->usableSize - 4); testcase( nSize==pPage->maxLocal ); testcase( nSize==pPage->maxLocal+1 ); if( nSize>pPage->maxLocal ){ nSize = minLocal; } - nSize += 4 + (u16)(pIter - pCell); + nSize += 4; } - assert( nSize==debuginfo.nSize || CORRUPT_DB ); + nSize += (u32)(pIter - pCell); + + /* The minimum size of any cell is 4 bytes. */ + if( nSize<4 ){ + nSize = 4; + } + + assert( nSize==debuginfo.nSize ); return (u16)nSize; } #ifdef SQLITE_DEBUG /* This variation on cellSizePtr() is used inside of assert() statements @@ -53298,10 +52626,11 @@ static void ptrmapPutOvflPtr(MemPage *pPage, u8 *pCell, int *pRC){ CellInfo info; if( *pRC ) return; assert( pCell!=0 ); btreeParseCellPtr(pPage, pCell, &info); + assert( (info.nData+(pPage->intKey?0:info.nKey))==info.nPayload ); if( info.iOverflow ){ Pgno ovfl = get4byte(&pCell[info.iOverflow]); ptrmapPut(pPage->pBt, ovfl, PTRMAP_OVERFLOW1, pPage->pgno, pRC); } } @@ -53314,11 +52643,11 @@ ** big FreeBlk that occurs in between the header and cell ** pointer array and the cell content area. */ static int defragmentPage(MemPage *pPage){ int i; /* Loop counter */ - int pc; /* Address of the i-th cell */ + int pc; /* Address of a i-th cell */ int hdr; /* Offset to the page header */ int size; /* Size of a cell */ int usableSize; /* Number of usable bytes on a page */ int cellOffset; /* Offset to the cell pointer array */ int cbrk; /* Offset to the cell content area */ @@ -53405,10 +52734,11 @@ ** also end up needing a new cell pointer. */ static int allocateSpace(MemPage *pPage, int nByte, int *pIdx){ const int hdr = pPage->hdrOffset; /* Local cache of pPage->hdrOffset */ u8 * const data = pPage->aData; /* Local cache of pPage->aData */ + int nFrag; /* Number of fragmented bytes on pPage */ int top; /* First byte of cell content area */ int gap; /* First byte of gap between cell pointers and cell content */ int rc; /* Integer return code */ int usableSize; /* Usable size of the page */ @@ -53419,30 +52749,29 @@ assert( pPage->nFree>=nByte ); assert( pPage->nOverflow==0 ); usableSize = pPage->pBt->usableSize; assert( nByte < usableSize-8 ); + nFrag = data[hdr+7]; assert( pPage->cellOffset == hdr + 12 - 4*pPage->leaf ); gap = pPage->cellOffset + 2*pPage->nCell; - assert( gap<=65536 ); - top = get2byte(&data[hdr+5]); - if( gap>top ){ - if( top==0 ){ - top = 65536; - }else{ - return SQLITE_CORRUPT_BKPT; - } - } - - /* If there is enough space between gap and top for one more cell pointer - ** array entry offset, and if the freelist is not empty, then search the - ** freelist looking for a free slot big enough to satisfy the request. - */ + top = get2byteNotZero(&data[hdr+5]); + if( gap>top ) return SQLITE_CORRUPT_BKPT; testcase( gap+2==top ); testcase( gap+1==top ); testcase( gap==top ); - if( gap+2<=top && (data[hdr+1] || data[hdr+2]) ){ + + if( nFrag>=60 ){ + /* Always defragment highly fragmented pages */ + rc = defragmentPage(pPage); + if( rc ) return rc; + top = get2byteNotZero(&data[hdr+5]); + }else if( gap+2<=top ){ + /* Search the freelist looking for a free slot big enough to satisfy + ** the request. The allocation is made from the first free slot in + ** the list that is large enough to accommodate it. + */ int pc, addr; for(addr=hdr+1; (pc = get2byte(&data[addr]))>0; addr=pc){ int size; /* Size of the free slot */ if( pc>usableSize-4 || pc iLast ) return SQLITE_CORRUPT_BKPT; - assert( iFreeBlk>iPtr || iFreeBlk==0 ); - - /* At this point: - ** iFreeBlk: First freeblock after iStart, or zero if none - ** iPtr: The address of a pointer iFreeBlk - ** - ** Check to see if iFreeBlk should be coalesced onto the end of iStart. - */ - if( iFreeBlk && iEnd+3>=iFreeBlk ){ - nFrag = iFreeBlk - iEnd; - if( iEnd>iFreeBlk ) return SQLITE_CORRUPT_BKPT; - iEnd = iFreeBlk + get2byte(&data[iFreeBlk+2]); - iSize = iEnd - iStart; - iFreeBlk = get2byte(&data[iFreeBlk]); - } - - /* If iPtr is another freeblock (that is, if iPtr is not the freelist - ** pointer in the page header) then check to see if iStart should be - ** coalesced onto the end of iPtr. - */ - if( iPtr>hdr+1 ){ - int iPtrEnd = iPtr + get2byte(&data[iPtr+2]); - if( iPtrEnd+3>=iStart ){ - if( iPtrEnd>iStart ) return SQLITE_CORRUPT_BKPT; - nFrag += iStart - iPtrEnd; - iSize = iEnd - iPtr; - iStart = iPtr; - } - } - if( nFrag>data[hdr+7] ) return SQLITE_CORRUPT_BKPT; - data[hdr+7] -= nFrag; - } - if( iStart==get2byte(&data[hdr+5]) ){ - /* The new freeblock is at the beginning of the cell content area, - ** so just extend the cell content area rather than create another - ** freelist entry */ - if( iPtr!=hdr+1 ) return SQLITE_CORRUPT_BKPT; - put2byte(&data[hdr+1], iFreeBlk); - put2byte(&data[hdr+5], iEnd); - }else{ - /* Insert the new freeblock into the freelist */ - put2byte(&data[iPtr], iStart); - put2byte(&data[iStart], iFreeBlk); - put2byte(&data[iStart+2], iSize); - } - pPage->nFree += iOrigSize; + addr = hdr + 1; + iLast = pPage->pBt->usableSize - 4; + assert( start<=iLast ); + while( (pbegin = get2byte(&data[addr])) 0 ){ + if( pbegin nCell ); assert( cursorHoldsMutex(pCur) ); - assert( eOp!=2 || offset==0 ); /* Always start from beginning for eOp==2 */ + assert( eOp!=2 || offset==0 ); /* Always start from beginning for eOp==2 */ getCellInfo(pCur); - aPayload = pCur->info.pPayload; + aPayload = pCur->info.pCell + pCur->info.nHeader; + nKey = (pPage->intKey ? 0 : (int)pCur->info.nKey); #ifdef SQLITE_DIRECT_OVERFLOW_READ - bEnd = offset+amt==pCur->info.nPayload; + bEnd = (offset+amt==nKey+pCur->info.nData); #endif - assert( offset+amt <= pCur->info.nPayload ); - if( &aPayload[pCur->info.nLocal] > &pPage->aData[pBt->usableSize] ){ + if( NEVER(offset+amt > nKey+pCur->info.nData) + || &aPayload[pCur->info.nLocal] > &pPage->aData[pBt->usableSize] + ){ /* Trying to read or write past the end of the data is an error */ return SQLITE_CORRUPT_BKPT; } /* Check if data must be read/written to/from the btree page itself. */ @@ -56267,13 +55578,11 @@ /* If the overflow page-list cache has been allocated and the ** entry for the first required overflow page is valid, skip ** directly to it. */ - if( (pCur->curFlags & BTCF_ValidOvfl)!=0 - && pCur->aOverflow[offset/ovflSize] - ){ + if( (pCur->curFlags & BTCF_ValidOvfl)!=0 && pCur->aOverflow[offset/ovflSize] ){ iIdx = (offset/ovflSize); nextPage = pCur->aOverflow[iIdx]; offset = (offset%ovflSize); } @@ -56322,11 +55631,10 @@ ** 2) data is required from the start of this overflow page, and ** 3) the database is file-backed, and ** 4) there is no open write-transaction, and ** 5) the database is not a WAL database, ** 6) all data from the page is being read. - ** 7) at least 4 bytes have already been read into the output buffer ** ** then data can be read directly from the database file into the ** output buffer, bypassing the page-cache altogether. This speeds ** up loading large records that span many overflow pages. */ @@ -56334,15 +55642,13 @@ && offset==0 /* (2) */ && (bEnd || a==ovflSize) /* (6) */ && pBt->inTransaction==TRANS_READ /* (4) */ && (fd = sqlite3PagerFile(pBt->pPager))->pMethods /* (3) */ && pBt->pPage1->aData[19]==0x01 /* (5) */ - && &pBuf[-4]>=pBufStart /* (7) */ ){ u8 aSave[4]; u8 *aWrite = &pBuf[-4]; - assert( aWrite>=pBufStart ); /* hence (7) */ memcpy(aSave, aWrite, 4); rc = sqlite3OsRead(fd, aWrite, a+4, (i64)pBt->pageSize*(nextPage-1)); nextPage = get4byte(aWrite); memcpy(aWrite, aSave, 4); }else @@ -56373,11 +55679,11 @@ return rc; } /* ** Read part of the key associated with cursor pCur. Exactly -** "amt" bytes will be transferred into pBuf[]. The transfer +** "amt" bytes will be transfered into pBuf[]. The transfer ** begins at "offset". ** ** The caller must ensure that pCur is pointing to a valid row ** in the table. ** @@ -56450,11 +55756,11 @@ assert( sqlite3_mutex_held(pCur->pBtree->db->mutex) ); assert( cursorHoldsMutex(pCur) ); assert( pCur->aiIdx[pCur->iPage] apPage[pCur->iPage]->nCell ); assert( pCur->info.nSize>0 ); *pAmt = pCur->info.nLocal; - return (void*)pCur->info.pPayload; + return (void*)(pCur->info.pCell + pCur->info.nHeader); } /* ** For the entry that cursor pCur is point to, return as @@ -56693,20 +55999,21 @@ int rc = SQLITE_OK; MemPage *pPage = 0; assert( cursorHoldsMutex(pCur) ); assert( pCur->eState==CURSOR_VALID ); - while( !(pPage = pCur->apPage[pCur->iPage])->leaf ){ + while( rc==SQLITE_OK && !(pPage = pCur->apPage[pCur->iPage])->leaf ){ pgno = get4byte(&pPage->aData[pPage->hdrOffset+8]); pCur->aiIdx[pCur->iPage] = pPage->nCell; rc = moveToChild(pCur, pgno); - if( rc ) return rc; } - pCur->aiIdx[pCur->iPage] = pPage->nCell-1; - assert( pCur->info.nSize==0 ); - assert( (pCur->curFlags & BTCF_ValidNKey)==0 ); - return SQLITE_OK; + if( rc==SQLITE_OK ){ + pCur->aiIdx[pCur->iPage] = pPage->nCell-1; + pCur->info.nSize = 0; + pCur->curFlags &= ~BTCF_ValidNKey; + } + return rc; } /* Move the cursor to the first entry in the table. Return SQLITE_OK ** on success. Set *pRes to 0 if the cursor actually points to something ** or set *pRes to 1 if the table is empty. @@ -56833,11 +56140,11 @@ } } if( pIdxKey ){ xRecordCompare = sqlite3VdbeFindCompare(pIdxKey); - pIdxKey->errCode = 0; + pIdxKey->isCorrupt = 0; assert( pIdxKey->default_rc==1 || pIdxKey->default_rc==0 || pIdxKey->default_rc==-1 ); }else{ @@ -56878,11 +56185,11 @@ pCur->aiIdx[pCur->iPage] = (u16)idx; if( xRecordCompare==0 ){ for(;;){ i64 nCellKey; pCell = findCell(pPage, idx) + pPage->childPtrSize; - if( pPage->intKeyLeaf ){ + if( pPage->hasData ){ while( 0x80 <= *(pCell++) ){ if( pCell>=pPage->aDataEnd ) return SQLITE_CORRUPT_BKPT; } } getVarint(pCell, (u64*)&nCellKey); @@ -56926,18 +56233,18 @@ if( nCell<=pPage->max1bytePayload ){ /* This branch runs if the record-size field of the cell is a ** single byte varint and the record fits entirely on the main ** b-tree page. */ testcase( pCell+nCell+1==pPage->aDataEnd ); - c = xRecordCompare(nCell, (void*)&pCell[1], pIdxKey); + c = xRecordCompare(nCell, (void*)&pCell[1], pIdxKey, 0); }else if( !(pCell[1] & 0x80) && (nCell = ((nCell&0x7f)<<7) + pCell[1])<=pPage->maxLocal ){ /* The record-size field is a 2 byte varint and the record ** fits entirely on the main b-tree page. */ testcase( pCell+nCell+2==pPage->aDataEnd ); - c = xRecordCompare(nCell, (void*)&pCell[2], pIdxKey); + c = xRecordCompare(nCell, (void*)&pCell[2], pIdxKey, 0); }else{ /* The record flows over onto one or more overflow pages. In ** this case the whole cell needs to be parsed, a buffer allocated ** and accessPayload() used to retrieve the record into the ** buffer before VdbeRecordCompare() can be called. */ @@ -56954,27 +56261,24 @@ rc = accessPayload(pCur, 0, nCell, (unsigned char*)pCellKey, 2); if( rc ){ sqlite3_free(pCellKey); goto moveto_finish; } - c = xRecordCompare(nCell, pCellKey, pIdxKey); + c = xRecordCompare(nCell, pCellKey, pIdxKey, 0); sqlite3_free(pCellKey); } - assert( - (pIdxKey->errCode!=SQLITE_CORRUPT || c==0) - && (pIdxKey->errCode!=SQLITE_NOMEM || pCur->pBtree->db->mallocFailed) - ); + assert( pIdxKey->isCorrupt==0 || c==0 ); if( c<0 ){ lwr = idx+1; }else if( c>0 ){ upr = idx-1; }else{ assert( c==0 ); *pRes = 0; rc = SQLITE_OK; pCur->aiIdx[pCur->iPage] = (u16)idx; - if( pIdxKey->errCode ) rc = SQLITE_CORRUPT; + if( pIdxKey->isCorrupt ) rc = SQLITE_CORRUPT; goto moveto_finish; } if( lwr>upr ) break; assert( lwr+upr>=0 ); idx = (lwr+upr)>>1; /* idx = (lwr+upr)/2 */ @@ -57025,37 +56329,33 @@ ** Advance the cursor to the next entry in the database. If ** successful then set *pRes=0. If the cursor ** was already pointing to the last entry in the database before ** this routine was called, then set *pRes=1. ** -** The main entry point is sqlite3BtreeNext(). That routine is optimized -** for the common case of merely incrementing the cell counter BtCursor.aiIdx -** to the next cell on the current page. The (slower) btreeNext() helper -** routine is called when it is necessary to move to a different page or -** to restore the cursor. -** ** The calling function will set *pRes to 0 or 1. The initial *pRes value ** will be 1 if the cursor being stepped corresponds to an SQL index and ** if this routine could have been skipped if that SQL index had been ** a unique index. Otherwise the caller will have set *pRes to zero. ** Zero is the common case. The btree implementation is free to use the ** initial *pRes value as a hint to improve performance, but the current ** SQLite btree implementation does not. (Note that the comdb2 btree ** implementation does use this hint, however.) */ -static SQLITE_NOINLINE int btreeNext(BtCursor *pCur, int *pRes){ +SQLITE_PRIVATE int sqlite3BtreeNext(BtCursor *pCur, int *pRes){ int rc; int idx; MemPage *pPage; assert( cursorHoldsMutex(pCur) ); + assert( pRes!=0 ); + assert( *pRes==0 || *pRes==1 ); assert( pCur->skipNext==0 || pCur->eState!=CURSOR_VALID ); - assert( *pRes==0 ); if( pCur->eState!=CURSOR_VALID ){ - assert( (pCur->curFlags & BTCF_ValidOvfl)==0 ); + invalidateOverflowCache(pCur); rc = restoreCursorPosition(pCur); if( rc!=SQLITE_OK ){ + *pRes = 0; return rc; } if( CURSOR_INVALID==pCur->eState ){ *pRes = 1; return SQLITE_OK; @@ -57063,10 +56363,11 @@ if( pCur->skipNext ){ assert( pCur->eState==CURSOR_VALID || pCur->eState==CURSOR_SKIPNEXT ); pCur->eState = CURSOR_VALID; if( pCur->skipNext>0 ){ pCur->skipNext = 0; + *pRes = 0; return SQLITE_OK; } pCur->skipNext = 0; } } @@ -57080,15 +56381,22 @@ ** the page while cursor pCur is holding a reference to it. Which can ** only happen if the database is corrupt in such a way as to link the ** page into more than one b-tree structure. */ testcase( idx>pPage->nCell ); + pCur->info.nSize = 0; + pCur->curFlags &= ~(BTCF_ValidNKey|BTCF_ValidOvfl); if( idx>=pPage->nCell ){ if( !pPage->leaf ){ rc = moveToChild(pCur, get4byte(&pPage->aData[pPage->hdrOffset+8])); - if( rc ) return rc; - return moveToLeftmost(pCur); + if( rc ){ + *pRes = 0; + return rc; + } + rc = moveToLeftmost(pCur); + *pRes = 0; + return rc; } do{ if( pCur->iPage==0 ){ *pRes = 1; pCur->eState = CURSOR_INVALID; @@ -57095,79 +56403,58 @@ return SQLITE_OK; } moveToParent(pCur); pPage = pCur->apPage[pCur->iPage]; }while( pCur->aiIdx[pCur->iPage]>=pPage->nCell ); - if( pPage->intKey ){ - return sqlite3BtreeNext(pCur, pRes); - }else{ - return SQLITE_OK; - } - } - if( pPage->leaf ){ - return SQLITE_OK; - }else{ - return moveToLeftmost(pCur); - } -} -SQLITE_PRIVATE int sqlite3BtreeNext(BtCursor *pCur, int *pRes){ - MemPage *pPage; - assert( cursorHoldsMutex(pCur) ); - assert( pRes!=0 ); - assert( *pRes==0 || *pRes==1 ); - assert( pCur->skipNext==0 || pCur->eState!=CURSOR_VALID ); - pCur->info.nSize = 0; - pCur->curFlags &= ~(BTCF_ValidNKey|BTCF_ValidOvfl); - *pRes = 0; - if( pCur->eState!=CURSOR_VALID ) return btreeNext(pCur, pRes); - pPage = pCur->apPage[pCur->iPage]; - if( (++pCur->aiIdx[pCur->iPage])>=pPage->nCell ){ - pCur->aiIdx[pCur->iPage]--; - return btreeNext(pCur, pRes); - } - if( pPage->leaf ){ - return SQLITE_OK; - }else{ - return moveToLeftmost(pCur); - } -} + *pRes = 0; + if( pPage->intKey ){ + rc = sqlite3BtreeNext(pCur, pRes); + }else{ + rc = SQLITE_OK; + } + return rc; + } + *pRes = 0; + if( pPage->leaf ){ + return SQLITE_OK; + } + rc = moveToLeftmost(pCur); + return rc; +} + /* ** Step the cursor to the back to the previous entry in the database. If ** successful then set *pRes=0. If the cursor ** was already pointing to the first entry in the database before ** this routine was called, then set *pRes=1. ** -** The main entry point is sqlite3BtreePrevious(). That routine is optimized -** for the common case of merely decrementing the cell counter BtCursor.aiIdx -** to the previous cell on the current page. The (slower) btreePrevious() -** helper routine is called when it is necessary to move to a different page -** or to restore the cursor. -** ** The calling function will set *pRes to 0 or 1. The initial *pRes value ** will be 1 if the cursor being stepped corresponds to an SQL index and ** if this routine could have been skipped if that SQL index had been ** a unique index. Otherwise the caller will have set *pRes to zero. ** Zero is the common case. The btree implementation is free to use the ** initial *pRes value as a hint to improve performance, but the current ** SQLite btree implementation does not. (Note that the comdb2 btree ** implementation does use this hint, however.) */ -static SQLITE_NOINLINE int btreePrevious(BtCursor *pCur, int *pRes){ +SQLITE_PRIVATE int sqlite3BtreePrevious(BtCursor *pCur, int *pRes){ int rc; MemPage *pPage; assert( cursorHoldsMutex(pCur) ); assert( pRes!=0 ); - assert( *pRes==0 ); + assert( *pRes==0 || *pRes==1 ); assert( pCur->skipNext==0 || pCur->eState!=CURSOR_VALID ); - assert( (pCur->curFlags & (BTCF_AtLast|BTCF_ValidOvfl|BTCF_ValidNKey))==0 ); - assert( pCur->info.nSize==0 ); + pCur->curFlags &= ~(BTCF_AtLast|BTCF_ValidOvfl); if( pCur->eState!=CURSOR_VALID ){ - rc = restoreCursorPosition(pCur); - if( rc!=SQLITE_OK ){ - return rc; + if( ALWAYS(pCur->eState>=CURSOR_REQUIRESEEK) ){ + rc = btreeRestoreCursorPosition(pCur); + if( rc!=SQLITE_OK ){ + *pRes = 0; + return rc; + } } if( CURSOR_INVALID==pCur->eState ){ *pRes = 1; return SQLITE_OK; } @@ -57174,10 +56461,11 @@ if( pCur->skipNext ){ assert( pCur->eState==CURSOR_VALID || pCur->eState==CURSOR_SKIPNEXT ); pCur->eState = CURSOR_VALID; if( pCur->skipNext<0 ){ pCur->skipNext = 0; + *pRes = 0; return SQLITE_OK; } pCur->skipNext = 0; } } @@ -57185,11 +56473,14 @@ pPage = pCur->apPage[pCur->iPage]; assert( pPage->isInit ); if( !pPage->leaf ){ int idx = pCur->aiIdx[pCur->iPage]; rc = moveToChild(pCur, get4byte(findCell(pPage, idx))); - if( rc ) return rc; + if( rc ){ + *pRes = 0; + return rc; + } rc = moveToRightmost(pCur); }else{ while( pCur->aiIdx[pCur->iPage]==0 ){ if( pCur->iPage==0 ){ pCur->eState = CURSOR_INVALID; @@ -57196,39 +56487,23 @@ *pRes = 1; return SQLITE_OK; } moveToParent(pCur); } - assert( pCur->info.nSize==0 ); - assert( (pCur->curFlags & (BTCF_ValidNKey|BTCF_ValidOvfl))==0 ); + pCur->info.nSize = 0; + pCur->curFlags &= ~(BTCF_ValidNKey|BTCF_ValidOvfl); pCur->aiIdx[pCur->iPage]--; pPage = pCur->apPage[pCur->iPage]; if( pPage->intKey && !pPage->leaf ){ rc = sqlite3BtreePrevious(pCur, pRes); }else{ rc = SQLITE_OK; } } - return rc; -} -SQLITE_PRIVATE int sqlite3BtreePrevious(BtCursor *pCur, int *pRes){ - assert( cursorHoldsMutex(pCur) ); - assert( pRes!=0 ); - assert( *pRes==0 || *pRes==1 ); - assert( pCur->skipNext==0 || pCur->eState!=CURSOR_VALID ); *pRes = 0; - pCur->curFlags &= ~(BTCF_AtLast|BTCF_ValidOvfl|BTCF_ValidNKey); - pCur->info.nSize = 0; - if( pCur->eState!=CURSOR_VALID - || pCur->aiIdx[pCur->iPage]==0 - || pCur->apPage[pCur->iPage]->leaf==0 - ){ - return btreePrevious(pCur, pRes); - } - pCur->aiIdx[pCur->iPage]--; - return SQLITE_OK; + return rc; } /* ** Allocate a new page from the database file. ** @@ -57466,11 +56741,11 @@ if( rc ) goto end_allocate_page; if( closest pDbPage); if( rc!=SQLITE_OK ){ releasePage(*ppPage); @@ -57499,11 +56774,11 @@ ** content for any page that really does lie past the end of the database ** file on disk. So the effects of disabling the no-content optimization ** here are confined to those pages that lie between the end of the ** database image and the end of the database file. */ - int bNoContent = (0==IfNotOmitAV(pBt->bDoTruncate))? PAGER_GET_NOCONTENT:0; + int bNoContent = (0==IfNotOmitAV(pBt->bDoTruncate)) ? PAGER_GET_NOCONTENT : 0; rc = sqlite3PagerWrite(pBt->pPage1->pDbPage); if( rc ) return rc; pBt->nPage++; if( pBt->nPage==PENDING_BYTE_PAGE(pBt) ) pBt->nPage++; @@ -57698,29 +56973,22 @@ *pRC = freePage2(pPage->pBt, pPage, pPage->pgno); } } /* -** Free any overflow pages associated with the given Cell. Write the -** local Cell size (the number of bytes on the original page, omitting -** overflow) into *pnSize. +** Free any overflow pages associated with the given Cell. */ -static int clearCell( - MemPage *pPage, /* The page that contains the Cell */ - unsigned char *pCell, /* First byte of the Cell */ - u16 *pnSize /* Write the size of the Cell here */ -){ +static int clearCell(MemPage *pPage, unsigned char *pCell){ BtShared *pBt = pPage->pBt; CellInfo info; Pgno ovflPgno; int rc; int nOvfl; u32 ovflPageSize; assert( sqlite3_mutex_held(pPage->pBt->mutex) ); btreeParseCellPtr(pPage, pCell, &info); - *pnSize = info.nSize; if( info.iOverflow==0 ){ return SQLITE_OK; /* No overflow pages. Return without doing anything */ } if( pCell+info.iOverflow+3 > pPage->aData+pPage->maskPage ){ return SQLITE_CORRUPT_BKPT; /* Cell extends past end of page */ @@ -57800,87 +57068,54 @@ unsigned char *pPrior; unsigned char *pPayload; BtShared *pBt = pPage->pBt; Pgno pgnoOvfl = 0; int nHeader; + CellInfo info; assert( sqlite3_mutex_held(pPage->pBt->mutex) ); /* pPage is not necessarily writeable since pCell might be auxiliary ** buffer space that is separate from the pPage buffer area */ assert( pCell aData || pCell>=&pPage->aData[pBt->pageSize] || sqlite3PagerIswriteable(pPage->pDbPage) ); /* Fill in the header. */ - nHeader = pPage->childPtrSize; - nPayload = nData + nZero; - if( pPage->intKeyLeaf ){ - nHeader += putVarint32(&pCell[nHeader], nPayload); + nHeader = 0; + if( !pPage->leaf ){ + nHeader += 4; + } + if( pPage->hasData ){ + nHeader += putVarint32(&pCell[nHeader], nData+nZero); }else{ - assert( nData==0 ); - assert( nZero==0 ); + nData = nZero = 0; } nHeader += putVarint(&pCell[nHeader], *(u64*)&nKey); + btreeParseCellPtr(pPage, pCell, &info); + assert( info.nHeader==nHeader ); + assert( info.nKey==nKey ); + assert( info.nData==(u32)(nData+nZero) ); - /* Fill in the payload size */ + /* Fill in the payload */ + nPayload = nData + nZero; if( pPage->intKey ){ pSrc = pData; nSrc = nData; nData = 0; }else{ if( NEVER(nKey>0x7fffffff || pKey==0) ){ return SQLITE_CORRUPT_BKPT; } - nPayload = (int)nKey; + nPayload += (int)nKey; pSrc = pKey; nSrc = (int)nKey; } - if( nPayload<=pPage->maxLocal ){ - n = nHeader + nPayload; - testcase( n==3 ); - testcase( n==4 ); - if( n<4 ) n = 4; - *pnSize = n; - spaceLeft = nPayload; - pPrior = pCell; - }else{ - int mn = pPage->minLocal; - n = mn + (nPayload - mn) % (pPage->pBt->usableSize - 4); - testcase( n==pPage->maxLocal ); - testcase( n==pPage->maxLocal+1 ); - if( n > pPage->maxLocal ) n = mn; - spaceLeft = n; - *pnSize = n + nHeader + 4; - pPrior = &pCell[nHeader+n]; - } + *pnSize = info.nSize; + spaceLeft = info.nLocal; pPayload = &pCell[nHeader]; - - /* At this point variables should be set as follows: - ** - ** nPayload Total payload size in bytes - ** pPayload Begin writing payload here - ** spaceLeft Space available at pPayload. If nPayload>spaceLeft, - ** that means content must spill into overflow pages. - ** *pnSize Size of the local cell (not counting overflow pages) - ** pPrior Where to write the pgno of the first overflow page - ** - ** Use a call to btreeParseCellPtr() to verify that the values above - ** were computed correctly. - */ -#if SQLITE_DEBUG - { - CellInfo info; - btreeParseCellPtr(pPage, pCell, &info); - assert( nHeader=(int)(info.pPayload - pCell) ); - assert( info.nKey==nKey ); - assert( *pnSize == info.nSize ); - assert( spaceLeft == info.nLocal ); - assert( pPrior == &pCell[info.iOverflow] ); - } -#endif - - /* Write the payload into the local Cell and any extra into overflow pages */ + pPrior = &pCell[info.iOverflow]; + while( nPayload>0 ){ if( spaceLeft==0 ){ #ifndef SQLITE_OMIT_AUTOVACUUM Pgno pgnoPtrmap = pgnoOvfl; /* Overflow page pointer-map entry page */ if( pBt->autoVacuum ){ @@ -58108,11 +57343,11 @@ /* ** Add a list of cells to a page. The page should be initially empty. ** The cells are guaranteed to fit on the page. */ static void assemblePage( - MemPage *pPage, /* The page to be assembled */ + MemPage *pPage, /* The page to be assemblied */ int nCell, /* The number of cells to add to this page */ u8 **apCell, /* Pointers to cell bodies */ u16 *aSize /* Sizes of the cells */ ){ int i; /* Loop counter */ @@ -58584,11 +57819,11 @@ ** ** leafCorrection: 4 if pPage is a leaf. 0 if pPage is not a leaf. ** leafData: 1 if pPage holds key+data and pParent holds only keys. */ leafCorrection = apOld[0]->leaf*4; - leafData = apOld[0]->intKeyLeaf; + leafData = apOld[0]->hasData; for(i=0; i autoVacuum && aRoot[i]>1 ){ - checkPtrmap(&sCheck, aRoot[i], PTRMAP_ROOTPAGE, 0); + checkPtrmap(&sCheck, aRoot[i], PTRMAP_ROOTPAGE, 0, 0); } #endif - sCheck.zPfx = "List of tree roots: "; - checkTreePage(&sCheck, aRoot[i], NULL, NULL); - sCheck.zPfx = 0; + checkTreePage(&sCheck, aRoot[i], "List of tree roots: ", NULL, NULL); } /* Make sure every page in the file is referenced */ for(i=1; i<=sCheck.nPage && sCheck.mxErr; i++){ #ifdef SQLITE_OMIT_AUTOVACUUM if( getPageReferenced(&sCheck, i)==0 ){ - checkAppendMsg(&sCheck, "Page %d is never used", i); + checkAppendMsg(&sCheck, 0, "Page %d is never used", i); } #else /* If the database supports auto-vacuum, make sure no tables contain ** references to pointer-map pages. */ if( getPageReferenced(&sCheck, i)==0 && (PTRMAP_PAGENO(pBt, i)!=i || !pBt->autoVacuum) ){ - checkAppendMsg(&sCheck, "Page %d is never used", i); + checkAppendMsg(&sCheck, 0, "Page %d is never used", i); } if( getPageReferenced(&sCheck, i)!=0 && (PTRMAP_PAGENO(pBt, i)==i && pBt->autoVacuum) ){ - checkAppendMsg(&sCheck, "Pointer map page %d is referenced", i); + checkAppendMsg(&sCheck, 0, "Pointer map page %d is referenced", i); } #endif } /* Make sure this analysis did not leave any unref() pages. ** This is an internal consistency check; an integrity check ** of the integrity check. */ if( NEVER(nRef != sqlite3PagerRefcount(pBt->pPager)) ){ - checkAppendMsg(&sCheck, + checkAppendMsg(&sCheck, 0, "Outstanding page count goes from %d to %d during this analysis", nRef, sqlite3PagerRefcount(pBt->pPager) ); } @@ -60763,11 +59980,11 @@ /* Save the positions of all other cursors open on this table. This is ** required in case any of them are holding references to an xFetch ** version of the b-tree page modified by the accessPayload call below. ** - ** Note that pCsr must be open on a INTKEY table and saveCursorPosition() + ** Note that pCsr must be open on a BTREE_INTKEY table and saveCursorPosition() ** and hence saveAllCursors() cannot fail on a BTREE_INTKEY table, hence ** saveAllCursors can only return SQLITE_OK. */ VVA_ONLY(rc =) saveAllCursors(pCsr->pBt, pCsr->pgnoRoot, pCsr); assert( rc==SQLITE_OK ); @@ -60938,16 +60155,16 @@ if( i==1 ){ Parse *pParse; int rc = 0; pParse = sqlite3StackAllocZero(pErrorDb, sizeof(*pParse)); if( pParse==0 ){ - sqlite3ErrorWithMsg(pErrorDb, SQLITE_NOMEM, "out of memory"); + sqlite3Error(pErrorDb, SQLITE_NOMEM, "out of memory"); rc = SQLITE_NOMEM; }else{ pParse->db = pDb; if( sqlite3OpenTempDatabase(pParse) ){ - sqlite3ErrorWithMsg(pErrorDb, pParse->rc, "%s", pParse->zErrMsg); + sqlite3Error(pErrorDb, pParse->rc, "%s", pParse->zErrMsg); rc = SQLITE_ERROR; } sqlite3DbFree(pErrorDb, pParse->zErrMsg); sqlite3ParserReset(pParse); sqlite3StackFree(pErrorDb, pParse); @@ -60956,11 +60173,11 @@ return 0; } } if( i<0 ){ - sqlite3ErrorWithMsg(pErrorDb, SQLITE_ERROR, "unknown database %s", zDb); + sqlite3Error(pErrorDb, SQLITE_ERROR, "unknown database %s", zDb); return 0; } return pDb->aDb[i].pBt; } @@ -61001,11 +60218,11 @@ */ sqlite3_mutex_enter(pSrcDb->mutex); sqlite3_mutex_enter(pDestDb->mutex); if( pSrcDb==pDestDb ){ - sqlite3ErrorWithMsg( + sqlite3Error( pDestDb, SQLITE_ERROR, "source and destination must be distinct" ); p = 0; }else { /* Allocate space for a new sqlite3_backup object... @@ -61012,11 +60229,11 @@ ** EVIDENCE-OF: R-64852-21591 The sqlite3_backup object is created by a ** call to sqlite3_backup_init() and is destroyed by a call to ** sqlite3_backup_finish(). */ p = (sqlite3_backup *)sqlite3MallocZero(sizeof(sqlite3_backup)); if( !p ){ - sqlite3Error(pDestDb, SQLITE_NOMEM); + sqlite3Error(pDestDb, SQLITE_NOMEM, 0); } } /* If the allocation succeeded, populate the new object. */ if( p ){ @@ -61453,11 +60670,11 @@ sqlite3BtreeRollback(p->pDest, SQLITE_OK); /* Set the error code of the destination database handle. */ rc = (p->rc==SQLITE_DONE) ? SQLITE_OK : p->rc; if( p->pDestDb ){ - sqlite3Error(p->pDestDb, rc); + sqlite3Error(p->pDestDb, rc, 0); /* Exit the mutexes and free the backup context structure. */ sqlite3LeaveMutexAndCloseZombie(p->pDestDb); } sqlite3BtreeLeave(p->pSrc); @@ -61626,44 +60843,33 @@ ** ** This routine is intended for use inside of assert() statements, like ** this: assert( sqlite3VdbeCheckMemInvariants(pMem) ); */ SQLITE_PRIVATE int sqlite3VdbeCheckMemInvariants(Mem *p){ - /* If MEM_Dyn is set then Mem.xDel!=0. - ** Mem.xDel is might not be initialized if MEM_Dyn is clear. + /* The MEM_Dyn bit is set if and only if Mem.xDel is a non-NULL destructor + ** function for Mem.z */ assert( (p->flags & MEM_Dyn)==0 || p->xDel!=0 ); - - /* MEM_Dyn may only be set if Mem.szMalloc==0. In this way we - ** ensure that if Mem.szMalloc>0 then it is safe to do - ** Mem.z = Mem.zMalloc without having to check Mem.flags&MEM_Dyn. - ** That saves a few cycles in inner loops. */ - assert( (p->flags & MEM_Dyn)==0 || p->szMalloc==0 ); - - /* Cannot be both MEM_Int and MEM_Real at the same time */ - assert( (p->flags & (MEM_Int|MEM_Real))!=(MEM_Int|MEM_Real) ); - - /* The szMalloc field holds the correct memory allocation size */ - assert( p->szMalloc==0 - || p->szMalloc==sqlite3DbMallocSize(p->db,p->zMalloc) ); + assert( (p->flags & MEM_Dyn)!=0 || p->xDel==0 ); /* If p holds a string or blob, the Mem.z must point to exactly ** one of the following: ** ** (1) Memory in Mem.zMalloc and managed by the Mem object ** (2) Memory to be freed using Mem.xDel - ** (3) An ephemeral string or blob + ** (3) An ephermal string or blob ** (4) A static string or blob */ - if( (p->flags & (MEM_Str|MEM_Blob)) && p->n>0 ){ + if( (p->flags & (MEM_Str|MEM_Blob)) && p->z!=0 ){ assert( - ((p->szMalloc>0 && p->z==p->zMalloc)? 1 : 0) + + ((p->z==p->zMalloc)? 1 : 0) + ((p->flags&MEM_Dyn)!=0 ? 1 : 0) + ((p->flags&MEM_Ephem)!=0 ? 1 : 0) + ((p->flags&MEM_Static)!=0 ? 1 : 0) == 1 ); } + return 1; } #endif @@ -61713,37 +60919,33 @@ ** If the bPreserve argument is true, then copy of the content of ** pMem->z into the new allocation. pMem must be either a string or ** blob if bPreserve is true. If bPreserve is false, any prior content ** in pMem->z is discarded. */ -SQLITE_PRIVATE SQLITE_NOINLINE int sqlite3VdbeMemGrow(Mem *pMem, int n, int bPreserve){ +SQLITE_PRIVATE int sqlite3VdbeMemGrow(Mem *pMem, int n, int bPreserve){ assert( sqlite3VdbeCheckMemInvariants(pMem) ); assert( (pMem->flags&MEM_RowSet)==0 ); /* If the bPreserve flag is set to true, then the memory cell must already ** contain a valid string or blob value. */ assert( bPreserve==0 || pMem->flags&(MEM_Blob|MEM_Str) ); testcase( bPreserve && pMem->z==0 ); - assert( pMem->szMalloc==0 - || pMem->szMalloc==sqlite3DbMallocSize(pMem->db, pMem->zMalloc) ); - if( pMem->szMalloc zMalloc==0 || sqlite3DbMallocSize(pMem->db, pMem->zMalloc) szMalloc>0 && pMem->z==pMem->zMalloc ){ + if( bPreserve && pMem->z==pMem->zMalloc ){ pMem->z = pMem->zMalloc = sqlite3DbReallocOrFree(pMem->db, pMem->z, n); bPreserve = 0; }else{ - if( pMem->szMalloc>0 ) sqlite3DbFree(pMem->db, pMem->zMalloc); + sqlite3DbFree(pMem->db, pMem->zMalloc); pMem->zMalloc = sqlite3DbMallocRaw(pMem->db, n); } if( pMem->zMalloc==0 ){ - sqlite3VdbeMemSetNull(pMem); + VdbeMemRelease(pMem); pMem->z = 0; - pMem->szMalloc = 0; + pMem->flags = MEM_Null; return SQLITE_NOMEM; - }else{ - pMem->szMalloc = sqlite3DbMallocSize(pMem->db, pMem->zMalloc); } } if( pMem->z && bPreserve && pMem->z!=pMem->zMalloc ){ memcpy(pMem->zMalloc, pMem->z, pMem->n); @@ -61753,51 +60955,29 @@ pMem->xDel((void *)(pMem->z)); } pMem->z = pMem->zMalloc; pMem->flags &= ~(MEM_Dyn|MEM_Ephem|MEM_Static); - return SQLITE_OK; -} - -/* -** Change the pMem->zMalloc allocation to be at least szNew bytes. -** If pMem->zMalloc already meets or exceeds the requested size, this -** routine is a no-op. -** -** Any prior string or blob content in the pMem object may be discarded. -** The pMem->xDel destructor is called, if it exists. Though MEM_Str -** and MEM_Blob values may be discarded, MEM_Int, MEM_Real, and MEM_Null -** values are preserved. -** -** Return SQLITE_OK on success or an error code (probably SQLITE_NOMEM) -** if unable to complete the resizing. -*/ -SQLITE_PRIVATE int sqlite3VdbeMemClearAndResize(Mem *pMem, int szNew){ - assert( szNew>0 ); - assert( (pMem->flags & MEM_Dyn)==0 || pMem->szMalloc==0 ); - if( pMem->szMalloc flags & MEM_Dyn)==0 ); - pMem->z = pMem->zMalloc; - pMem->flags &= (MEM_Null|MEM_Int|MEM_Real); - return SQLITE_OK; -} - -/* -** Change pMem so that its MEM_Str or MEM_Blob value is stored in -** MEM.zMalloc, where it can be safely written. + pMem->xDel = 0; + return SQLITE_OK; +} + +/* +** Make the given Mem object MEM_Dyn. In other words, make it so +** that any TEXT or BLOB content is stored in memory obtained from +** malloc(). In this way, we know that the memory is safe to be +** overwritten or altered. ** ** Return SQLITE_OK on success or SQLITE_NOMEM if malloc fails. */ SQLITE_PRIVATE int sqlite3VdbeMemMakeWriteable(Mem *pMem){ int f; assert( pMem->db==0 || sqlite3_mutex_held(pMem->db->mutex) ); assert( (pMem->flags&MEM_RowSet)==0 ); ExpandBlob(pMem); f = pMem->flags; - if( (f&(MEM_Str|MEM_Blob)) && (pMem->szMalloc==0 || pMem->z!=pMem->zMalloc) ){ + if( (f&(MEM_Str|MEM_Blob)) && pMem->z!=pMem->zMalloc ){ if( sqlite3VdbeMemGrow(pMem, pMem->n + 2, 1) ){ return SQLITE_NOMEM; } pMem->z[pMem->n] = 0; pMem->z[pMem->n+1] = 0; @@ -61836,54 +61016,44 @@ pMem->flags &= ~(MEM_Zero|MEM_Term); } return SQLITE_OK; } #endif + /* -** It is already known that pMem contains an unterminated string. -** Add the zero terminator. +** Make sure the given Mem is \u0000 terminated. */ -static SQLITE_NOINLINE int vdbeMemAddTerminator(Mem *pMem){ +SQLITE_PRIVATE int sqlite3VdbeMemNulTerminate(Mem *pMem){ + assert( pMem->db==0 || sqlite3_mutex_held(pMem->db->mutex) ); + if( (pMem->flags & MEM_Term)!=0 || (pMem->flags & MEM_Str)==0 ){ + return SQLITE_OK; /* Nothing to do */ + } if( sqlite3VdbeMemGrow(pMem, pMem->n+2, 1) ){ return SQLITE_NOMEM; } pMem->z[pMem->n] = 0; pMem->z[pMem->n+1] = 0; pMem->flags |= MEM_Term; return SQLITE_OK; } -/* -** Make sure the given Mem is \u0000 terminated. -*/ -SQLITE_PRIVATE int sqlite3VdbeMemNulTerminate(Mem *pMem){ - assert( pMem->db==0 || sqlite3_mutex_held(pMem->db->mutex) ); - testcase( (pMem->flags & (MEM_Term|MEM_Str))==(MEM_Term|MEM_Str) ); - testcase( (pMem->flags & (MEM_Term|MEM_Str))==0 ); - if( (pMem->flags & (MEM_Term|MEM_Str))!=MEM_Str ){ - return SQLITE_OK; /* Nothing to do */ - }else{ - return vdbeMemAddTerminator(pMem); - } -} - /* ** Add MEM_Str to the set of representations for the given Mem. Numbers ** are converted using sqlite3_snprintf(). Converting a BLOB to a string ** is a no-op. ** -** Existing representations MEM_Int and MEM_Real are invalidated if -** bForce is true but are retained if bForce is false. +** Existing representations MEM_Int and MEM_Real are *not* invalidated. ** ** A MEM_Null value will never be passed to this function. This function is ** used for converting values to text for returning to the user (i.e. via ** sqlite3_value_text()), or for ensuring that values to be used as btree ** keys are strings. In the former case a NULL pointer is returned the -** user and the latter is an internal programming error. +** user and the later is an internal programming error. */ -SQLITE_PRIVATE int sqlite3VdbeMemStringify(Mem *pMem, u8 enc, u8 bForce){ +SQLITE_PRIVATE int sqlite3VdbeMemStringify(Mem *pMem, int enc){ + int rc = SQLITE_OK; int fg = pMem->flags; const int nByte = 32; assert( pMem->db==0 || sqlite3_mutex_held(pMem->db->mutex) ); assert( !(fg&MEM_Zero) ); @@ -61891,32 +61061,31 @@ assert( fg&(MEM_Int|MEM_Real) ); assert( (pMem->flags&MEM_RowSet)==0 ); assert( EIGHT_BYTE_ALIGNMENT(pMem) ); - if( sqlite3VdbeMemClearAndResize(pMem, nByte) ){ + if( sqlite3VdbeMemGrow(pMem, nByte, 0) ){ return SQLITE_NOMEM; } - /* For a Real or Integer, use sqlite3_snprintf() to produce the UTF-8 + /* For a Real or Integer, use sqlite3_mprintf() to produce the UTF-8 ** string representation of the value. Then, if the required encoding ** is UTF-16le or UTF-16be do a translation. ** ** FIX ME: It would be better if sqlite3_snprintf() could do UTF-16. */ if( fg & MEM_Int ){ sqlite3_snprintf(nByte, pMem->z, "%lld", pMem->u.i); }else{ assert( fg & MEM_Real ); - sqlite3_snprintf(nByte, pMem->z, "%!.15g", pMem->u.r); + sqlite3_snprintf(nByte, pMem->z, "%!.15g", pMem->r); } pMem->n = sqlite3Strlen30(pMem->z); pMem->enc = SQLITE_UTF8; pMem->flags |= MEM_Str|MEM_Term; - if( bForce ) pMem->flags &= ~(MEM_Int|MEM_Real); sqlite3VdbeChangeEncoding(pMem, enc); - return SQLITE_OK; + return rc; } /* ** Memory cell pMem contains the context of an aggregate function. ** This routine calls the finalize method for that function. The @@ -61927,94 +61096,63 @@ */ SQLITE_PRIVATE int sqlite3VdbeMemFinalize(Mem *pMem, FuncDef *pFunc){ int rc = SQLITE_OK; if( ALWAYS(pFunc && pFunc->xFinalize) ){ sqlite3_context ctx; - Mem t; assert( (pMem->flags & MEM_Null)!=0 || pFunc==pMem->u.pDef ); assert( pMem->db==0 || sqlite3_mutex_held(pMem->db->mutex) ); memset(&ctx, 0, sizeof(ctx)); - memset(&t, 0, sizeof(t)); - t.flags = MEM_Null; - t.db = pMem->db; - ctx.pOut = &t; + ctx.s.flags = MEM_Null; + ctx.s.db = pMem->db; ctx.pMem = pMem; ctx.pFunc = pFunc; pFunc->xFinalize(&ctx); /* IMP: R-24505-23230 */ - assert( (pMem->flags & MEM_Dyn)==0 ); - if( pMem->szMalloc>0 ) sqlite3DbFree(pMem->db, pMem->zMalloc); - memcpy(pMem, &t, sizeof(t)); + assert( 0==(pMem->flags&MEM_Dyn) && !pMem->xDel ); + sqlite3DbFree(pMem->db, pMem->zMalloc); + memcpy(pMem, &ctx.s, sizeof(ctx.s)); rc = ctx.isError; } return rc; } /* -** If the memory cell contains a value that must be freed by -** invoking the external callback in Mem.xDel, then this routine -** will free that value. It also sets Mem.flags to MEM_Null. -** -** This is a helper routine for sqlite3VdbeMemSetNull() and -** for sqlite3VdbeMemRelease(). Use those other routines as the -** entry point for releasing Mem resources. +** If the memory cell contains a string value that must be freed by +** invoking an external callback, free it now. Calling this function +** does not free any Mem.zMalloc buffer. */ -static SQLITE_NOINLINE void vdbeMemClearExternAndSetNull(Mem *p){ +SQLITE_PRIVATE void sqlite3VdbeMemReleaseExternal(Mem *p){ assert( p->db==0 || sqlite3_mutex_held(p->db->mutex) ); - assert( VdbeMemDynamic(p) ); if( p->flags&MEM_Agg ){ sqlite3VdbeMemFinalize(p, p->u.pDef); assert( (p->flags & MEM_Agg)==0 ); - testcase( p->flags & MEM_Dyn ); - } - if( p->flags&MEM_Dyn ){ + sqlite3VdbeMemRelease(p); + }else if( p->flags&MEM_Dyn ){ assert( (p->flags&MEM_RowSet)==0 ); assert( p->xDel!=SQLITE_DYNAMIC && p->xDel!=0 ); p->xDel((void *)p->z); + p->xDel = 0; }else if( p->flags&MEM_RowSet ){ sqlite3RowSetClear(p->u.pRowSet); }else if( p->flags&MEM_Frame ){ - VdbeFrame *pFrame = p->u.pFrame; - pFrame->pParent = pFrame->v->pDelFrame; - pFrame->v->pDelFrame = pFrame; - } - p->flags = MEM_Null; -} - -/* -** Release memory held by the Mem p, both external memory cleared -** by p->xDel and memory in p->zMalloc. -** -** This is a helper routine invoked by sqlite3VdbeMemRelease() in -** the unusual case where there really is memory in p that needs -** to be freed. -*/ -static SQLITE_NOINLINE void vdbeMemClear(Mem *p){ - if( VdbeMemDynamic(p) ){ - vdbeMemClearExternAndSetNull(p); - } - if( p->szMalloc ){ - sqlite3DbFree(p->db, p->zMalloc); - p->szMalloc = 0; - } - p->z = 0; -} - -/* -** Release any memory resources held by the Mem. Both the memory that is -** free by Mem.xDel and the Mem.zMalloc allocation are freed. -** -** Use this routine prior to clean up prior to abandoning a Mem, or to -** reset a Mem back to its minimum memory utilization. -** -** Use sqlite3VdbeMemSetNull() to release just the Mem.xDel space -** prior to inserting new content into the Mem. + sqlite3VdbeMemSetNull(p); + } +} + +/* +** Release any memory held by the Mem. This may leave the Mem in an +** inconsistent state, for example with (Mem.z==0) and +** (Mem.flags==MEM_Str). */ SQLITE_PRIVATE void sqlite3VdbeMemRelease(Mem *p){ assert( sqlite3VdbeCheckMemInvariants(p) ); - if( VdbeMemDynamic(p) || p->szMalloc ){ - vdbeMemClear(p); + VdbeMemRelease(p); + if( p->zMalloc ){ + sqlite3DbFree(p->db, p->zMalloc); + p->zMalloc = 0; } + p->z = 0; + assert( p->xDel==0 ); /* Zeroed by VdbeMemRelease() above */ } /* ** Convert a 64-bit IEEE double into a 64-bit signed integer. ** If the double is out of range of a 64-bit signed integer then @@ -62049,11 +61187,11 @@ ** Return some kind of integer value which is the best we can do ** at representing the value that *pMem describes as an integer. ** If pMem is an integer, then the value is exact. If pMem is ** a floating-point then the value returned is the integer part. ** If pMem is a string or blob, then we make an attempt to convert -** it into an integer and return that. If pMem represents an +** it into a integer and return that. If pMem represents an ** an SQL-NULL value, return 0. ** ** If pMem represents a string value, its encoding might be changed. */ SQLITE_PRIVATE i64 sqlite3VdbeIntValue(Mem *pMem){ @@ -62062,14 +61200,15 @@ assert( EIGHT_BYTE_ALIGNMENT(pMem) ); flags = pMem->flags; if( flags & MEM_Int ){ return pMem->u.i; }else if( flags & MEM_Real ){ - return doubleToInt64(pMem->u.r); + return doubleToInt64(pMem->r); }else if( flags & (MEM_Str|MEM_Blob) ){ i64 value = 0; assert( pMem->z || pMem->n==0 ); + testcase( pMem->z==0 ); sqlite3Atoi64(pMem->z, &value, pMem->n, pMem->enc); return value; }else{ return 0; } @@ -62083,11 +61222,11 @@ */ SQLITE_PRIVATE double sqlite3VdbeRealValue(Mem *pMem){ assert( pMem->db==0 || sqlite3_mutex_held(pMem->db->mutex) ); assert( EIGHT_BYTE_ALIGNMENT(pMem) ); if( pMem->flags & MEM_Real ){ - return pMem->u.r; + return pMem->r; }else if( pMem->flags & MEM_Int ){ return (double)pMem->u.i; }else if( pMem->flags & (MEM_Str|MEM_Blob) ){ /* (double)0 In case of SQLITE_OMIT_FLOATING_POINT... */ double val = (double)0; @@ -62102,17 +61241,16 @@ /* ** The MEM structure is already a MEM_Real. Try to also make it a ** MEM_Int if we can. */ SQLITE_PRIVATE void sqlite3VdbeIntegerAffinity(Mem *pMem){ - i64 ix; assert( pMem->flags & MEM_Real ); assert( (pMem->flags & MEM_RowSet)==0 ); assert( pMem->db==0 || sqlite3_mutex_held(pMem->db->mutex) ); assert( EIGHT_BYTE_ALIGNMENT(pMem) ); - ix = doubleToInt64(pMem->u.r); + pMem->u.i = doubleToInt64(pMem->r); /* Only mark the value as an integer if ** ** (1) the round-trip conversion real->int->real is a no-op, and ** (2) The integer is neither the largest nor the smallest @@ -62120,13 +61258,15 @@ ** ** The second and third terms in the following conditional enforces ** the second condition under the assumption that addition overflow causes ** values to wrap around. */ - if( pMem->u.r==ix && ix>SMALLEST_INT64 && ix u.i = ix; - MemSetTypeFlag(pMem, MEM_Int); + if( pMem->r==(double)pMem->u.i + && pMem->u.i>SMALLEST_INT64 + && pMem->u.i flags |= MEM_Int; } } /* ** Convert pMem to type integer. Invalidate any prior representations. @@ -62147,11 +61287,11 @@ */ SQLITE_PRIVATE int sqlite3VdbeMemRealify(Mem *pMem){ assert( pMem->db==0 || sqlite3_mutex_held(pMem->db->mutex) ); assert( EIGHT_BYTE_ALIGNMENT(pMem) ); - pMem->u.r = sqlite3VdbeRealValue(pMem); + pMem->r = sqlite3VdbeRealValue(pMem); MemSetTypeFlag(pMem, MEM_Real); return SQLITE_OK; } /* @@ -62167,95 +61307,33 @@ assert( (pMem->flags & (MEM_Blob|MEM_Str))!=0 ); assert( pMem->db==0 || sqlite3_mutex_held(pMem->db->mutex) ); if( 0==sqlite3Atoi64(pMem->z, &pMem->u.i, pMem->n, pMem->enc) ){ MemSetTypeFlag(pMem, MEM_Int); }else{ - pMem->u.r = sqlite3VdbeRealValue(pMem); + pMem->r = sqlite3VdbeRealValue(pMem); MemSetTypeFlag(pMem, MEM_Real); sqlite3VdbeIntegerAffinity(pMem); } } assert( (pMem->flags & (MEM_Int|MEM_Real|MEM_Null))!=0 ); pMem->flags &= ~(MEM_Str|MEM_Blob); return SQLITE_OK; } - -/* -** Cast the datatype of the value in pMem according to the affinity -** "aff". Casting is different from applying affinity in that a cast -** is forced. In other words, the value is converted into the desired -** affinity even if that results in loss of data. This routine is -** used (for example) to implement the SQL "cast()" operator. -*/ -SQLITE_PRIVATE void sqlite3VdbeMemCast(Mem *pMem, u8 aff, u8 encoding){ - if( pMem->flags & MEM_Null ) return; - switch( aff ){ - case SQLITE_AFF_NONE: { /* Really a cast to BLOB */ - if( (pMem->flags & MEM_Blob)==0 ){ - sqlite3ValueApplyAffinity(pMem, SQLITE_AFF_TEXT, encoding); - assert( pMem->flags & MEM_Str || pMem->db->mallocFailed ); - MemSetTypeFlag(pMem, MEM_Blob); - }else{ - pMem->flags &= ~(MEM_TypeMask&~MEM_Blob); - } - break; - } - case SQLITE_AFF_NUMERIC: { - sqlite3VdbeMemNumerify(pMem); - break; - } - case SQLITE_AFF_INTEGER: { - sqlite3VdbeMemIntegerify(pMem); - break; - } - case SQLITE_AFF_REAL: { - sqlite3VdbeMemRealify(pMem); - break; - } - default: { - assert( aff==SQLITE_AFF_TEXT ); - assert( MEM_Str==(MEM_Blob>>3) ); - pMem->flags |= (pMem->flags&MEM_Blob)>>3; - sqlite3ValueApplyAffinity(pMem, SQLITE_AFF_TEXT, encoding); - assert( pMem->flags & MEM_Str || pMem->db->mallocFailed ); - pMem->flags &= ~(MEM_Int|MEM_Real|MEM_Blob|MEM_Zero); - break; - } - } -} - -/* -** Initialize bulk memory to be a consistent Mem object. -** -** The minimum amount of initialization feasible is performed. -*/ -SQLITE_PRIVATE void sqlite3VdbeMemInit(Mem *pMem, sqlite3 *db, u16 flags){ - assert( (flags & ~MEM_TypeMask)==0 ); - pMem->flags = flags; - pMem->db = db; - pMem->szMalloc = 0; -} - /* ** Delete any previous value and set the value stored in *pMem to NULL. -** -** This routine calls the Mem.xDel destructor to dispose of values that -** require the destructor. But it preserves the Mem.zMalloc memory allocation. -** To free all resources, use sqlite3VdbeMemRelease(), which both calls this -** routine to invoke the destructor and deallocates Mem.zMalloc. -** -** Use this routine to reset the Mem prior to insert a new value. -** -** Use sqlite3VdbeMemRelease() to complete erase the Mem prior to abandoning it. */ SQLITE_PRIVATE void sqlite3VdbeMemSetNull(Mem *pMem){ - if( VdbeMemDynamic(pMem) ){ - vdbeMemClearExternAndSetNull(pMem); - }else{ - pMem->flags = MEM_Null; + if( pMem->flags & MEM_Frame ){ + VdbeFrame *pFrame = pMem->u.pFrame; + pFrame->pParent = pFrame->v->pDelFrame; + pFrame->v->pDelFrame = pFrame; } + if( pMem->flags & MEM_RowSet ){ + sqlite3RowSetClear(pMem->u.pRowSet); + } + MemSetTypeFlag(pMem, MEM_Null); } SQLITE_PRIVATE void sqlite3ValueSetNull(sqlite3_value *p){ sqlite3VdbeMemSetNull((Mem*)p); } @@ -62268,46 +61346,41 @@ pMem->flags = MEM_Blob|MEM_Zero; pMem->n = 0; if( n<0 ) n = 0; pMem->u.nZero = n; pMem->enc = SQLITE_UTF8; - pMem->z = 0; -} - -/* -** The pMem is known to contain content that needs to be destroyed prior -** to a value change. So invoke the destructor, then set the value to -** a 64-bit integer. -*/ -static SQLITE_NOINLINE void vdbeReleaseAndSetInt64(Mem *pMem, i64 val){ - sqlite3VdbeMemSetNull(pMem); - pMem->u.i = val; - pMem->flags = MEM_Int; + +#ifdef SQLITE_OMIT_INCRBLOB + sqlite3VdbeMemGrow(pMem, n, 0); + if( pMem->z ){ + pMem->n = n; + memset(pMem->z, 0, n); + } +#endif } /* ** Delete any previous value and set the value stored in *pMem to val, ** manifest type INTEGER. */ SQLITE_PRIVATE void sqlite3VdbeMemSetInt64(Mem *pMem, i64 val){ - if( VdbeMemDynamic(pMem) ){ - vdbeReleaseAndSetInt64(pMem, val); - }else{ - pMem->u.i = val; - pMem->flags = MEM_Int; - } + sqlite3VdbeMemRelease(pMem); + pMem->u.i = val; + pMem->flags = MEM_Int; } #ifndef SQLITE_OMIT_FLOATING_POINT /* ** Delete any previous value and set the value stored in *pMem to val, ** manifest type REAL. */ SQLITE_PRIVATE void sqlite3VdbeMemSetDouble(Mem *pMem, double val){ - sqlite3VdbeMemSetNull(pMem); - if( !sqlite3IsNaN(val) ){ - pMem->u.r = val; + if( sqlite3IsNaN(val) ){ + sqlite3VdbeMemSetNull(pMem); + }else{ + sqlite3VdbeMemRelease(pMem); + pMem->r = val; pMem->flags = MEM_Real; } } #endif @@ -62321,15 +61394,14 @@ assert( (pMem->flags & MEM_RowSet)==0 ); sqlite3VdbeMemRelease(pMem); pMem->zMalloc = sqlite3DbMallocRaw(db, 64); if( db->mallocFailed ){ pMem->flags = MEM_Null; - pMem->szMalloc = 0; }else{ assert( pMem->zMalloc ); - pMem->szMalloc = sqlite3DbMallocSize(db, pMem->zMalloc); - pMem->u.pRowSet = sqlite3RowSetInit(db, pMem->zMalloc, pMem->szMalloc); + pMem->u.pRowSet = sqlite3RowSetInit(db, pMem->zMalloc, + sqlite3DbMallocSize(db, pMem->zMalloc)); assert( pMem->u.pRowSet!=0 ); pMem->flags = MEM_RowSet; } } @@ -62349,11 +61421,11 @@ return 0; } #ifdef SQLITE_DEBUG /* -** This routine prepares a memory cell for modification by breaking +** This routine prepares a memory cell for modication by breaking ** its link to a shallow copy and by marking any current shallow ** copies of this cell as invalid. ** ** This is used for testing and debugging only - to make sure shallow ** copies are not misused. @@ -62382,13 +61454,13 @@ ** pFrom->z is used, then pTo->z points to the same thing as pFrom->z ** and flags gets srcType (either MEM_Ephem or MEM_Static). */ SQLITE_PRIVATE void sqlite3VdbeMemShallowCopy(Mem *pTo, const Mem *pFrom, int srcType){ assert( (pFrom->flags & MEM_RowSet)==0 ); - assert( pTo->db==pFrom->db ); - if( VdbeMemDynamic(pTo) ) vdbeMemClearExternAndSetNull(pTo); + VdbeMemRelease(pTo); memcpy(pTo, pFrom, MEMCELLSIZE); + pTo->xDel = 0; if( (pFrom->flags&MEM_Static)==0 ){ pTo->flags &= ~(MEM_Dyn|MEM_Static|MEM_Ephem); assert( srcType==MEM_Ephem || srcType==MEM_Static ); pTo->flags |= srcType; } @@ -62399,15 +61471,16 @@ ** freed before the copy is made. */ SQLITE_PRIVATE int sqlite3VdbeMemCopy(Mem *pTo, const Mem *pFrom){ int rc = SQLITE_OK; - assert( pTo->db==pFrom->db ); assert( (pFrom->flags & MEM_RowSet)==0 ); - if( VdbeMemDynamic(pTo) ) vdbeMemClearExternAndSetNull(pTo); + VdbeMemRelease(pTo); memcpy(pTo, pFrom, MEMCELLSIZE); pTo->flags &= ~MEM_Dyn; + pTo->xDel = 0; + if( pTo->flags&(MEM_Str|MEM_Blob) ){ if( 0==(pFrom->flags&MEM_Static) ){ pTo->flags |= MEM_Ephem; rc = sqlite3VdbeMemMakeWriteable(pTo); } @@ -62428,11 +61501,12 @@ assert( pFrom->db==0 || pTo->db==0 || pFrom->db==pTo->db ); sqlite3VdbeMemRelease(pTo); memcpy(pTo, pFrom, sizeof(Mem)); pFrom->flags = MEM_Null; - pFrom->szMalloc = 0; + pFrom->xDel = 0; + pFrom->zMalloc = 0; } /* ** Change the value of a Mem to be a string or a BLOB. ** @@ -62475,12 +61549,11 @@ } flags = (enc==0?MEM_Blob:MEM_Str); if( nByte<0 ){ assert( enc!=0 ); if( enc==SQLITE_UTF8 ){ - nByte = sqlite3Strlen30(z); - if( nByte>iLimit ) nByte = iLimit+1; + for(nByte=0; nByte<=iLimit && z[nByte]; nByte++){} }else{ for(nByte=0; nByte<=iLimit && (z[nByte] | z[nByte+1]); nByte+=2){} } flags |= MEM_Term; } @@ -62495,21 +61568,18 @@ nAlloc += (enc==SQLITE_UTF8?1:2); } if( nByte>iLimit ){ return SQLITE_TOOBIG; } - testcase( nAlloc==0 ); - testcase( nAlloc==31 ); - testcase( nAlloc==32 ); - if( sqlite3VdbeMemClearAndResize(pMem, MAX(nAlloc,32)) ){ + if( sqlite3VdbeMemGrow(pMem, nAlloc, 0) ){ return SQLITE_NOMEM; } memcpy(pMem->z, z, nAlloc); }else if( xDel==SQLITE_DYNAMIC ){ sqlite3VdbeMemRelease(pMem); pMem->zMalloc = pMem->z = (char *)z; - pMem->szMalloc = sqlite3DbMallocSize(pMem->db, pMem->zMalloc); + pMem->xDel = 0; }else{ sqlite3VdbeMemRelease(pMem); pMem->z = (char *)z; pMem->xDel = xDel; flags |= ((xDel==SQLITE_STATIC)?MEM_Static:MEM_Dyn); @@ -62537,15 +61607,12 @@ ** The data or key is taken from the entry that pCur is currently pointing ** to. offset and amt determine what portion of the data or key to retrieve. ** key is true to get the key or false to get data. The result is written ** into the pMem element. ** -** The pMem object must have been initialized. This routine will use -** pMem->zMalloc to hold the content from the btree, if possible. New -** pMem->zMalloc space will be allocated if necessary. The calling routine -** is responsible for making sure that the pMem object is eventually -** destroyed. +** The pMem structure is assumed to be uninitialized. Any prior content +** is overwritten without being freed. ** ** If this routine fails for any reason (malloc returns NULL or unable ** to read from the disk) then the pMem is left in an inconsistent state. */ SQLITE_PRIVATE int sqlite3VdbeMemFromBtree( @@ -62558,11 +61625,10 @@ char *zData; /* Data from the btree layer */ u32 available = 0; /* Number of bytes available on the local btree page */ int rc = SQLITE_OK; /* Return code */ assert( sqlite3BtreeCursorIsValid(pCur) ); - assert( !VdbeMemDynamic(pMem) ); /* Note: the calls to BtreeKeyFetch() and DataFetch() below assert() ** that both the BtShared and database handle mutexes are held. */ assert( (pMem->flags & MEM_RowSet)==0 ); if( key ){ @@ -62571,74 +61637,33 @@ zData = (char *)sqlite3BtreeDataFetch(pCur, &available); } assert( zData!=0 ); if( offset+amt<=available ){ + sqlite3VdbeMemRelease(pMem); pMem->z = &zData[offset]; pMem->flags = MEM_Blob|MEM_Ephem; pMem->n = (int)amt; - }else{ - pMem->flags = MEM_Null; - if( SQLITE_OK==(rc = sqlite3VdbeMemClearAndResize(pMem, amt+2)) ){ - if( key ){ - rc = sqlite3BtreeKey(pCur, offset, amt, pMem->z); - }else{ - rc = sqlite3BtreeData(pCur, offset, amt, pMem->z); - } - if( rc==SQLITE_OK ){ - pMem->z[amt] = 0; - pMem->z[amt+1] = 0; - pMem->flags = MEM_Blob|MEM_Term; - pMem->n = (int)amt; - }else{ - sqlite3VdbeMemRelease(pMem); - } + }else if( SQLITE_OK==(rc = sqlite3VdbeMemGrow(pMem, amt+2, 0)) ){ + if( key ){ + rc = sqlite3BtreeKey(pCur, offset, amt, pMem->z); + }else{ + rc = sqlite3BtreeData(pCur, offset, amt, pMem->z); + } + if( rc==SQLITE_OK ){ + pMem->z[amt] = 0; + pMem->z[amt+1] = 0; + pMem->flags = MEM_Blob|MEM_Term; + pMem->n = (int)amt; + }else{ + sqlite3VdbeMemRelease(pMem); } } return rc; } -/* -** The pVal argument is known to be a value other than NULL. -** Convert it into a string with encoding enc and return a pointer -** to a zero-terminated version of that string. -*/ -static SQLITE_NOINLINE const void *valueToText(sqlite3_value* pVal, u8 enc){ - assert( pVal!=0 ); - assert( pVal->db==0 || sqlite3_mutex_held(pVal->db->mutex) ); - assert( (enc&3)==(enc&~SQLITE_UTF16_ALIGNED) ); - assert( (pVal->flags & MEM_RowSet)==0 ); - assert( (pVal->flags & (MEM_Null))==0 ); - if( pVal->flags & (MEM_Blob|MEM_Str) ){ - pVal->flags |= MEM_Str; - if( pVal->flags & MEM_Zero ){ - sqlite3VdbeMemExpandBlob(pVal); - } - if( pVal->enc != (enc & ~SQLITE_UTF16_ALIGNED) ){ - sqlite3VdbeChangeEncoding(pVal, enc & ~SQLITE_UTF16_ALIGNED); - } - if( (enc & SQLITE_UTF16_ALIGNED)!=0 && 1==(1&SQLITE_PTR_TO_INT(pVal->z)) ){ - assert( (pVal->flags & (MEM_Ephem|MEM_Static))!=0 ); - if( sqlite3VdbeMemMakeWriteable(pVal)!=SQLITE_OK ){ - return 0; - } - } - sqlite3VdbeMemNulTerminate(pVal); /* IMP: R-31275-44060 */ - }else{ - sqlite3VdbeMemStringify(pVal, enc, 0); - assert( 0==(1&SQLITE_PTR_TO_INT(pVal->z)) ); - } - assert(pVal->enc==(enc & ~SQLITE_UTF16_ALIGNED) || pVal->db==0 - || pVal->db->mallocFailed ); - if( pVal->enc==(enc & ~SQLITE_UTF16_ALIGNED) ){ - return pVal->z; - }else{ - return 0; - } -} - /* This function is only available internally, it is not part of the ** external API. It works in a similar way to sqlite3_value_text(), ** except the data returned is in the encoding specified by the second ** parameter, which must be one of SQLITE_UTF16BE, SQLITE_UTF16LE or ** SQLITE_UTF8. @@ -62647,20 +61672,42 @@ ** If that is the case, then the result must be aligned on an even byte ** boundary. */ SQLITE_PRIVATE const void *sqlite3ValueText(sqlite3_value* pVal, u8 enc){ if( !pVal ) return 0; + assert( pVal->db==0 || sqlite3_mutex_held(pVal->db->mutex) ); assert( (enc&3)==(enc&~SQLITE_UTF16_ALIGNED) ); assert( (pVal->flags & MEM_RowSet)==0 ); - if( (pVal->flags&(MEM_Str|MEM_Term))==(MEM_Str|MEM_Term) && pVal->enc==enc ){ - return pVal->z; - } + if( pVal->flags&MEM_Null ){ return 0; } - return valueToText(pVal, enc); + assert( (MEM_Blob>>3) == MEM_Str ); + pVal->flags |= (pVal->flags & MEM_Blob)>>3; + ExpandBlob(pVal); + if( pVal->flags&MEM_Str ){ + sqlite3VdbeChangeEncoding(pVal, enc & ~SQLITE_UTF16_ALIGNED); + if( (enc & SQLITE_UTF16_ALIGNED)!=0 && 1==(1&SQLITE_PTR_TO_INT(pVal->z)) ){ + assert( (pVal->flags & (MEM_Ephem|MEM_Static))!=0 ); + if( sqlite3VdbeMemMakeWriteable(pVal)!=SQLITE_OK ){ + return 0; + } + } + sqlite3VdbeMemNulTerminate(pVal); /* IMP: R-31275-44060 */ + }else{ + assert( (pVal->flags&MEM_Blob)==0 ); + sqlite3VdbeMemStringify(pVal, enc); + assert( 0==(1&SQLITE_PTR_TO_INT(pVal->z)) ); + } + assert(pVal->enc==(enc & ~SQLITE_UTF16_ALIGNED) || pVal->db==0 + || pVal->db->mallocFailed ); + if( pVal->enc==(enc & ~SQLITE_UTF16_ALIGNED) ){ + return pVal->z; + }else{ + return 0; + } } /* ** Create a new sqlite3_value object. */ @@ -62763,24 +61810,13 @@ if( !pExpr ){ *ppVal = 0; return SQLITE_OK; } - while( (op = pExpr->op)==TK_UPLUS ) pExpr = pExpr->pLeft; + op = pExpr->op; if( NEVER(op==TK_REGISTER) ) op = pExpr->op2; - if( op==TK_CAST ){ - u8 aff = sqlite3AffinityType(pExpr->u.zToken,0); - rc = valueFromExpr(db, pExpr->pLeft, enc, aff, ppVal, pCtx); - testcase( rc!=SQLITE_OK ); - if( *ppVal ){ - sqlite3VdbeMemCast(*ppVal, aff, SQLITE_UTF8); - sqlite3ValueApplyAffinity(*ppVal, affinity, SQLITE_UTF8); - } - return rc; - } - /* Handle negative integers in a single step. This is needed in the ** case when the value is -9223372036854775808. */ if( op==TK_UMINUS && (pExpr->pLeft->op==TK_INTEGER || pExpr->pLeft->op==TK_FLOAT) ){ @@ -62813,18 +61849,18 @@ /* This branch happens for multiple negative signs. Ex: -(-5) */ if( SQLITE_OK==sqlite3ValueFromExpr(db,pExpr->pLeft,enc,affinity,&pVal) && pVal!=0 ){ sqlite3VdbeMemNumerify(pVal); - if( pVal->flags & MEM_Real ){ - pVal->u.r = -pVal->u.r; - }else if( pVal->u.i==SMALLEST_INT64 ){ - pVal->u.r = -(double)SMALLEST_INT64; - MemSetTypeFlag(pVal, MEM_Real); + if( pVal->u.i==SMALLEST_INT64 ){ + pVal->flags &= ~MEM_Int; + pVal->flags |= MEM_Real; + pVal->r = (double)SMALLEST_INT64; }else{ pVal->u.i = -pVal->u.i; } + pVal->r = -pVal->r; sqlite3ValueApplyAffinity(pVal, affinity, enc); } }else if( op==TK_NULL ){ pVal = valueNew(db, pCtx); if( pVal==0 ) goto no_mem; @@ -62911,11 +61947,11 @@ aRet = sqlite3DbMallocRaw(db, nRet); if( aRet==0 ){ sqlite3_result_error_nomem(context); }else{ aRet[0] = nSerial+1; - putVarint32(&aRet[1], iSerial); + sqlite3PutVarint(&aRet[1], iSerial); sqlite3VdbeSerialPut(&aRet[1+nSerial], argv[0], iSerial); sqlite3_result_blob(context, aRet, nRet, SQLITE_TRANSIENT); sqlite3DbFree(db, aRet); } } @@ -63128,11 +62164,11 @@ int i; int nCol = pRec->pKeyInfo->nField+pRec->pKeyInfo->nXField; Mem *aMem = pRec->aMem; sqlite3 *db = aMem[0].db; for(i=0; i zMalloc); + } return; } - do{ + for(pEnd=&p[N]; p mallocFailed = malloc_failed; } } /* @@ -64601,11 +63638,11 @@ pMem->flags = MEM_Int; pMem->u.i = pOp->p3; /* P3 */ pMem++; - if( sqlite3VdbeMemClearAndResize(pMem, 32) ){ /* P4 */ + if( sqlite3VdbeMemGrow(pMem, 32, 0) ){ /* P4 */ assert( p->db->mallocFailed ); return SQLITE_ERROR; } pMem->flags = MEM_Str|MEM_Term; zP4 = displayP4(pOp, pMem->z, 32); @@ -64617,11 +63654,11 @@ pMem->enc = SQLITE_UTF8; } pMem++; if( p->explain==1 ){ - if( sqlite3VdbeMemClearAndResize(pMem, 4) ){ + if( sqlite3VdbeMemGrow(pMem, 4, 0) ){ assert( p->db->mallocFailed ); return SQLITE_ERROR; } pMem->flags = MEM_Str|MEM_Term; pMem->n = 2; @@ -64628,11 +63665,11 @@ sqlite3_snprintf(3, pMem->z, "%.2x", pOp->p5); /* P5 */ pMem->enc = SQLITE_UTF8; pMem++; #ifdef SQLITE_ENABLE_EXPLAIN_COMMENTS - if( sqlite3VdbeMemClearAndResize(pMem, 500) ){ + if( sqlite3VdbeMemGrow(pMem, 500, 0) ){ assert( p->db->mallocFailed ); return SQLITE_ERROR; } pMem->flags = MEM_Str|MEM_Term; pMem->n = displayComment(pOp, zP4, pMem->z, 500); @@ -64781,17 +63818,17 @@ } /* ** Prepare a virtual machine for execution for the first time after ** creating the virtual machine. This involves things such -** as allocating registers and initializing the program counter. +** as allocating stack space and initializing the program counter. ** After the VDBE has be prepped, it can be executed by one or more ** calls to sqlite3VdbeExec(). ** -** This function may be called exactly once on each virtual machine. +** This function may be called exact once on a each virtual machine. ** After this routine is called the VM has been "packaged" and is ready -** to run. After this routine is called, further calls to +** to run. After this routine is called, futher calls to ** sqlite3VdbeAddOp() functions are prohibited. This routine disconnects ** the Vdbe from the Parse object that helped generate it so that the ** the Vdbe becomes an independent entity and the Parse object can be ** destroyed. ** @@ -64993,11 +64030,15 @@ sqlite3VdbeDeleteAuxData(p, -1, 0); assert( p->pAuxData==0 ); } /* -** Clean up the VM after a single run. +** Clean up the VM after execution. +** +** This routine will automatically close any cursors, lists, and/or +** sorters that were left open. It also deletes the values of +** variables in the aVar[] array. */ static void Cleanup(Vdbe *p){ sqlite3 *db = p->db; #ifdef SQLITE_DEBUG @@ -65161,11 +64202,11 @@ } } /* The complex case - There is a multi-file write-transaction active. ** This requires a master journal file to ensure the transaction is - ** committed atomically. + ** committed atomicly. */ #ifndef SQLITE_OMIT_DISKIO else{ sqlite3_vfs *pVfs = db->pVfs; int needSync = 0; @@ -65668,11 +64709,11 @@ sqlite3ValueSetStr(db->pErr, -1, p->zErrMsg, SQLITE_UTF8, SQLITE_TRANSIENT); sqlite3EndBenignMalloc(); db->mallocFailed = mallocFailed; db->errCode = rc; }else{ - sqlite3Error(db, rc); + sqlite3Error(db, rc, 0); } return rc; } #ifdef SQLITE_ENABLE_SQLLOG @@ -65731,11 +64772,11 @@ }else if( p->rc && p->expired ){ /* The expired flag was set on the VDBE before the first call ** to sqlite3_step(). For consistency (since sqlite3_step() was ** called), set the database error in this case as well. */ - sqlite3ErrorWithMsg(db, p->rc, p->zErrMsg ? "%s" : 0, p->zErrMsg); + sqlite3Error(db, p->rc, p->zErrMsg ? "%s" : 0, p->zErrMsg); sqlite3DbFree(db, p->zErrMsg); p->zErrMsg = 0; } /* Reclaim all memory used by the VDBE @@ -65809,11 +64850,11 @@ ** ** * the associated function parameter is the 32nd or later (counting ** from left to right), or ** ** * the corresponding bit in argument mask is clear (where the first -** function parameter corresponds to bit 0 etc.). +** function parameter corrsponds to bit 0 etc.). */ SQLITE_PRIVATE void sqlite3VdbeDeleteAuxData(Vdbe *pVdbe, int iOp, int mask){ AuxData **pp = &pVdbe->pAuxData; while( *pp ){ AuxData *pAux = *pp; @@ -65854,10 +64895,14 @@ for(i=p->nzVar-1; i>=0; i--) sqlite3DbFree(db, p->azVar[i]); vdbeFreeOpArray(db, p->aOp, p->nOp); sqlite3DbFree(db, p->aColName); sqlite3DbFree(db, p->zSql); sqlite3DbFree(db, p->pFree); +#if defined(SQLITE_ENABLE_TREE_EXPLAIN) + sqlite3DbFree(db, p->zExplain); + sqlite3DbFree(db, p->pExplain); +#endif } /* ** Delete an entire VDBE. */ @@ -65880,52 +64925,10 @@ p->magic = VDBE_MAGIC_DEAD; p->db = 0; sqlite3DbFree(db, p); } -/* -** The cursor "p" has a pending seek operation that has not yet been -** carried out. Seek the cursor now. If an error occurs, return -** the appropriate error code. -*/ -static int SQLITE_NOINLINE handleDeferredMoveto(VdbeCursor *p){ - int res, rc; -#ifdef SQLITE_TEST - extern int sqlite3_search_count; -#endif - assert( p->deferredMoveto ); - assert( p->isTable ); - rc = sqlite3BtreeMovetoUnpacked(p->pCursor, 0, p->movetoTarget, 0, &res); - if( rc ) return rc; - p->lastRowid = p->movetoTarget; - if( res!=0 ) return SQLITE_CORRUPT_BKPT; - p->rowidIsValid = 1; -#ifdef SQLITE_TEST - sqlite3_search_count++; -#endif - p->deferredMoveto = 0; - p->cacheStatus = CACHE_STALE; - return SQLITE_OK; -} - -/* -** Something has moved cursor "p" out of place. Maybe the row it was -** pointed to was deleted out from under it. Or maybe the btree was -** rebalanced. Whatever the cause, try to restore "p" to the place it -** is supposed to be pointing. If the row was deleted out from under the -** cursor, set the cursor to point to a NULL row. -*/ -static int SQLITE_NOINLINE handleMovedCursor(VdbeCursor *p){ - int isDifferentRow, rc; - assert( p->pCursor!=0 ); - assert( sqlite3BtreeCursorHasMoved(p->pCursor) ); - rc = sqlite3BtreeCursorRestore(p->pCursor, &isDifferentRow); - p->cacheStatus = CACHE_STALE; - if( isDifferentRow ) p->nullRow = 1; - return rc; -} - /* ** Make sure the cursor p is ready to read or write the row to which it ** was last positioned. Return an error code if an OOM fault or I/O error ** prevents us from positioning the cursor to its correct position. ** @@ -65937,14 +64940,33 @@ ** If the cursor is already pointing to the correct row and that row has ** not been deleted out from under the cursor, then this routine is a no-op. */ SQLITE_PRIVATE int sqlite3VdbeCursorMoveto(VdbeCursor *p){ if( p->deferredMoveto ){ - return handleDeferredMoveto(p); - } - if( sqlite3BtreeCursorHasMoved(p->pCursor) ){ - return handleMovedCursor(p); + int res, rc; +#ifdef SQLITE_TEST + extern int sqlite3_search_count; +#endif + assert( p->isTable ); + rc = sqlite3BtreeMovetoUnpacked(p->pCursor, 0, p->movetoTarget, 0, &res); + if( rc ) return rc; + p->lastRowid = p->movetoTarget; + if( res!=0 ) return SQLITE_CORRUPT_BKPT; + p->rowidIsValid = 1; +#ifdef SQLITE_TEST + sqlite3_search_count++; +#endif + p->deferredMoveto = 0; + p->cacheStatus = CACHE_STALE; + }else if( p->pCursor ){ + int hasMoved; + int rc = sqlite3BtreeCursorHasMoved(p->pCursor, &hasMoved); + if( rc ) return rc; + if( hasMoved ){ + p->cacheStatus = CACHE_STALE; + if( hasMoved==2 ) p->nullRow = 1; + } } return SQLITE_OK; } /* @@ -66116,22 +65138,21 @@ /* Integer and Real */ if( serial_type<=7 && serial_type>0 ){ u64 v; u32 i; if( serial_type==7 ){ - assert( sizeof(v)==sizeof(pMem->u.r) ); - memcpy(&v, &pMem->u.r, sizeof(v)); + assert( sizeof(v)==sizeof(pMem->r) ); + memcpy(&v, &pMem->r, sizeof(v)); swapMixedEndianFloat(v); }else{ v = pMem->u.i; } len = i = sqlite3VdbeSerialTypeLen(serial_type); - assert( i>0 ); - do{ - buf[--i] = (u8)(v&0xFF); + while( i-- ){ + buf[i] = (u8)(v&0xFF); v >>= 8; - }while( i ); + } return len; } /* String or blob */ if( serial_type>=12 ){ @@ -66151,58 +65172,22 @@ */ #define ONE_BYTE_INT(x) ((i8)(x)[0]) #define TWO_BYTE_INT(x) (256*(i8)((x)[0])|(x)[1]) #define THREE_BYTE_INT(x) (65536*(i8)((x)[0])|((x)[1]<<8)|(x)[2]) #define FOUR_BYTE_UINT(x) (((u32)(x)[0]<<24)|((x)[1]<<16)|((x)[2]<<8)|(x)[3]) -#define FOUR_BYTE_INT(x) (16777216*(i8)((x)[0])|((x)[1]<<16)|((x)[2]<<8)|(x)[3]) /* ** Deserialize the data blob pointed to by buf as serial type serial_type ** and store the result in pMem. Return the number of bytes read. -** -** This function is implemented as two separate routines for performance. -** The few cases that require local variables are broken out into a separate -** routine so that in most cases the overhead of moving the stack pointer -** is avoided. -*/ -static u32 SQLITE_NOINLINE serialGet( - const unsigned char *buf, /* Buffer to deserialize from */ - u32 serial_type, /* Serial type to deserialize */ - Mem *pMem /* Memory cell to write value into */ -){ - u64 x = FOUR_BYTE_UINT(buf); - u32 y = FOUR_BYTE_UINT(buf+4); - x = (x<<32) + y; - if( serial_type==6 ){ - pMem->u.i = *(i64*)&x; - pMem->flags = MEM_Int; - testcase( pMem->u.i<0 ); - }else{ -#if !defined(NDEBUG) && !defined(SQLITE_OMIT_FLOATING_POINT) - /* Verify that integers and floating point values use the same - ** byte order. Or, that if SQLITE_MIXED_ENDIAN_64BIT_FLOAT is - ** defined that 64-bit floating point values really are mixed - ** endian. - */ - static const u64 t1 = ((u64)0x3ff00000)<<32; - static const double r1 = 1.0; - u64 t2 = t1; - swapMixedEndianFloat(t2); - assert( sizeof(r1)==sizeof(t2) && memcmp(&r1, &t2, sizeof(r1))==0 ); -#endif - assert( sizeof(x)==8 && sizeof(pMem->u.r)==8 ); - swapMixedEndianFloat(x); - memcpy(&pMem->u.r, &x, sizeof(x)); - pMem->flags = sqlite3IsNaN(pMem->u.r) ? MEM_Null : MEM_Real; - } - return 8; -} +*/ SQLITE_PRIVATE u32 sqlite3VdbeSerialGet( const unsigned char *buf, /* Buffer to deserialize from */ u32 serial_type, /* Serial type to deserialize */ Mem *pMem /* Memory cell to write value into */ ){ + u64 x; + u32 y; switch( serial_type ){ case 10: /* Reserved for future use */ case 11: /* Reserved for future use */ case 0: { /* NULL */ pMem->flags = MEM_Null; @@ -66225,11 +65210,12 @@ pMem->flags = MEM_Int; testcase( pMem->u.i<0 ); return 3; } case 4: { /* 4-byte signed integer */ - pMem->u.i = FOUR_BYTE_INT(buf); + y = FOUR_BYTE_UINT(buf); + pMem->u.i = (i64)*(int*)&y; pMem->flags = MEM_Int; testcase( pMem->u.i<0 ); return 4; } case 5: { /* 6-byte signed integer */ @@ -66238,30 +65224,56 @@ testcase( pMem->u.i<0 ); return 6; } case 6: /* 8-byte signed integer */ case 7: { /* IEEE floating point */ - /* These use local variables, so do them in a separate routine - ** to avoid having to move the frame pointer in the common case */ - return serialGet(buf,serial_type,pMem); +#if !defined(NDEBUG) && !defined(SQLITE_OMIT_FLOATING_POINT) + /* Verify that integers and floating point values use the same + ** byte order. Or, that if SQLITE_MIXED_ENDIAN_64BIT_FLOAT is + ** defined that 64-bit floating point values really are mixed + ** endian. + */ + static const u64 t1 = ((u64)0x3ff00000)<<32; + static const double r1 = 1.0; + u64 t2 = t1; + swapMixedEndianFloat(t2); + assert( sizeof(r1)==sizeof(t2) && memcmp(&r1, &t2, sizeof(r1))==0 ); +#endif + x = FOUR_BYTE_UINT(buf); + y = FOUR_BYTE_UINT(buf+4); + x = (x<<32) | y; + if( serial_type==6 ){ + pMem->u.i = *(i64*)&x; + pMem->flags = MEM_Int; + testcase( pMem->u.i<0 ); + }else{ + assert( sizeof(x)==8 && sizeof(pMem->r)==8 ); + swapMixedEndianFloat(x); + memcpy(&pMem->r, &x, sizeof(x)); + pMem->flags = sqlite3IsNaN(pMem->r) ? MEM_Null : MEM_Real; + } + return 8; } case 8: /* Integer 0 */ case 9: { /* Integer 1 */ pMem->u.i = serial_type-8; pMem->flags = MEM_Int; return 0; } default: { static const u16 aFlag[] = { MEM_Blob|MEM_Ephem, MEM_Str|MEM_Ephem }; + u32 len = (serial_type-12)/2; pMem->z = (char *)buf; - pMem->n = (serial_type-12)/2; + pMem->n = len; + pMem->xDel = 0; pMem->flags = aFlag[serial_type&1]; - return pMem->n; + return len; } } return 0; } + /* ** This routine is used to allocate sufficient space for an UnpackedRecord ** structure large enough to be used with sqlite3VdbeRecordUnpack() if ** the first argument is a pointer to KeyInfo structure pKeyInfo. ** @@ -66327,21 +65339,21 @@ p->default_rc = 0; assert( EIGHT_BYTE_ALIGNMENT(pMem) ); idx = getVarint32(aKey, szHdr); d = szHdr; u = 0; - while( idx nField && d<=nKey ){ u32 serial_type; idx += getVarint32(&aKey[idx], serial_type); pMem->enc = pKeyInfo->enc; pMem->db = pKeyInfo->db; /* pMem->flags = 0; // sqlite3VdbeSerialGet() will set this for us */ - pMem->szMalloc = 0; + pMem->zMalloc = 0; d += sqlite3VdbeSerialGet(&aKey[d], serial_type, pMem); pMem++; - if( (++u)>=p->nField ) break; + u++; } assert( u<=pKeyInfo->nField + 1 ); p->nField = u; } @@ -66351,18 +65363,14 @@ ** as the sqlite3VdbeRecordCompare() routine. Unlike VdbeRecordCompare(), ** this function deserializes and compares values using the ** sqlite3VdbeSerialGet() and sqlite3MemCompare() functions. It is used ** in assert() statements to ensure that the optimized code in ** sqlite3VdbeRecordCompare() returns results with these two primitives. -** -** Return true if the result of comparison is equivalent to desiredResult. -** Return false if there is a disagreement. */ static int vdbeRecordCompareDebug( int nKey1, const void *pKey1, /* Left key */ - const UnpackedRecord *pPKey2, /* Right key */ - int desiredResult /* Correct answer */ + const UnpackedRecord *pPKey2 /* Right key */ ){ u32 d1; /* Offset into aKey[] of next data element */ u32 idx1; /* Offset into aKey[] of next header element */ u32 szHdr1; /* Number of bytes in header */ int i = 0; @@ -66370,15 +65378,14 @@ const unsigned char *aKey1 = (const unsigned char *)pKey1; KeyInfo *pKeyInfo; Mem mem1; pKeyInfo = pPKey2->pKeyInfo; - if( pKeyInfo->db==0 ) return 1; mem1.enc = pKeyInfo->enc; mem1.db = pKeyInfo->db; /* mem1.flags = 0; // Will be initialized by sqlite3VdbeSerialGet() */ - VVA_ONLY( mem1.szMalloc = 0; ) /* Only needed by assert() statements */ + VVA_ONLY( mem1.zMalloc = 0; ) /* Only needed by assert() statements */ /* Compilers may complain that mem1.u.i is potentially uninitialized. ** We could initialize it, as shown here, to silence those complaints. ** But in fact, mem1.u.i will never actually be used uninitialized, and doing ** the unnecessary initialization has a measurable negative performance @@ -66417,37 +65424,29 @@ /* Do the comparison */ rc = sqlite3MemCompare(&mem1, &pPKey2->aMem[i], pKeyInfo->aColl[i]); if( rc!=0 ){ - assert( mem1.szMalloc==0 ); /* See comment below */ + assert( mem1.zMalloc==0 ); /* See comment below */ if( pKeyInfo->aSortOrder[i] ){ rc = -rc; /* Invert the result for DESC sort order. */ } - goto debugCompareEnd; + return rc; } i++; }while( idx1 nField ); /* No memory allocation is ever used on mem1. Prove this using ** the following assert(). If the assert() fails, it indicates a ** memory leak and a need to call sqlite3VdbeMemRelease(&mem1). */ - assert( mem1.szMalloc==0 ); + assert( mem1.zMalloc==0 ); /* rc==0 here means that one of the keys ran out of fields and - ** all the fields up to that point were equal. Return the default_rc + ** all the fields up to that point were equal. Return the the default_rc ** value. */ - rc = pPKey2->default_rc; - -debugCompareEnd: - if( desiredResult==0 && rc==0 ) return 1; - if( desiredResult<0 && rc<0 ) return 1; - if( desiredResult>0 && rc>0 ) return 1; - if( CORRUPT_DB ) return 1; - if( pKeyInfo->db->mallocFailed ) return 1; - return 0; + return pPKey2->default_rc; } #endif /* ** Both *pMem1 and *pMem2 contain string values. Compare the two values @@ -66456,12 +65455,11 @@ ** *pMem2, respectively. Similar in spirit to "rc = (*pMem1) - (*pMem2);". */ static int vdbeCompareMemString( const Mem *pMem1, const Mem *pMem2, - const CollSeq *pColl, - u8 *prcErr /* If an OOM occurs, set to SQLITE_NOMEM */ + const CollSeq *pColl ){ if( pMem1->enc==pColl->enc ){ /* The strings are already in the correct encoding. Call the ** comparison function directly */ return pColl->xCmp(pColl->pUser,pMem1->n,pMem1->z,pMem2->n,pMem2->z); @@ -66469,37 +65467,24 @@ int rc; const void *v1, *v2; int n1, n2; Mem c1; Mem c2; - sqlite3VdbeMemInit(&c1, pMem1->db, MEM_Null); - sqlite3VdbeMemInit(&c2, pMem1->db, MEM_Null); + memset(&c1, 0, sizeof(c1)); + memset(&c2, 0, sizeof(c2)); sqlite3VdbeMemShallowCopy(&c1, pMem1, MEM_Ephem); sqlite3VdbeMemShallowCopy(&c2, pMem2, MEM_Ephem); v1 = sqlite3ValueText((sqlite3_value*)&c1, pColl->enc); n1 = v1==0 ? 0 : c1.n; v2 = sqlite3ValueText((sqlite3_value*)&c2, pColl->enc); n2 = v2==0 ? 0 : c2.n; rc = pColl->xCmp(pColl->pUser, n1, v1, n2, v2); sqlite3VdbeMemRelease(&c1); sqlite3VdbeMemRelease(&c2); - if( (v1==0 || v2==0) && prcErr ) *prcErr = SQLITE_NOMEM; return rc; } } - -/* -** Compare two blobs. Return negative, zero, or positive if the first -** is less than, equal to, or greater than the second, respectively. -** If one blob is a prefix of the other, then the shorter is the lessor. -*/ -static SQLITE_NOINLINE int sqlite3BlobCompare(const Mem *pB1, const Mem *pB2){ - int c = memcmp(pB1->z, pB2->z, pB1->n>pB2->n ? pB2->n : pB1->n); - if( c ) return c; - return pB1->n - pB2->n; -} - /* ** Compare the values contained by the two memory cells, returning ** negative, zero or positive if pMem1 is less than, equal to, or greater ** than pMem2. Sorting order is NULL's first, followed by numbers (integers @@ -66507,10 +65492,11 @@ ** sequence pColl and finally blob's ordered by memcmp(). ** ** Two NULL values are considered equal by this function. */ SQLITE_PRIVATE int sqlite3MemCompare(const Mem *pMem1, const Mem *pMem2, const CollSeq *pColl){ + int rc; int f1, f2; int combined_flags; f1 = pMem1->flags; f2 = pMem2->flags; @@ -66534,18 +65520,18 @@ if( pMem1->u.i < pMem2->u.i ) return -1; if( pMem1->u.i > pMem2->u.i ) return 1; return 0; } if( (f1&MEM_Real)!=0 ){ - r1 = pMem1->u.r; + r1 = pMem1->r; }else if( (f1&MEM_Int)!=0 ){ r1 = (double)pMem1->u.i; }else{ return 1; } if( (f2&MEM_Real)!=0 ){ - r2 = pMem2->u.r; + r2 = pMem2->r; }else if( (f2&MEM_Int)!=0 ){ r2 = (double)pMem2->u.i; }else{ return -1; } @@ -66574,18 +65560,22 @@ ** compiled (this was not always the case). */ assert( !pColl || pColl->xCmp ); if( pColl ){ - return vdbeCompareMemString(pMem1, pMem2, pColl, 0); + return vdbeCompareMemString(pMem1, pMem2, pColl); } /* If a NULL pointer was passed as the collate function, fall through ** to the blob case and use memcmp(). */ } /* Both values must be blobs. Compare using memcmp(). */ - return sqlite3BlobCompare(pMem1, pMem2); + rc = memcmp(pMem1->z, pMem2->z, (pMem1->n>pMem2->n)?pMem2->n:pMem1->n); + if( rc==0 ){ + rc = pMem1->n - pMem2->n; + } + return rc; } /* ** The first argument passed to this function is a serial-type that @@ -66631,11 +65621,11 @@ /* ** This function compares the two table rows or index records ** specified by {nKey1, pKey1} and pPKey2. It returns a negative, zero ** or positive integer if key1 is less than, equal to or ** greater than key2. The {nKey1, pKey1} key must be a blob -** created by the OP_MakeRecord opcode of the VDBE. The pPKey2 +** created by th OP_MakeRecord opcode of the VDBE. The pPKey2 ** key must be a parsed key such as obtained from ** sqlite3VdbeParseRecord. ** ** If argument bSkip is non-zero, it is assumed that the caller has already ** determined that the first fields of the keys are equal. @@ -66642,16 +65632,14 @@ ** ** Key1 and Key2 do not have to contain the same number of fields. If all ** fields that appear in both keys are equal, then pPKey2->default_rc is ** returned. ** -** If database corruption is discovered, set pPKey2->errCode to -** SQLITE_CORRUPT and return 0. If an OOM error is encountered, -** pPKey2->errCode is set to SQLITE_NOMEM and, if it is not NULL, the -** malloc-failed flag set on database handle (pPKey2->pKeyInfo->db). +** If database corruption is discovered, set pPKey2->isCorrupt to non-zero +** and return 0. */ -static int vdbeRecordCompareWithSkip( +SQLITE_PRIVATE int sqlite3VdbeRecordCompare( int nKey1, const void *pKey1, /* Left key */ UnpackedRecord *pPKey2, /* Right key */ int bSkip /* If true, skip the first field */ ){ u32 d1; /* Offset into aKey[] of next data element */ @@ -66676,17 +65664,17 @@ pRhs++; }else{ idx1 = getVarint32(aKey1, szHdr1); d1 = szHdr1; if( d1>(unsigned)nKey1 ){ - pPKey2->errCode = (u8)SQLITE_CORRUPT_BKPT; + pPKey2->isCorrupt = (u8)SQLITE_CORRUPT_BKPT; return 0; /* Corruption */ } i = 0; } - VVA_ONLY( mem1.szMalloc = 0; ) /* Only needed by assert() statements */ + VVA_ONLY( mem1.zMalloc = 0; ) /* Only needed by assert() statements */ assert( pPKey2->pKeyInfo->nField+pPKey2->pKeyInfo->nXField>=pPKey2->nField || CORRUPT_DB ); assert( pPKey2->pKeyInfo->aSortOrder!=0 ); assert( pPKey2->pKeyInfo->nField>0 ); assert( idx1<=szHdr1 || CORRUPT_DB ); @@ -66702,13 +65690,13 @@ }else if( serial_type==0 ){ rc = -1; }else if( serial_type==7 ){ double rhs = (double)pRhs->u.i; sqlite3VdbeSerialGet(&aKey1[d1], serial_type, &mem1); - if( mem1.u.r rhs ){ + }else if( mem1.r>rhs ){ rc = +1; } }else{ i64 lhs = vdbeRecordDecodeInt(serial_type, &aKey1[d1]); i64 rhs = pRhs->u.i; @@ -66726,15 +65714,15 @@ if( serial_type>=12 ){ rc = +1; }else if( serial_type==0 ){ rc = -1; }else{ - double rhs = pRhs->u.r; + double rhs = pRhs->r; double lhs; sqlite3VdbeSerialGet(&aKey1[d1], serial_type, &mem1); if( serial_type==7 ){ - lhs = mem1.u.r; + lhs = mem1.r; }else{ lhs = (double)mem1.u.i; } if( lhs (unsigned)nKey1 ){ - pPKey2->errCode = (u8)SQLITE_CORRUPT_BKPT; + pPKey2->isCorrupt = (u8)SQLITE_CORRUPT_BKPT; return 0; /* Corruption */ }else if( pKeyInfo->aColl[i] ){ mem1.enc = pKeyInfo->enc; mem1.db = pKeyInfo->db; mem1.flags = MEM_Str; mem1.z = (char*)&aKey1[d1]; - rc = vdbeCompareMemString( - &mem1, pRhs, pKeyInfo->aColl[i], &pPKey2->errCode - ); + rc = vdbeCompareMemString(&mem1, pRhs, pKeyInfo->aColl[i]); }else{ int nCmp = MIN(mem1.n, pRhs->n); rc = memcmp(&aKey1[d1], pRhs->z, nCmp); if( rc==0 ) rc = mem1.n - pRhs->n; } @@ -66784,11 +65770,11 @@ }else{ int nStr = (serial_type - 12) / 2; testcase( (d1+nStr)==(unsigned)nKey1 ); testcase( (d1+nStr+1)==(unsigned)nKey1 ); if( (d1+nStr) > (unsigned)nKey1 ){ - pPKey2->errCode = (u8)SQLITE_CORRUPT_BKPT; + pPKey2->isCorrupt = (u8)SQLITE_CORRUPT_BKPT; return 0; /* Corruption */ }else{ int nCmp = MIN(nStr, pRhs->n); rc = memcmp(&aKey1[d1], pRhs->z, nCmp); if( rc==0 ) rc = nStr - pRhs->n; @@ -66804,12 +65790,16 @@ if( rc!=0 ){ if( pKeyInfo->aSortOrder[i] ){ rc = -rc; } - assert( vdbeRecordCompareDebug(nKey1, pKey1, pPKey2, rc) ); - assert( mem1.szMalloc==0 ); /* See comment below */ + assert( CORRUPT_DB + || (rc<0 && vdbeRecordCompareDebug(nKey1, pKey1, pPKey2)<0) + || (rc>0 && vdbeRecordCompareDebug(nKey1, pKey1, pPKey2)>0) + || pKeyInfo->db->mallocFailed + ); + assert( mem1.zMalloc==0 ); /* See comment below */ return rc; } i++; pRhs++; @@ -66818,28 +65808,21 @@ }while( idx1<(unsigned)szHdr1 && i nField && d1<=(unsigned)nKey1 ); /* No memory allocation is ever used on mem1. Prove this using ** the following assert(). If the assert() fails, it indicates a ** memory leak and a need to call sqlite3VdbeMemRelease(&mem1). */ - assert( mem1.szMalloc==0 ); + assert( mem1.zMalloc==0 ); /* rc==0 here means that one or both of the keys ran out of fields and - ** all the fields up to that point were equal. Return the default_rc + ** all the fields up to that point were equal. Return the the default_rc ** value. */ assert( CORRUPT_DB - || vdbeRecordCompareDebug(nKey1, pKey1, pPKey2, pPKey2->default_rc) + || pPKey2->default_rc==vdbeRecordCompareDebug(nKey1, pKey1, pPKey2) || pKeyInfo->db->mallocFailed ); return pPKey2->default_rc; } -SQLITE_PRIVATE int sqlite3VdbeRecordCompare( - int nKey1, const void *pKey1, /* Left key */ - UnpackedRecord *pPKey2 /* Right key */ -){ - return vdbeRecordCompareWithSkip(nKey1, pKey1, pPKey2, 0); -} - /* ** This function is an optimized version of sqlite3VdbeRecordCompare() ** that (a) the first field of pPKey2 is an integer, and (b) the ** size-of-header varint at the start of (pKey1/nKey1) fits in a single @@ -66848,20 +65831,23 @@ ** To avoid concerns about buffer overreads, this routine is only used ** on schemas where the maximum valid header size is 63 bytes or less. */ static int vdbeRecordCompareInt( int nKey1, const void *pKey1, /* Left key */ - UnpackedRecord *pPKey2 /* Right key */ + UnpackedRecord *pPKey2, /* Right key */ + int bSkip /* Ignored */ ){ const u8 *aKey = &((const u8*)pKey1)[*(const u8*)pKey1 & 0x3F]; int serial_type = ((const u8*)pKey1)[1]; int res; u32 y; u64 x; i64 v = pPKey2->aMem[0].u.i; i64 lhs; + UNUSED_PARAMETER(bSkip); + assert( bSkip==0 ); assert( (*(u8*)pKey1)<=0x3F || CORRUPT_DB ); switch( serial_type ){ case 1: { /* 1-byte signed integer */ lhs = ONE_BYTE_INT(aKey); testcase( lhs<0 ); @@ -66907,31 +65893,35 @@ ** statement (since the range of switch targets now starts at zero and ** is contiguous) but does not cause any duplicate code to be generated ** (as gcc is clever enough to combine the two like cases). Other ** compilers might be similar. */ case 0: case 7: - return sqlite3VdbeRecordCompare(nKey1, pKey1, pPKey2); + return sqlite3VdbeRecordCompare(nKey1, pKey1, pPKey2, 0); default: - return sqlite3VdbeRecordCompare(nKey1, pKey1, pPKey2); + return sqlite3VdbeRecordCompare(nKey1, pKey1, pPKey2, 0); } if( v>lhs ){ res = pPKey2->r1; }else if( v r2; }else if( pPKey2->nField>1 ){ /* The first fields of the two keys are equal. Compare the trailing ** fields. */ - res = vdbeRecordCompareWithSkip(nKey1, pKey1, pPKey2, 1); + res = sqlite3VdbeRecordCompare(nKey1, pKey1, pPKey2, 1); }else{ /* The first fields of the two keys are equal and there are no trailing ** fields. Return pPKey2->default_rc in this case. */ res = pPKey2->default_rc; } - assert( vdbeRecordCompareDebug(nKey1, pKey1, pPKey2, res) ); + assert( (res==0 && vdbeRecordCompareDebug(nKey1, pKey1, pPKey2)==0) + || (res<0 && vdbeRecordCompareDebug(nKey1, pKey1, pPKey2)<0) + || (res>0 && vdbeRecordCompareDebug(nKey1, pKey1, pPKey2)>0) + || CORRUPT_DB + ); return res; } /* ** This function is an optimized version of sqlite3VdbeRecordCompare() @@ -66939,17 +65929,21 @@ ** uses the collation sequence BINARY and (c) that the size-of-header varint ** at the start of (pKey1/nKey1) fits in a single byte. */ static int vdbeRecordCompareString( int nKey1, const void *pKey1, /* Left key */ - UnpackedRecord *pPKey2 /* Right key */ + UnpackedRecord *pPKey2, /* Right key */ + int bSkip ){ const u8 *aKey1 = (const u8*)pKey1; int serial_type; int res; + UNUSED_PARAMETER(bSkip); + assert( bSkip==0 ); getVarint32(&aKey1[1], serial_type); + if( serial_type<12 ){ res = pPKey2->r1; /* (pKey1/nKey1) is a number or a null */ }else if( !(serial_type & 0x01) ){ res = pPKey2->r2; /* (pKey1/nKey1) is a blob */ }else{ @@ -66957,21 +65951,21 @@ int nStr; int szHdr = aKey1[0]; nStr = (serial_type-12) / 2; if( (szHdr + nStr) > nKey1 ){ - pPKey2->errCode = (u8)SQLITE_CORRUPT_BKPT; + pPKey2->isCorrupt = (u8)SQLITE_CORRUPT_BKPT; return 0; /* Corruption */ } nCmp = MIN( pPKey2->aMem[0].n, nStr ); res = memcmp(&aKey1[szHdr], pPKey2->aMem[0].z, nCmp); if( res==0 ){ res = nStr - pPKey2->aMem[0].n; if( res==0 ){ if( pPKey2->nField>1 ){ - res = vdbeRecordCompareWithSkip(nKey1, pKey1, pPKey2, 1); + res = sqlite3VdbeRecordCompare(nKey1, pKey1, pPKey2, 1); }else{ res = pPKey2->default_rc; } }else if( res>0 ){ res = pPKey2->r2; @@ -66983,11 +65977,13 @@ }else{ res = pPKey2->r1; } } - assert( vdbeRecordCompareDebug(nKey1, pKey1, pPKey2, res) + assert( (res==0 && vdbeRecordCompareDebug(nKey1, pKey1, pPKey2)==0) + || (res<0 && vdbeRecordCompareDebug(nKey1, pKey1, pPKey2)<0) + || (res>0 && vdbeRecordCompareDebug(nKey1, pKey1, pPKey2)>0) || CORRUPT_DB || pPKey2->pKeyInfo->db->mallocFailed ); return res; } @@ -67048,10 +66044,12 @@ int rc; u32 szHdr; /* Size of the header */ u32 typeRowid; /* Serial type of the rowid */ u32 lenRowid; /* Size of the rowid */ Mem m, v; + + UNUSED_PARAMETER(db); /* Get the size of the index entry. Only indices entries of less ** than 2GiB are support - anything large must be database corruption. ** Any corruption is detected in sqlite3BtreeParseCellPtr(), though, so ** this code can safely assume that nCellKey is 32-bits @@ -67060,11 +66058,11 @@ VVA_ONLY(rc =) sqlite3BtreeKeySize(pCur, &nCellKey); assert( rc==SQLITE_OK ); /* pCur is always valid so KeySize cannot fail */ assert( (nCellKey & SQLITE_MAX_U32)==(u64)nCellKey ); /* Read in the complete content of the index entry */ - sqlite3VdbeMemInit(&m, db, 0); + memset(&m, 0, sizeof(m)); rc = sqlite3VdbeMemFromBtree(pCur, 0, (u32)nCellKey, 1, &m); if( rc ){ return rc; } @@ -67103,11 +66101,11 @@ return SQLITE_OK; /* Jump here if database corruption is detected after m has been ** allocated. Free the m object and return SQLITE_CORRUPT. */ idx_rowid_corruption: - testcase( m.szMalloc!=0 ); + testcase( m.zMalloc!=0 ); sqlite3VdbeMemRelease(&m); return SQLITE_CORRUPT_BKPT; } /* @@ -67120,11 +66118,10 @@ ** omits the rowid at the end. The rowid at the end of the index entry ** is ignored as well. Hence, this routine only compares the prefixes ** of the keys prior to the final rowid, not the entire key. */ SQLITE_PRIVATE int sqlite3VdbeIdxKeyCompare( - sqlite3 *db, /* Database connection */ VdbeCursor *pC, /* The cursor to compare against */ UnpackedRecord *pUnpacked, /* Unpacked version of key */ int *res /* Write the comparison result here */ ){ i64 nCellKey = 0; @@ -67139,16 +66136,16 @@ ** that btreeParseCellPtr() and sqlite3GetVarint32() are implemented */ if( nCellKey<=0 || nCellKey>0x7fffffff ){ *res = 0; return SQLITE_CORRUPT_BKPT; } - sqlite3VdbeMemInit(&m, db, 0); + memset(&m, 0, sizeof(m)); rc = sqlite3VdbeMemFromBtree(pC->pCursor, 0, (u32)nCellKey, 1, &m); if( rc ){ return rc; } - *res = sqlite3VdbeRecordCompare(m.n, m.z, pUnpacked); + *res = sqlite3VdbeRecordCompare(m.n, m.z, pUnpacked, 0); sqlite3VdbeMemRelease(&m); return SQLITE_OK; } /* @@ -67458,185 +66455,136 @@ /**************************** sqlite3_result_ ******************************* ** The following routines are used by user-defined functions to specify ** the function result. ** -** The setStrOrError() function calls sqlite3VdbeMemSetStr() to store the +** The setStrOrError() funtion calls sqlite3VdbeMemSetStr() to store the ** result as a string or blob but if the string or blob is too large, it ** then sets the error code to SQLITE_TOOBIG -** -** The invokeValueDestructor(P,X) routine invokes destructor function X() -** on value P is not going to be used and need to be destroyed. */ static void setResultStrOrError( sqlite3_context *pCtx, /* Function context */ const char *z, /* String pointer */ int n, /* Bytes in string, or negative */ u8 enc, /* Encoding of z. 0 for BLOBs */ void (*xDel)(void*) /* Destructor function */ ){ - if( sqlite3VdbeMemSetStr(pCtx->pOut, z, n, enc, xDel)==SQLITE_TOOBIG ){ + if( sqlite3VdbeMemSetStr(&pCtx->s, z, n, enc, xDel)==SQLITE_TOOBIG ){ sqlite3_result_error_toobig(pCtx); } } -static int invokeValueDestructor( - const void *p, /* Value to destroy */ - void (*xDel)(void*), /* The destructor */ - sqlite3_context *pCtx /* Set a SQLITE_TOOBIG error if no NULL */ -){ - assert( xDel!=SQLITE_DYNAMIC ); - if( xDel==0 ){ - /* noop */ - }else if( xDel==SQLITE_TRANSIENT ){ - /* noop */ - }else{ - xDel((void*)p); - } - if( pCtx ) sqlite3_result_error_toobig(pCtx); - return SQLITE_TOOBIG; -} SQLITE_API void sqlite3_result_blob( sqlite3_context *pCtx, const void *z, int n, void (*xDel)(void *) ){ assert( n>=0 ); - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); setResultStrOrError(pCtx, z, n, 0, xDel); } -SQLITE_API void sqlite3_result_blob64( - sqlite3_context *pCtx, - const void *z, - sqlite3_uint64 n, - void (*xDel)(void *) -){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); - assert( xDel!=SQLITE_DYNAMIC ); - if( n>0x7fffffff ){ - (void)invokeValueDestructor(z, xDel, pCtx); - }else{ - setResultStrOrError(pCtx, z, (int)n, 0, xDel); - } -} SQLITE_API void sqlite3_result_double(sqlite3_context *pCtx, double rVal){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); - sqlite3VdbeMemSetDouble(pCtx->pOut, rVal); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); + sqlite3VdbeMemSetDouble(&pCtx->s, rVal); } SQLITE_API void sqlite3_result_error(sqlite3_context *pCtx, const char *z, int n){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); pCtx->isError = SQLITE_ERROR; pCtx->fErrorOrAux = 1; - sqlite3VdbeMemSetStr(pCtx->pOut, z, n, SQLITE_UTF8, SQLITE_TRANSIENT); + sqlite3VdbeMemSetStr(&pCtx->s, z, n, SQLITE_UTF8, SQLITE_TRANSIENT); } #ifndef SQLITE_OMIT_UTF16 SQLITE_API void sqlite3_result_error16(sqlite3_context *pCtx, const void *z, int n){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); pCtx->isError = SQLITE_ERROR; pCtx->fErrorOrAux = 1; - sqlite3VdbeMemSetStr(pCtx->pOut, z, n, SQLITE_UTF16NATIVE, SQLITE_TRANSIENT); + sqlite3VdbeMemSetStr(&pCtx->s, z, n, SQLITE_UTF16NATIVE, SQLITE_TRANSIENT); } #endif SQLITE_API void sqlite3_result_int(sqlite3_context *pCtx, int iVal){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); - sqlite3VdbeMemSetInt64(pCtx->pOut, (i64)iVal); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); + sqlite3VdbeMemSetInt64(&pCtx->s, (i64)iVal); } SQLITE_API void sqlite3_result_int64(sqlite3_context *pCtx, i64 iVal){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); - sqlite3VdbeMemSetInt64(pCtx->pOut, iVal); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); + sqlite3VdbeMemSetInt64(&pCtx->s, iVal); } SQLITE_API void sqlite3_result_null(sqlite3_context *pCtx){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); - sqlite3VdbeMemSetNull(pCtx->pOut); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); + sqlite3VdbeMemSetNull(&pCtx->s); } SQLITE_API void sqlite3_result_text( sqlite3_context *pCtx, const char *z, int n, void (*xDel)(void *) ){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); setResultStrOrError(pCtx, z, n, SQLITE_UTF8, xDel); } -SQLITE_API void sqlite3_result_text64( - sqlite3_context *pCtx, - const char *z, - sqlite3_uint64 n, - void (*xDel)(void *), - unsigned char enc -){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); - assert( xDel!=SQLITE_DYNAMIC ); - if( enc==SQLITE_UTF16 ) enc = SQLITE_UTF16NATIVE; - if( n>0x7fffffff ){ - (void)invokeValueDestructor(z, xDel, pCtx); - }else{ - setResultStrOrError(pCtx, z, (int)n, enc, xDel); - } -} #ifndef SQLITE_OMIT_UTF16 SQLITE_API void sqlite3_result_text16( sqlite3_context *pCtx, const void *z, int n, void (*xDel)(void *) ){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); setResultStrOrError(pCtx, z, n, SQLITE_UTF16NATIVE, xDel); } SQLITE_API void sqlite3_result_text16be( sqlite3_context *pCtx, const void *z, int n, void (*xDel)(void *) ){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); setResultStrOrError(pCtx, z, n, SQLITE_UTF16BE, xDel); } SQLITE_API void sqlite3_result_text16le( sqlite3_context *pCtx, const void *z, int n, void (*xDel)(void *) ){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); setResultStrOrError(pCtx, z, n, SQLITE_UTF16LE, xDel); } #endif /* SQLITE_OMIT_UTF16 */ SQLITE_API void sqlite3_result_value(sqlite3_context *pCtx, sqlite3_value *pValue){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); - sqlite3VdbeMemCopy(pCtx->pOut, pValue); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); + sqlite3VdbeMemCopy(&pCtx->s, pValue); } SQLITE_API void sqlite3_result_zeroblob(sqlite3_context *pCtx, int n){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); - sqlite3VdbeMemSetZeroBlob(pCtx->pOut, n); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); + sqlite3VdbeMemSetZeroBlob(&pCtx->s, n); } SQLITE_API void sqlite3_result_error_code(sqlite3_context *pCtx, int errCode){ pCtx->isError = errCode; pCtx->fErrorOrAux = 1; - if( pCtx->pOut->flags & MEM_Null ){ - sqlite3VdbeMemSetStr(pCtx->pOut, sqlite3ErrStr(errCode), -1, + if( pCtx->s.flags & MEM_Null ){ + sqlite3VdbeMemSetStr(&pCtx->s, sqlite3ErrStr(errCode), -1, SQLITE_UTF8, SQLITE_STATIC); } } /* Force an SQLITE_TOOBIG error. */ SQLITE_API void sqlite3_result_error_toobig(sqlite3_context *pCtx){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); pCtx->isError = SQLITE_TOOBIG; pCtx->fErrorOrAux = 1; - sqlite3VdbeMemSetStr(pCtx->pOut, "string or blob too big", -1, + sqlite3VdbeMemSetStr(&pCtx->s, "string or blob too big", -1, SQLITE_UTF8, SQLITE_STATIC); } /* An SQLITE_NOMEM error. */ SQLITE_API void sqlite3_result_error_nomem(sqlite3_context *pCtx){ - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); - sqlite3VdbeMemSetNull(pCtx->pOut); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); + sqlite3VdbeMemSetNull(&pCtx->s); pCtx->isError = SQLITE_NOMEM; pCtx->fErrorOrAux = 1; - pCtx->pOut->db->mallocFailed = 1; + pCtx->s.db->mallocFailed = 1; } /* ** This function is called after a transaction has been committed. It ** invokes callbacks registered with sqlite3_wal_hook() as required. @@ -67808,16 +66756,14 @@ } db = v->db; sqlite3_mutex_enter(db->mutex); v->doingRerun = 0; while( (rc = sqlite3Step(v))==SQLITE_SCHEMA - && cnt++ < SQLITE_MAX_SCHEMA_RETRY ){ - int savedPc = v->pc; - rc2 = rc = sqlite3Reprepare(v); - if( rc!=SQLITE_OK) break; + && cnt++ < SQLITE_MAX_SCHEMA_RETRY + && (rc2 = rc = sqlite3Reprepare(v))==SQLITE_OK ){ sqlite3_reset(pStmt); - if( savedPc>=0 ) v->doingRerun = 1; + v->doingRerun = 1; assert( v->expired==0 ); } if( rc2!=SQLITE_OK ){ /* This case occurs after failing to recompile an sql statement. ** The error message from the SQL compiler has already been loaded @@ -67863,21 +66809,21 @@ ** sqlite3_create_function16() routines that originally registered the ** application defined function. */ SQLITE_API sqlite3 *sqlite3_context_db_handle(sqlite3_context *p){ assert( p && p->pFunc ); - return p->pOut->db; + return p->s.db; } /* ** Return the current time for a statement */ SQLITE_PRIVATE sqlite3_int64 sqlite3StmtCurrentTime(sqlite3_context *p){ Vdbe *v = p->pVdbe; int rc; if( v->iCurrentTime==0 ){ - rc = sqlite3OsCurrentTimeInt64(p->pOut->db->pVfs, &v->iCurrentTime); + rc = sqlite3OsCurrentTimeInt64(p->s.db->pVfs, &v->iCurrentTime); if( rc ) v->iCurrentTime = 0; } return v->iCurrentTime; } @@ -67902,63 +66848,54 @@ sqlite3_result_error(context, zErr, -1); sqlite3_free(zErr); } /* -** Create a new aggregate context for p and return a pointer to -** its pMem->z element. +** Allocate or return the aggregate context for a user function. A new +** context is allocated on the first call. Subsequent calls return the +** same context that was returned on prior calls. */ -static SQLITE_NOINLINE void *createAggContext(sqlite3_context *p, int nByte){ - Mem *pMem = p->pMem; - assert( (pMem->flags & MEM_Agg)==0 ); - if( nByte<=0 ){ - sqlite3VdbeMemSetNull(pMem); - pMem->z = 0; - }else{ - sqlite3VdbeMemClearAndResize(pMem, nByte); - pMem->flags = MEM_Agg; - pMem->u.pDef = p->pFunc; - if( pMem->z ){ - memset(pMem->z, 0, nByte); +SQLITE_API void *sqlite3_aggregate_context(sqlite3_context *p, int nByte){ + Mem *pMem; + assert( p && p->pFunc && p->pFunc->xStep ); + assert( sqlite3_mutex_held(p->s.db->mutex) ); + pMem = p->pMem; + testcase( nByte<0 ); + if( (pMem->flags & MEM_Agg)==0 ){ + if( nByte<=0 ){ + sqlite3VdbeMemReleaseExternal(pMem); + pMem->flags = MEM_Null; + pMem->z = 0; + }else{ + sqlite3VdbeMemGrow(pMem, nByte, 0); + pMem->flags = MEM_Agg; + pMem->u.pDef = p->pFunc; + if( pMem->z ){ + memset(pMem->z, 0, nByte); + } } } return (void*)pMem->z; } /* -** Allocate or return the aggregate context for a user function. A new -** context is allocated on the first call. Subsequent calls return the -** same context that was returned on prior calls. -*/ -SQLITE_API void *sqlite3_aggregate_context(sqlite3_context *p, int nByte){ - assert( p && p->pFunc && p->pFunc->xStep ); - assert( sqlite3_mutex_held(p->pOut->db->mutex) ); - testcase( nByte<0 ); - if( (p->pMem->flags & MEM_Agg)==0 ){ - return createAggContext(p, nByte); - }else{ - return (void*)p->pMem->z; - } -} - -/* -** Return the auxiliary data pointer, if any, for the iArg'th argument to +** Return the auxilary data pointer, if any, for the iArg'th argument to ** the user-function defined by pCtx. */ SQLITE_API void *sqlite3_get_auxdata(sqlite3_context *pCtx, int iArg){ AuxData *pAuxData; - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); for(pAuxData=pCtx->pVdbe->pAuxData; pAuxData; pAuxData=pAuxData->pNext){ if( pAuxData->iOp==pCtx->iOp && pAuxData->iArg==iArg ) break; } return (pAuxData ? pAuxData->pAux : 0); } /* -** Set the auxiliary data pointer and delete function, for the iArg'th +** Set the auxilary data pointer and delete function, for the iArg'th ** argument to the user-function defined by pCtx. Any previous value is ** deleted by calling the delete function specified when it was set. */ SQLITE_API void sqlite3_set_auxdata( sqlite3_context *pCtx, @@ -67967,11 +66904,11 @@ void (*xDelete)(void*) ){ AuxData *pAuxData; Vdbe *pVdbe = pCtx->pVdbe; - assert( sqlite3_mutex_held(pCtx->pOut->db->mutex) ); + assert( sqlite3_mutex_held(pCtx->s.db->mutex) ); if( iArg<0 ) goto failed; for(pAuxData=pVdbe->pAuxData; pAuxData; pAuxData=pAuxData->pNext){ if( pAuxData->iOp==pCtx->iOp && pAuxData->iArg==iArg ) break; } @@ -68000,11 +66937,11 @@ } } #ifndef SQLITE_OMIT_DEPRECATED /* -** Return the number of times the Step function of an aggregate has been +** Return the number of times the Step function of a aggregate has been ** called. ** ** This function is deprecated. Do not use it for new code. It is ** provide only to avoid breaking legacy code. New aggregate function ** implementations should keep their own counts within their aggregate @@ -68049,26 +66986,15 @@ ** __attribute__((aligned(8))) macro. */ static const Mem nullMem #if defined(SQLITE_DEBUG) && defined(__GNUC__) __attribute__((aligned(8))) #endif - = { - /* .u = */ {0}, - /* .flags = */ MEM_Null, - /* .enc = */ 0, - /* .n = */ 0, - /* .z = */ 0, - /* .zMalloc = */ 0, - /* .szMalloc = */ 0, - /* .iPadding1 = */ 0, - /* .db = */ 0, - /* .xDel = */ 0, + = {0, "", (double)0, {0}, 0, MEM_Null, 0, #ifdef SQLITE_DEBUG - /* .pScopyFrom = */ 0, - /* .pFiller = */ 0, + 0, 0, /* pScopyFrom, pFiller */ #endif - }; + 0, 0 }; return &nullMem; } /* ** Check to see if column iCol of the given statement is valid. If @@ -68085,11 +67011,11 @@ sqlite3_mutex_enter(pVm->db->mutex); pOut = &pVm->pResultSet[i]; }else{ if( pVm && ALWAYS(pVm->db) ){ sqlite3_mutex_enter(pVm->db->mutex); - sqlite3Error(pVm->db, SQLITE_RANGE); + sqlite3Error(pVm->db, SQLITE_RANGE, 0); } pOut = (Mem*)columnNullValue(); } return pOut; } @@ -68281,11 +67207,11 @@ #ifdef SQLITE_ENABLE_COLUMN_METADATA /* ** Return the name of the database from which a result column derives. ** NULL is returned if the result column is an expression or constant or -** anything else which is not an unambiguous reference to a database column. +** anything else which is not an unabiguous reference to a database column. */ SQLITE_API const char *sqlite3_column_database_name(sqlite3_stmt *pStmt, int N){ return columnName( pStmt, N, (const void*(*)(Mem*))sqlite3_value_text, COLNAME_DATABASE); } @@ -68297,11 +67223,11 @@ #endif /* SQLITE_OMIT_UTF16 */ /* ** Return the name of the table from which a result column derives. ** NULL is returned if the result column is an expression or constant or -** anything else which is not an unambiguous reference to a database column. +** anything else which is not an unabiguous reference to a database column. */ SQLITE_API const char *sqlite3_column_table_name(sqlite3_stmt *pStmt, int N){ return columnName( pStmt, N, (const void*(*)(Mem*))sqlite3_value_text, COLNAME_TABLE); } @@ -68313,11 +67239,11 @@ #endif /* SQLITE_OMIT_UTF16 */ /* ** Return the name of the table column from which a result column derives. ** NULL is returned if the result column is an expression or constant or -** anything else which is not an unambiguous reference to a database column. +** anything else which is not an unabiguous reference to a database column. */ SQLITE_API const char *sqlite3_column_origin_name(sqlite3_stmt *pStmt, int N){ return columnName( pStmt, N, (const void*(*)(Mem*))sqlite3_value_text, COLNAME_COLUMN); } @@ -68350,26 +67276,26 @@ if( vdbeSafetyNotNull(p) ){ return SQLITE_MISUSE_BKPT; } sqlite3_mutex_enter(p->db->mutex); if( p->magic!=VDBE_MAGIC_RUN || p->pc>=0 ){ - sqlite3Error(p->db, SQLITE_MISUSE); + sqlite3Error(p->db, SQLITE_MISUSE, 0); sqlite3_mutex_leave(p->db->mutex); sqlite3_log(SQLITE_MISUSE, "bind on a busy prepared statement: [%s]", p->zSql); return SQLITE_MISUSE_BKPT; } if( i<1 || i>p->nVar ){ - sqlite3Error(p->db, SQLITE_RANGE); + sqlite3Error(p->db, SQLITE_RANGE, 0); sqlite3_mutex_leave(p->db->mutex); return SQLITE_RANGE; } i--; pVar = &p->aVar[i]; sqlite3VdbeMemRelease(pVar); pVar->flags = MEM_Null; - sqlite3Error(p->db, SQLITE_OK); + sqlite3Error(p->db, SQLITE_OK, 0); /* If the bit corresponding to this variable in Vdbe.expmask is set, then ** binding a new value to this variable invalidates the current query plan. ** ** IMPLEMENTATION-OF: R-48440-37595 If the specific value bound to host @@ -68407,11 +67333,11 @@ pVar = &p->aVar[i-1]; rc = sqlite3VdbeMemSetStr(pVar, zData, nData, encoding, xDel); if( rc==SQLITE_OK && encoding!=0 ){ rc = sqlite3VdbeChangeEncoding(pVar, ENC(p->db)); } - sqlite3Error(p->db, rc); + sqlite3Error(p->db, rc, 0); rc = sqlite3ApiExit(p->db, rc); } sqlite3_mutex_leave(p->db->mutex); }else if( xDel!=SQLITE_STATIC && xDel!=SQLITE_TRANSIENT ){ xDel((void*)zData); @@ -68430,24 +67356,10 @@ int nData, void (*xDel)(void*) ){ return bindText(pStmt, i, zData, nData, xDel, 0); } -SQLITE_API int sqlite3_bind_blob64( - sqlite3_stmt *pStmt, - int i, - const void *zData, - sqlite3_uint64 nData, - void (*xDel)(void*) -){ - assert( xDel!=SQLITE_DYNAMIC ); - if( nData>0x7fffffff ){ - return invokeValueDestructor(zData, xDel, 0); - }else{ - return bindText(pStmt, i, zData, (int)nData, xDel, 0); - } -} SQLITE_API int sqlite3_bind_double(sqlite3_stmt *pStmt, int i, double rValue){ int rc; Vdbe *p = (Vdbe *)pStmt; rc = vdbeUnbind(p, i); if( rc==SQLITE_OK ){ @@ -68485,26 +67397,10 @@ int nData, void (*xDel)(void*) ){ return bindText(pStmt, i, zData, nData, xDel, SQLITE_UTF8); } -SQLITE_API int sqlite3_bind_text64( - sqlite3_stmt *pStmt, - int i, - const char *zData, - sqlite3_uint64 nData, - void (*xDel)(void*), - unsigned char enc -){ - assert( xDel!=SQLITE_DYNAMIC ); - if( nData>0x7fffffff ){ - return invokeValueDestructor(zData, xDel, 0); - }else{ - if( enc==SQLITE_UTF16 ) enc = SQLITE_UTF16NATIVE; - return bindText(pStmt, i, zData, (int)nData, xDel, enc); - } -} #ifndef SQLITE_OMIT_UTF16 SQLITE_API int sqlite3_bind_text16( sqlite3_stmt *pStmt, int i, const void *zData, @@ -68520,11 +67416,11 @@ case SQLITE_INTEGER: { rc = sqlite3_bind_int64(pStmt, i, pValue->u.i); break; } case SQLITE_FLOAT: { - rc = sqlite3_bind_double(pStmt, i, pValue->u.r); + rc = sqlite3_bind_double(pStmt, i, pValue->r); break; } case SQLITE_BLOB: { if( pValue->flags & MEM_Zero ){ rc = sqlite3_bind_zeroblob(pStmt, i, pValue->u.nZero); @@ -68623,11 +67519,11 @@ #ifndef SQLITE_OMIT_DEPRECATED /* ** Deprecated external interface. Internal/core SQLite code ** should call sqlite3TransferBindings. ** -** It is misuse to call this routine with statements from different +** Is is misuse to call this routine with statements from different ** database connections. But as this is a deprecated interface, we ** will not bother to check for that condition. ** ** If the two statements contain a different number of bindings, then ** an SQLITE_ERROR is returned. Nothing else can go wrong, so otherwise @@ -68767,11 +67663,11 @@ ** is eventually freed. ** ** ALGORITHM: Scan the input string looking for host parameters in any of ** these forms: ?, ?N, $A, @A, :A. Take care to avoid text within ** string literals, quoted identifier names, and comments. For text forms, -** the host parameter index is found by scanning the prepared +** the host parameter index is found by scanning the perpared ** statement for the corresponding OP_Variable opcode. Once the host ** parameter index is known, locate the value in p->aVar[]. Then render ** the value as a literal in place of the host parameter name. */ SQLITE_PRIVATE char *sqlite3VdbeExpandSql( @@ -68830,11 +67726,11 @@ if( pVar->flags & MEM_Null ){ sqlite3StrAccumAppend(&out, "NULL", 4); }else if( pVar->flags & MEM_Int ){ sqlite3XPrintf(&out, 0, "%lld", pVar->u.i); }else if( pVar->flags & MEM_Real ){ - sqlite3XPrintf(&out, 0, "%!.15g", pVar->u.r); + sqlite3XPrintf(&out, 0, "%!.15g", pVar->r); }else if( pVar->flags & MEM_Str ){ int nOut; /* Number of bytes of the string text to include in output */ #ifndef SQLITE_OMIT_UTF16 u8 enc = ENC(db); Mem utf8; @@ -68886,10 +67782,125 @@ } return sqlite3StrAccumFinish(&out); } #endif /* #ifndef SQLITE_OMIT_TRACE */ + +/***************************************************************************** +** The following code implements the data-structure explaining logic +** for the Vdbe. +*/ + +#if defined(SQLITE_ENABLE_TREE_EXPLAIN) + +/* +** Allocate a new Explain object +*/ +SQLITE_PRIVATE void sqlite3ExplainBegin(Vdbe *pVdbe){ + if( pVdbe ){ + Explain *p; + sqlite3BeginBenignMalloc(); + p = (Explain *)sqlite3MallocZero( sizeof(Explain) ); + if( p ){ + p->pVdbe = pVdbe; + sqlite3_free(pVdbe->pExplain); + pVdbe->pExplain = p; + sqlite3StrAccumInit(&p->str, p->zBase, sizeof(p->zBase), + SQLITE_MAX_LENGTH); + p->str.useMalloc = 2; + }else{ + sqlite3EndBenignMalloc(); + } + } +} + +/* +** Return true if the Explain ends with a new-line. +*/ +static int endsWithNL(Explain *p){ + return p && p->str.zText && p->str.nChar + && p->str.zText[p->str.nChar-1]=='\n'; +} + +/* +** Append text to the indentation +*/ +SQLITE_PRIVATE void sqlite3ExplainPrintf(Vdbe *pVdbe, const char *zFormat, ...){ + Explain *p; + if( pVdbe && (p = pVdbe->pExplain)!=0 ){ + va_list ap; + if( p->nIndent && endsWithNL(p) ){ + int n = p->nIndent; + if( n>ArraySize(p->aIndent) ) n = ArraySize(p->aIndent); + sqlite3AppendSpace(&p->str, p->aIndent[n-1]); + } + va_start(ap, zFormat); + sqlite3VXPrintf(&p->str, SQLITE_PRINTF_INTERNAL, zFormat, ap); + va_end(ap); + } +} + +/* +** Append a '\n' if there is not already one. +*/ +SQLITE_PRIVATE void sqlite3ExplainNL(Vdbe *pVdbe){ + Explain *p; + if( pVdbe && (p = pVdbe->pExplain)!=0 && !endsWithNL(p) ){ + sqlite3StrAccumAppend(&p->str, "\n", 1); + } +} + +/* +** Push a new indentation level. Subsequent lines will be indented +** so that they begin at the current cursor position. +*/ +SQLITE_PRIVATE void sqlite3ExplainPush(Vdbe *pVdbe){ + Explain *p; + if( pVdbe && (p = pVdbe->pExplain)!=0 ){ + if( p->str.zText && p->nIndent aIndent[p->nIndent-1] ){ + x = p->aIndent[p->nIndent-1]; + } + p->aIndent[p->nIndent] = x; + } + p->nIndent++; + } +} + +/* +** Pop the indentation stack by one level. +*/ +SQLITE_PRIVATE void sqlite3ExplainPop(Vdbe *p){ + if( p && p->pExplain ) p->pExplain->nIndent--; +} + +/* +** Free the indentation structure +*/ +SQLITE_PRIVATE void sqlite3ExplainFinish(Vdbe *pVdbe){ + if( pVdbe && pVdbe->pExplain ){ + sqlite3_free(pVdbe->zExplain); + sqlite3ExplainNL(pVdbe); + pVdbe->zExplain = sqlite3StrAccumFinish(&pVdbe->pExplain->str); + sqlite3_free(pVdbe->pExplain); + pVdbe->pExplain = 0; + sqlite3EndBenignMalloc(); + } +} + +/* +** Return the explanation of a virtual machine. +*/ +SQLITE_PRIVATE const char *sqlite3VdbeExplanation(Vdbe *pVdbe){ + return (pVdbe && pVdbe->zExplain) ? pVdbe->zExplain : 0; +} +#endif /* defined(SQLITE_DEBUG) */ /************** End of vdbetrace.c *******************************************/ /************** Begin file vdbe.c ********************************************/ /* ** 2001 September 15 @@ -69035,11 +68046,11 @@ /* ** Convert the given register into a string if it isn't one ** already. Return non-zero if a malloc() fails. */ #define Stringify(P, enc) \ - if(((P)->flags&(MEM_Str|MEM_Blob))==0 && sqlite3VdbeMemStringify(P,enc,0)) \ + if(((P)->flags&(MEM_Str|MEM_Blob))==0 && sqlite3VdbeMemStringify(P,enc)) \ { goto no_mem; } /* ** An ephemeral string value (signified by the MEM_Ephem flag) contains ** a pointer to a dynamically allocated string where some other entity @@ -69098,11 +68109,11 @@ assert( iCur nCursor ); if( p->apCsr[iCur] ){ sqlite3VdbeFreeCursor(p, p->apCsr[iCur]); p->apCsr[iCur] = 0; } - if( SQLITE_OK==sqlite3VdbeMemClearAndResize(pMem, nByte) ){ + if( SQLITE_OK==sqlite3VdbeMemGrow(pMem, nByte, 0) ){ p->apCsr[iCur] = pCx = (VdbeCursor*)pMem->z; memset(pCx, 0, sizeof(VdbeCursor)); pCx->iDb = iDb; pCx->nField = nField; if( isBtreeCursor ){ @@ -69117,35 +68128,27 @@ /* ** Try to convert a value into a numeric representation if we can ** do so without loss of information. In other words, if the string ** looks like a number, convert it into a number. If it does not ** look like a number, leave it alone. -** -** If the bTryForInt flag is true, then extra effort is made to give -** an integer representation. Strings that look like floating point -** values but which have no fractional component (example: '48.00') -** will have a MEM_Int representation when bTryForInt is true. -** -** If bTryForInt is false, then if the input string contains a decimal -** point or exponential notation, the result is only MEM_Real, even -** if there is an exact integer representation of the quantity. */ -static void applyNumericAffinity(Mem *pRec, int bTryForInt){ +static void applyNumericAffinity(Mem *pRec){ double rValue; i64 iValue; u8 enc = pRec->enc; - assert( (pRec->flags & (MEM_Str|MEM_Int|MEM_Real))==MEM_Str ); + if( (pRec->flags&MEM_Str)==0 ) return; if( sqlite3AtoF(pRec->z, &rValue, pRec->n, enc)==0 ) return; if( 0==sqlite3Atoi64(pRec->z, &iValue, pRec->n, enc) ){ pRec->u.i = iValue; pRec->flags |= MEM_Int; }else{ - pRec->u.r = rValue; + pRec->r = rValue; pRec->flags |= MEM_Real; - if( bTryForInt ) sqlite3VdbeIntegerAffinity(pRec); } } +#define ApplyNumericAffinity(X) \ + if(((X)->flags&(MEM_Real|MEM_Int))==0){applyNumericAffinity(X);} /* ** Processing is determine by the affinity parameter: ** ** SQLITE_AFF_INTEGER: @@ -69166,27 +68169,25 @@ static void applyAffinity( Mem *pRec, /* The value to apply affinity to */ char affinity, /* The affinity to be applied */ u8 enc /* Use this text encoding */ ){ - if( affinity>=SQLITE_AFF_NUMERIC ){ - assert( affinity==SQLITE_AFF_INTEGER || affinity==SQLITE_AFF_REAL - || affinity==SQLITE_AFF_NUMERIC ); - if( (pRec->flags & MEM_Int)==0 ){ - if( (pRec->flags & MEM_Real)==0 ){ - if( pRec->flags & MEM_Str ) applyNumericAffinity(pRec,1); - }else{ - sqlite3VdbeIntegerAffinity(pRec); - } - } - }else if( affinity==SQLITE_AFF_TEXT ){ + if( affinity==SQLITE_AFF_TEXT ){ /* Only attempt the conversion to TEXT if there is an integer or real ** representation (blob and NULL do not get converted) but no string ** representation. */ if( 0==(pRec->flags&MEM_Str) && (pRec->flags&(MEM_Real|MEM_Int)) ){ - sqlite3VdbeMemStringify(pRec, enc, 1); + sqlite3VdbeMemStringify(pRec, enc); + } + pRec->flags &= ~(MEM_Real|MEM_Int); + }else if( affinity!=SQLITE_AFF_NONE ){ + assert( affinity==SQLITE_AFF_INTEGER || affinity==SQLITE_AFF_REAL + || affinity==SQLITE_AFF_NUMERIC ); + ApplyNumericAffinity(pRec); + if( pRec->flags & MEM_Real ){ + sqlite3VdbeIntegerAffinity(pRec); } } } /* @@ -69197,11 +68198,11 @@ */ SQLITE_API int sqlite3_value_numeric_type(sqlite3_value *pVal){ int eType = sqlite3_value_type(pVal); if( eType==SQLITE_TEXT ){ Mem *pMem = (Mem*)pVal; - applyNumericAffinity(pMem, 0); + applyNumericAffinity(pMem); eType = sqlite3_value_type(pVal); } return eType; } @@ -69215,41 +68216,29 @@ u8 enc ){ applyAffinity((Mem *)pVal, affinity, enc); } -/* -** pMem currently only holds a string type (or maybe a BLOB that we can -** interpret as a string if we want to). Compute its corresponding -** numeric type, if has one. Set the pMem->u.r and pMem->u.i fields -** accordingly. -*/ -static u16 SQLITE_NOINLINE computeNumericType(Mem *pMem){ - assert( (pMem->flags & (MEM_Int|MEM_Real))==0 ); - assert( (pMem->flags & (MEM_Str|MEM_Blob))!=0 ); - if( sqlite3AtoF(pMem->z, &pMem->u.r, pMem->n, pMem->enc)==0 ){ - return 0; - } - if( sqlite3Atoi64(pMem->z, &pMem->u.i, pMem->n, pMem->enc)==SQLITE_OK ){ - return MEM_Int; - } - return MEM_Real; -} - /* ** Return the numeric type for pMem, either MEM_Int or MEM_Real or both or ** none. ** ** Unlike applyNumericAffinity(), this routine does not modify pMem->flags. -** But it does set pMem->u.r and pMem->u.i appropriately. +** But it does set pMem->r and pMem->u.i appropriately. */ static u16 numericType(Mem *pMem){ if( pMem->flags & (MEM_Int|MEM_Real) ){ return pMem->flags & (MEM_Int|MEM_Real); } if( pMem->flags & (MEM_Str|MEM_Blob) ){ - return computeNumericType(pMem); + if( sqlite3AtoF(pMem->z, &pMem->r, pMem->n, pMem->enc)==0 ){ + return 0; + } + if( sqlite3Atoi64(pMem->z, &pMem->u.i, pMem->n, pMem->enc)==SQLITE_OK ){ + return MEM_Int; + } + return MEM_Real; } return 0; } #ifdef SQLITE_DEBUG @@ -69348,11 +68337,11 @@ printf(" si:%lld", p->u.i); }else if( p->flags & MEM_Int ){ printf(" i:%lld", p->u.i); #ifndef SQLITE_OMIT_FLOATING_POINT }else if( p->flags & MEM_Real ){ - printf(" r:%g", p->u.r); + printf(" r:%g", p->r); #endif }else if( p->flags & MEM_RowSet ){ printf(" (rowset)"); }else{ char zBuf[200]; @@ -69618,11 +68607,11 @@ if( pOp->opflags & OPFLG_OUT2_PRERELEASE ){ assert( pOp->p2>0 ); assert( pOp->p2<=(p->nMem-p->nCursor) ); pOut = &aMem[pOp->p2]; memAboutToChange(p, pOut); - if( VdbeMemDynamic(pOut) ) sqlite3VdbeMemSetNull(pOut); + VdbeMemRelease(pOut); pOut->flags = MEM_Int; } /* Sanity checking on other operands */ #ifdef SQLITE_DEBUG @@ -69980,11 +68969,11 @@ ** Write that value into register P2. */ case OP_Real: { /* same as TK_FLOAT, out2-prerelease */ pOut->flags = MEM_Real; assert( !sqlite3IsNaN(*pOp->p4.pReal) ); - pOut->u.r = *pOp->p4.pReal; + pOut->r = *pOp->p4.pReal; break; } #endif /* Opcode: String8 * P2 * P4 * @@ -70003,13 +68992,13 @@ #ifndef SQLITE_OMIT_UTF16 if( encoding!=SQLITE_UTF8 ){ rc = sqlite3VdbeMemSetStr(pOut, pOp->p4.z, -1, SQLITE_UTF8, SQLITE_STATIC); if( rc==SQLITE_TOOBIG ) goto too_big; if( SQLITE_OK!=sqlite3VdbeChangeEncoding(pOut, encoding) ) goto no_mem; - assert( pOut->szMalloc>0 && pOut->zMalloc==pOut->z ); + assert( pOut->zMalloc==pOut->z ); assert( VdbeMemDynamic(pOut)==0 ); - pOut->szMalloc = 0; + pOut->zMalloc = 0; pOut->flags |= MEM_Static; if( pOp->p4type==P4_DYNAMIC ){ sqlite3DbFree(db, pOp->p4.z); } pOp->p4type = P4_DYNAMIC; @@ -70057,11 +69046,11 @@ assert( pOp->p3<=(p->nMem-p->nCursor) ); pOut->flags = nullFlag = pOp->p1 ? (MEM_Null|MEM_Cleared) : MEM_Null; while( cnt>0 ){ pOut++; memAboutToChange(p, pOut); - sqlite3VdbeMemSetNull(pOut); + VdbeMemRelease(pOut); pOut->flags = nullFlag; cnt--; } break; } @@ -70125,10 +69114,11 @@ ** left holding a NULL. It is an error for register ranges ** P1..P1+P3-1 and P2..P2+P3-1 to overlap. It is an error ** for P3 to be less than 1. */ case OP_Move: { + char *zMalloc; /* Holding variable for allocated memory */ int n; /* Number of registers left to copy */ int p1; /* Register to copy from */ int p2; /* Register to copy to */ n = pOp->p3; @@ -70142,16 +69132,21 @@ do{ assert( pOut<=&aMem[(p->nMem-p->nCursor)] ); assert( pIn1<=&aMem[(p->nMem-p->nCursor)] ); assert( memIsValid(pIn1) ); memAboutToChange(p, pOut); - sqlite3VdbeMemMove(pOut, pIn1); + VdbeMemRelease(pOut); + zMalloc = pOut->zMalloc; + memcpy(pOut, pIn1, sizeof(Mem)); #ifdef SQLITE_DEBUG if( pOut->pScopyFrom>=&aMem[p1] && pOut->pScopyFrom<&aMem[p1+pOp->p3] ){ pOut->pScopyFrom += p1 - pOp->p2; } #endif + pIn1->flags = MEM_Undefined; + pIn1->xDel = 0; + pIn1->zMalloc = zMalloc; REGISTER_TRACE(p2++, pOut); pIn1++; pOut++; }while( --n ); break; @@ -70452,11 +69447,11 @@ MemSetTypeFlag(pOut, MEM_Int); #else if( sqlite3IsNaN(rB) ){ goto arithmetic_result_is_null; } - pOut->u.r = rB; + pOut->r = rB; MemSetTypeFlag(pOut, MEM_Real); if( ((type1|type2)&MEM_Real)==0 && !bIntint ){ sqlite3VdbeIntegerAffinity(pOut); } #endif @@ -70517,12 +69512,12 @@ n = pOp->p5; apVal = p->apArg; assert( apVal || n==0 ); assert( pOp->p3>0 && pOp->p3<=(p->nMem-p->nCursor) ); - ctx.pOut = &aMem[pOp->p3]; - memAboutToChange(p, ctx.pOut); + pOut = &aMem[pOp->p3]; + memAboutToChange(p, pOut); assert( n==0 || (pOp->p2>0 && pOp->p2+n<=(p->nMem-p->nCursor)+1) ); assert( pOp->p3 p2 || pOp->p3>=pOp->p2+n ); pArg = &aMem[pOp->p2]; for(i=0; i -
-**
- TEXT -**
- BLOB -**
- NUMERIC -**
- INTEGER -**
- REAL -**
nCursor ); pC = p->apCsr[pOp->p1]; assert( pC!=0 ); assert( p2 nField ); - aOffset = pC->aType + pC->nField; + aType = pC->aType; + aOffset = aType + pC->nField; #ifndef SQLITE_OMIT_VIRTUALTABLE assert( pC->pVtabCursor==0 ); /* OP_Column never called on virtual table */ #endif pCrsr = pC->pCursor; assert( pCrsr!=0 || pC->pseudoTableReg>0 ); /* pCrsr NULL on PseudoTables */ @@ -71335,11 +70416,11 @@ goto op_column_error; } } /* Make sure at least the first p2+1 entries of the header have been - ** parsed and valid information is in aOffset[] and pC->aType[]. + ** parsed and valid information is in aOffset[] and aType[]. */ if( pC->nHdrParsed<=p2 ){ /* If there is more header available for parsing in the record, try ** to extract additional fields up through the p2+1-th field */ @@ -71355,11 +70436,11 @@ zData = (u8*)sMem.z; }else{ zData = pC->aRow; } - /* Fill in pC->aType[i] and aOffset[i] values through the p2-th field. */ + /* Fill in aType[i] and aOffset[i] values through the p2-th field. */ i = pC->nHdrParsed; offset = aOffset[i]; zHdr = zData + pC->iHdrOffset; zEndHdr = zData + aOffset[0]; assert( i<=p2 && zHdr aType[i] = t; + aType[i] = t; szField = sqlite3VdbeSerialTypeLen(t); offset += szField; if( offset aType[p2] are + ** reach this point if aOffset[p2], aOffset[p2+1], and aType[p2] are ** all valid. */ assert( p2 nHdrParsed ); assert( rc==SQLITE_OK ); assert( sqlite3VdbeCheckMemInvariants(pDest) ); - if( VdbeMemDynamic(pDest) ) sqlite3VdbeMemSetNull(pDest); - t = pC->aType[p2]; if( pC->szRow>=aOffset[p2+1] ){ /* This is the common case where the desired content fits on the original ** page - where the content is not on an overflow page */ - sqlite3VdbeSerialGet(pC->aRow+aOffset[p2], t, pDest); + VdbeMemRelease(pDest); + sqlite3VdbeSerialGet(pC->aRow+aOffset[p2], aType[p2], pDest); }else{ /* This branch happens only when content is on overflow pages */ + t = aType[p2]; if( ((pOp->p5 & (OPFLAG_LENGTHARG|OPFLAG_TYPEOFARG))!=0 && ((t>=12 && (t&1)==0) || (pOp->p5 & OPFLAG_TYPEOFARG)!=0)) || (len = sqlite3VdbeSerialTypeLen(t))==0 ){ - /* Content is irrelevant for - ** 1. the typeof() function, - ** 2. the length(X) function if X is a blob, and - ** 3. if the content length is zero. - ** So we might as well use bogus content rather than reading - ** content from disk. NULL will work for the value for strings - ** and blobs and whatever is in the payloadSize64 variable - ** will work for everything else. */ - sqlite3VdbeSerialGet(t<=13 ? (u8*)&payloadSize64 : 0, t, pDest); + /* Content is irrelevant for the typeof() function and for + ** the length(X) function if X is a blob. So we might as well use + ** bogus content rather than reading content from disk. NULL works + ** for text and blob and whatever is in the payloadSize64 variable + ** will work for everything else. Content is also irrelevant if + ** the content length is 0. */ + zData = t<=13 ? (u8*)&payloadSize64 : 0; + sMem.zMalloc = 0; }else{ + memset(&sMem, 0, sizeof(sMem)); + sqlite3VdbeMemMove(&sMem, pDest); rc = sqlite3VdbeMemFromBtree(pCrsr, aOffset[p2], len, !pC->isTable, - pDest); + &sMem); if( rc!=SQLITE_OK ){ goto op_column_error; } - sqlite3VdbeSerialGet((const u8*)pDest->z, t, pDest); - pDest->flags &= ~MEM_Ephem; + zData = (u8*)sMem.z; + } + sqlite3VdbeSerialGet(zData, t, pDest); + /* If we dynamically allocated space to hold the data (in the + ** sqlite3VdbeMemFromBtree() call above) then transfer control of that + ** dynamically allocated space over to the pDest structure. + ** This prevents a memory copy. */ + if( sMem.zMalloc ){ + assert( sMem.z==sMem.zMalloc ); + assert( VdbeMemDynamic(pDest)==0 ); + assert( (pDest->flags & (MEM_Blob|MEM_Str))==0 || pDest->z==sMem.z ); + pDest->flags &= ~(MEM_Ephem|MEM_Static); + pDest->flags |= MEM_Term; + pDest->z = sMem.z; + pDest->zMalloc = sMem.zMalloc; } } pDest->enc = encoding; op_column_out: - /* If the column value is an ephemeral string, go ahead and persist - ** that string in case the cursor moves before the column value is - ** used. The following code does the equivalent of Deephemeralize() - ** but does it faster. */ - if( (pDest->flags & MEM_Ephem)!=0 && pDest->z ){ - fx = pDest->flags & (MEM_Str|MEM_Blob); - assert( fx!=0 ); - zData = (const u8*)pDest->z; - len = pDest->n; - if( sqlite3VdbeMemClearAndResize(pDest, len+2) ) goto no_mem; - memcpy(pDest->z, zData, len); - pDest->z[len] = 0; - pDest->z[len+1] = 0; - pDest->flags = fx|MEM_Term; - } + Deephemeralize(pDest); op_column_error: UPDATE_MAX_BLOBSIZE(pDest); REGISTER_TRACE(pOp->p3, pDest); break; } @@ -71544,11 +70625,11 @@ ** ------------------------------------------------------------------------ ** | hdr-size | type 0 | type 1 | ... | type N-1 | data0 | ... | data N-1 | ** ------------------------------------------------------------------------ ** ** Data(0) is taken from register P1. Data(1) comes from register P1+1 - ** and so forth. + ** and so froth. ** ** Each type field is a varint representing the serial type of the ** corresponding data element (see sqlite3VdbeSerialType()). The ** hdr-size field is also a varint which is the offset from the beginning ** of the record to data0. @@ -71620,13 +70701,13 @@ } /* Make sure the output register has a buffer large enough to store ** the new record. The output register (pOp->p3) is not allowed to ** be one of the input registers (because the following call to - ** sqlite3VdbeMemClearAndResize() could clobber the value before it is used). + ** sqlite3VdbeMemGrow() could clobber the value before it is used). */ - if( sqlite3VdbeMemClearAndResize(pOut, (int)nByte) ){ + if( sqlite3VdbeMemGrow(pOut, (int)nByte, 0) ){ goto no_mem; } zNewRecord = (u8 *)pOut->z; /* Write the record */ @@ -71643,10 +70724,11 @@ assert( j==nByte ); assert( pOp->p3>0 && pOp->p3<=(p->nMem-p->nCursor) ); pOut->n = (int)nByte; pOut->flags = MEM_Blob; + pOut->xDel = 0; if( nZero ){ pOut->u.nZero = nZero; pOut->flags |= MEM_Zero; } pOut->enc = SQLITE_UTF8; /* In case the blob is ever converted to text */ @@ -72257,10 +71339,14 @@ pCur->pgnoRoot = p2; rc = sqlite3BtreeCursor(pX, p2, wrFlag, pKeyInfo, pCur->pCursor); pCur->pKeyInfo = pKeyInfo; assert( OPFLAG_BULKCSR==BTREE_BULKLOAD ); sqlite3BtreeCursorHints(pCur->pCursor, (pOp->p5 & OPFLAG_BULKCSR)); + + /* Since it performs no memory allocation or IO, the only value that + ** sqlite3BtreeCursor() may return is SQLITE_OK. */ + assert( rc==SQLITE_OK ); /* Set the VdbeCursor.isTable variable. Previous versions of ** SQLite used to check if the root-page flags were sane at this point ** and report database corruption if they were not, but this check has ** since moved into the btree layer. */ @@ -72341,19 +71427,15 @@ } pCx->isOrdered = (pOp->p5!=BTREE_UNORDERED); break; } -/* Opcode: SorterOpen P1 P2 P3 P4 * +/* Opcode: SorterOpen P1 P2 * P4 * ** ** This opcode works like OP_OpenEphemeral except that it opens ** a transient index that is specifically designed to sort large ** tables using an external merge-sort algorithm. -** -** If argument P3 is non-zero, then it indicates that the sorter may -** assume that a stable sort considering the first P3 fields of each -** key is sufficient to produce the required results. */ case OP_SorterOpen: { VdbeCursor *pCx; assert( pOp->p1>=0 ); @@ -72361,29 +71443,11 @@ pCx = allocateCursor(p, pOp->p1, pOp->p2, -1, 1); if( pCx==0 ) goto no_mem; pCx->pKeyInfo = pOp->p4.pKeyInfo; assert( pCx->pKeyInfo->db==db ); assert( pCx->pKeyInfo->enc==ENC(db) ); - rc = sqlite3VdbeSorterInit(db, pOp->p3, pCx); - break; -} - -/* Opcode: SequenceTest P1 P2 * * * -** Synopsis: if( cursor[P1].ctr++ ) pc = P2 -** -** P1 is a sorter cursor. If the sequence counter is currently zero, jump -** to P2. Regardless of whether or not the jump is taken, increment the -** the sequence value. -*/ -case OP_SequenceTest: { - VdbeCursor *pC; - assert( pOp->p1>=0 && pOp->p1 nCursor ); - pC = p->apCsr[pOp->p1]; - assert( pC->pSorter ); - if( (pC->seqCount++)==0 ){ - pc = pOp->p2 - 1; - } + rc = sqlite3VdbeSorterInit(db, pCx); break; } /* Opcode: OpenPseudo P1 P2 P3 * * ** Synopsis: P3 columns in r[P2] @@ -72526,15 +71590,13 @@ pC->seekOp = pOp->opcode; #endif if( pC->isTable ){ /* The input value in P3 might be of any type: integer, real, string, ** blob, or NULL. But it needs to be an integer before we can do - ** the seek, so convert it. */ + ** the seek, so covert it. */ pIn3 = &aMem[pOp->p3]; - if( (pIn3->flags & (MEM_Int|MEM_Real|MEM_Str))==MEM_Str ){ - applyNumericAffinity(pIn3, 0); - } + ApplyNumericAffinity(pIn3); iKey = sqlite3VdbeIntValue(pIn3); pC->rowidIsValid = 0; /* If the P3 value could not be converted into an integer without ** loss of information, then special processing is required... */ @@ -72551,20 +71613,20 @@ ** is 4.9 and the integer approximation 5: ** ** (x > 4.9) -> (x >= 5) ** (x <= 4.9) -> (x < 5) */ - if( pIn3->u.r<(double)iKey ){ + if( pIn3->r<(double)iKey ){ assert( OP_SeekGE==(OP_SeekGT-1) ); assert( OP_SeekLT==(OP_SeekLE-1) ); assert( (OP_SeekLE & 0x0001)==(OP_SeekGT & 0x0001) ); if( (oc & 0x0001)==(OP_SeekGT & 0x0001) ) oc--; } /* If the approximation iKey is smaller than the actual real search ** term, substitute <= for < and > for >=. */ - else if( pIn3->u.r>(double)iKey ){ + else if( pIn3->r>(double)iKey ){ assert( OP_SeekLE==(OP_SeekLT+1) ); assert( OP_SeekGT==(OP_SeekGE+1) ); assert( (OP_SeekLT & 0x0001)==(OP_SeekGE & 0x0001) ); if( (oc & 0x0001)==(OP_SeekLT & 0x0001) ) oc++; } @@ -72991,18 +72053,29 @@ ** largest possible integer (9223372036854775807) then the database ** engine starts picking positive candidate ROWIDs at random until ** it finds one that is not previously used. */ assert( pOp->p3==0 ); /* We cannot be in random rowid mode if this is ** an AUTOINCREMENT table. */ + /* on the first attempt, simply do one more than previous */ + v = lastRowid; + v &= (MAX_ROWID>>1); /* ensure doesn't go negative */ + v++; /* ensure non-zero */ cnt = 0; - do{ - sqlite3_randomness(sizeof(v), &v); - v &= (MAX_ROWID>>1); v++; /* Ensure that v is greater than zero */ - }while( ((rc = sqlite3BtreeMovetoUnpacked(pC->pCursor, 0, (u64)v, + while( ((rc = sqlite3BtreeMovetoUnpacked(pC->pCursor, 0, (u64)v, 0, &res))==SQLITE_OK) && (res==0) - && (++cnt<100)); + && (++cnt<100)){ + /* collision - try another random rowid */ + sqlite3_randomness(sizeof(v), &v); + if( cnt<5 ){ + /* try "small" random rowids for the initial attempts */ + v &= 0xffffff; + }else{ + v &= (MAX_ROWID>>1); /* ensure doesn't go negative */ + } + v++; /* ensure non-zero */ + } if( rc==SQLITE_OK && res==0 ){ rc = SQLITE_FULL; /* IMP: R-38219-53002 */ goto abort_due_to_error; } assert( v>0 ); /* EV: R-40812-03570 */ @@ -73217,11 +72290,10 @@ pC = p->apCsr[pOp->p1]; assert( isSorter(pC) ); assert( pOp->p4type==P4_INT32 ); pIn3 = &aMem[pOp->p3]; nKeyCol = pOp->p4.i; - res = 0; rc = sqlite3VdbeSorterCompare(pC, pIn3, nKeyCol, &res); VdbeBranchTaken(res!=0,2); if( res ){ pc = pOp->p2-1; } @@ -73311,12 +72383,11 @@ assert( rc==SQLITE_OK ); /* DataSize() cannot fail */ if( n>(u32)db->aLimit[SQLITE_LIMIT_LENGTH] ){ goto too_big; } } - testcase( n==0 ); - if( sqlite3VdbeMemClearAndResize(pOut, MAX(n,32)) ){ + if( sqlite3VdbeMemGrow(pOut, n, 0) ){ goto no_mem; } pOut->n = n; MemSetTypeFlag(pOut, MEM_Blob); if( pC->isTable==0 ){ @@ -73483,11 +72554,11 @@ res = 1; #ifdef SQLITE_DEBUG pC->seekOp = OP_Rewind; #endif if( isSorter(pC) ){ - rc = sqlite3VdbeSorterRewind(pC, &res); + rc = sqlite3VdbeSorterRewind(db, pC, &res); }else{ pCrsr = pC->pCursor; assert( pCrsr ); rc = sqlite3BtreeFirst(pCrsr, &res); pC->deferredMoveto = 0; @@ -73661,11 +72732,11 @@ assert( pCrsr!=0 ); assert( pC->isTable==0 ); rc = ExpandBlob(pIn2); if( rc==SQLITE_OK ){ if( isSorter(pC) ){ - rc = sqlite3VdbeSorterWrite(pC, pIn2); + rc = sqlite3VdbeSorterWrite(db, pC, pIn2); }else{ nKey = pIn2->n; zKey = pIn2->z; rc = sqlite3BtreeInsert(pCrsr, zKey, nKey, "", 0, 0, pOp->p3, ((pOp->p5 & OPFLAG_USESEEKRESULT) ? pC->seekResult : 0) @@ -73822,11 +72893,11 @@ r.aMem = &aMem[pOp->p3]; #ifdef SQLITE_DEBUG { int i; for(i=0; i nAlloc*2); + int nNew = p->nAlloc*2; while( nByte>nNew ) nNew = nNew*2; - aNew = sqlite3Realloc(p->aAlloc, nNew); - if( !aNew ) return SQLITE_NOMEM; + p->aAlloc = sqlite3DbReallocOrFree(db, p->aAlloc, nNew); + if( !p->aAlloc ) return SQLITE_NOMEM; p->nAlloc = nNew; - p->aAlloc = aNew; } /* Copy as much data as is available in the buffer into the start of ** p->aAlloc[]. */ memcpy(p->aAlloc, &p->aBuffer[iBuf], nAvail); @@ -76447,17 +75235,17 @@ nRem = nByte - nAvail; /* The following loop copies up to p->nBuffer bytes per iteration into ** the p->aAlloc[] buffer. */ while( nRem>0 ){ - int rc; /* vdbePmaReadBlob() return code */ + int rc; /* vdbeSorterIterRead() return code */ int nCopy; /* Number of bytes to copy */ u8 *aNext; /* Pointer to buffer to copy data from */ nCopy = nRem; if( nRem>p->nBuffer ) nCopy = p->nBuffer; - rc = vdbePmaReadBlob(p, nCopy, &aNext); + rc = vdbeSorterIterRead(db, p, nCopy, &aNext); if( rc!=SQLITE_OK ) return rc; assert( aNext!=p->aAlloc ); memcpy(&p->aAlloc[nByte - nRem], aNext, nCopy); nRem -= nCopy; } @@ -76470,742 +75258,405 @@ /* ** Read a varint from the stream of data accessed by p. Set *pnOut to ** the value read. */ -static int vdbePmaReadVarint(PmaReader *p, u64 *pnOut){ +static int vdbeSorterIterVarint(sqlite3 *db, VdbeSorterIter *p, u64 *pnOut){ int iBuf; - if( p->aMap ){ - p->iReadOff += sqlite3GetVarint(&p->aMap[p->iReadOff], pnOut); - }else{ - iBuf = p->iReadOff % p->nBuffer; - if( iBuf && (p->nBuffer-iBuf)>=9 ){ - p->iReadOff += sqlite3GetVarint(&p->aBuffer[iBuf], pnOut); - }else{ - u8 aVarint[16], *a; - int i = 0, rc; - do{ - rc = vdbePmaReadBlob(p, 1, &a); - if( rc ) return rc; - aVarint[(i++)&0xf] = a[0]; - }while( (a[0]&0x80)!=0 ); - sqlite3GetVarint(aVarint, pnOut); - } - } - - return SQLITE_OK; -} - -/* -** Attempt to memory map file pFile. If successful, set *pp to point to the -** new mapping and return SQLITE_OK. If the mapping is not attempted -** (because the file is too large or the VFS layer is configured not to use -** mmap), return SQLITE_OK and set *pp to NULL. -** -** Or, if an error occurs, return an SQLite error code. The final value of -** *pp is undefined in this case. -*/ -static int vdbeSorterMapFile(SortSubtask *pTask, SorterFile *pFile, u8 **pp){ - int rc = SQLITE_OK; - if( pFile->iEof<=(i64)(pTask->pSorter->db->nMaxSorterMmap) ){ - sqlite3_file *pFd = pFile->pFd; - if( pFd->pMethods->iVersion>=3 ){ - rc = sqlite3OsFetch(pFd, 0, (int)pFile->iEof, (void**)pp); - testcase( rc!=SQLITE_OK ); - } - } - return rc; -} - -/* -** Attach PmaReader pReadr to file pFile (if it is not already attached to -** that file) and seek it to offset iOff within the file. Return SQLITE_OK -** if successful, or an SQLite error code if an error occurs. -*/ -static int vdbePmaReaderSeek( - SortSubtask *pTask, /* Task context */ - PmaReader *pReadr, /* Reader whose cursor is to be moved */ - SorterFile *pFile, /* Sorter file to read from */ - i64 iOff /* Offset in pFile */ -){ - int rc = SQLITE_OK; - - assert( pReadr->pIncr==0 || pReadr->pIncr->bEof==0 ); - - if( sqlite3FaultSim(201) ) return SQLITE_IOERR_READ; - if( pReadr->aMap ){ - sqlite3OsUnfetch(pReadr->pFd, 0, pReadr->aMap); - pReadr->aMap = 0; - } - pReadr->iReadOff = iOff; - pReadr->iEof = pFile->iEof; - pReadr->pFd = pFile->pFd; - - rc = vdbeSorterMapFile(pTask, pFile, &pReadr->aMap); - if( rc==SQLITE_OK && pReadr->aMap==0 ){ - int pgsz = pTask->pSorter->pgsz; - int iBuf = pReadr->iReadOff % pgsz; - if( pReadr->aBuffer==0 ){ - pReadr->aBuffer = (u8*)sqlite3Malloc(pgsz); - if( pReadr->aBuffer==0 ) rc = SQLITE_NOMEM; - pReadr->nBuffer = pgsz; - } - if( rc==SQLITE_OK && iBuf ){ - int nRead = pgsz - iBuf; - if( (pReadr->iReadOff + nRead) > pReadr->iEof ){ - nRead = (int)(pReadr->iEof - pReadr->iReadOff); + iBuf = p->iReadOff % p->nBuffer; + if( iBuf && (p->nBuffer-iBuf)>=9 ){ + p->iReadOff += sqlite3GetVarint(&p->aBuffer[iBuf], pnOut); + }else{ + u8 aVarint[16], *a; + int i = 0, rc; + do{ + rc = vdbeSorterIterRead(db, p, 1, &a); + if( rc ) return rc; + aVarint[(i++)&0xf] = a[0]; + }while( (a[0]&0x80)!=0 ); + sqlite3GetVarint(aVarint, pnOut); + } + + return SQLITE_OK; +} + + +/* +** Advance iterator pIter to the next key in its PMA. Return SQLITE_OK if +** no error occurs, or an SQLite error code if one does. +*/ +static int vdbeSorterIterNext( + sqlite3 *db, /* Database handle (for sqlite3DbMalloc() ) */ + VdbeSorterIter *pIter /* Iterator to advance */ +){ + int rc; /* Return Code */ + u64 nRec = 0; /* Size of record in bytes */ + + if( pIter->iReadOff>=pIter->iEof ){ + /* This is an EOF condition */ + vdbeSorterIterZero(db, pIter); + return SQLITE_OK; + } + + rc = vdbeSorterIterVarint(db, pIter, &nRec); + if( rc==SQLITE_OK ){ + pIter->nKey = (int)nRec; + rc = vdbeSorterIterRead(db, pIter, (int)nRec, &pIter->aKey); + } + + return rc; +} + +/* +** Initialize iterator pIter to scan through the PMA stored in file pFile +** starting at offset iStart and ending at offset iEof-1. This function +** leaves the iterator pointing to the first key in the PMA (or EOF if the +** PMA is empty). +*/ +static int vdbeSorterIterInit( + sqlite3 *db, /* Database handle */ + const VdbeSorter *pSorter, /* Sorter object */ + i64 iStart, /* Start offset in pFile */ + VdbeSorterIter *pIter, /* Iterator to populate */ + i64 *pnByte /* IN/OUT: Increment this value by PMA size */ +){ + int rc = SQLITE_OK; + int nBuf; + + nBuf = sqlite3BtreeGetPageSize(db->aDb[0].pBt); + + assert( pSorter->iWriteOff>iStart ); + assert( pIter->aAlloc==0 ); + assert( pIter->aBuffer==0 ); + pIter->pFile = pSorter->pTemp1; + pIter->iReadOff = iStart; + pIter->nAlloc = 128; + pIter->aAlloc = (u8 *)sqlite3DbMallocRaw(db, pIter->nAlloc); + pIter->nBuffer = nBuf; + pIter->aBuffer = (u8 *)sqlite3DbMallocRaw(db, nBuf); + + if( !pIter->aBuffer ){ + rc = SQLITE_NOMEM; + }else{ + int iBuf; + + iBuf = iStart % nBuf; + if( iBuf ){ + int nRead = nBuf - iBuf; + if( (iStart + nRead) > pSorter->iWriteOff ){ + nRead = (int)(pSorter->iWriteOff - iStart); } rc = sqlite3OsRead( - pReadr->pFd, &pReadr->aBuffer[iBuf], nRead, pReadr->iReadOff - ); - testcase( rc!=SQLITE_OK ); - } - } - - return rc; -} - -/* -** Advance PmaReader pReadr to the next key in its PMA. Return SQLITE_OK if -** no error occurs, or an SQLite error code if one does. -*/ -static int vdbePmaReaderNext(PmaReader *pReadr){ - int rc = SQLITE_OK; /* Return Code */ - u64 nRec = 0; /* Size of record in bytes */ - - - if( pReadr->iReadOff>=pReadr->iEof ){ - IncrMerger *pIncr = pReadr->pIncr; - int bEof = 1; - if( pIncr ){ - rc = vdbeIncrSwap(pIncr); - if( rc==SQLITE_OK && pIncr->bEof==0 ){ - rc = vdbePmaReaderSeek( - pIncr->pTask, pReadr, &pIncr->aFile[0], pIncr->iStartOff - ); - bEof = 0; - } - } - - if( bEof ){ - /* This is an EOF condition */ - vdbePmaReaderClear(pReadr); - testcase( rc!=SQLITE_OK ); - return rc; - } - } - - if( rc==SQLITE_OK ){ - rc = vdbePmaReadVarint(pReadr, &nRec); - } - if( rc==SQLITE_OK ){ - pReadr->nKey = (int)nRec; - rc = vdbePmaReadBlob(pReadr, (int)nRec, &pReadr->aKey); - testcase( rc!=SQLITE_OK ); - } - - return rc; -} - -/* -** Initialize PmaReader pReadr to scan through the PMA stored in file pFile -** starting at offset iStart and ending at offset iEof-1. This function -** leaves the PmaReader pointing to the first key in the PMA (or EOF if the -** PMA is empty). -** -** If the pnByte parameter is NULL, then it is assumed that the file -** contains a single PMA, and that that PMA omits the initial length varint. -*/ -static int vdbePmaReaderInit( - SortSubtask *pTask, /* Task context */ - SorterFile *pFile, /* Sorter file to read from */ - i64 iStart, /* Start offset in pFile */ - PmaReader *pReadr, /* PmaReader to populate */ - i64 *pnByte /* IN/OUT: Increment this value by PMA size */ -){ - int rc; - - assert( pFile->iEof>iStart ); - assert( pReadr->aAlloc==0 && pReadr->nAlloc==0 ); - assert( pReadr->aBuffer==0 ); - assert( pReadr->aMap==0 ); - - rc = vdbePmaReaderSeek(pTask, pReadr, pFile, iStart); - if( rc==SQLITE_OK ){ - u64 nByte; /* Size of PMA in bytes */ - rc = vdbePmaReadVarint(pReadr, &nByte); - pReadr->iEof = pReadr->iReadOff + nByte; - *pnByte += nByte; - } - - if( rc==SQLITE_OK ){ - rc = vdbePmaReaderNext(pReadr); + pSorter->pTemp1, &pIter->aBuffer[iBuf], nRead, iStart + ); + } + + if( rc==SQLITE_OK ){ + u64 nByte; /* Size of PMA in bytes */ + pIter->iEof = pSorter->iWriteOff; + rc = vdbeSorterIterVarint(db, pIter, &nByte); + pIter->iEof = pIter->iReadOff + nByte; + *pnByte += nByte; + } + } + + if( rc==SQLITE_OK ){ + rc = vdbeSorterIterNext(db, pIter); } return rc; } /* ** Compare key1 (buffer pKey1, size nKey1 bytes) with key2 (buffer pKey2, -** size nKey2 bytes). Use (pTask->pKeyInfo) for the collation sequences -** used by the comparison. Return the result of the comparison. -** -** Before returning, object (pTask->pUnpacked) is populated with the -** unpacked version of key2. Or, if pKey2 is passed a NULL pointer, then it -** is assumed that the (pTask->pUnpacked) structure already contains the -** unpacked key to use as key2. -** -** If an OOM error is encountered, (pTask->pUnpacked->error_rc) is set -** to SQLITE_NOMEM. -*/ -static int vdbeSorterCompare( - SortSubtask *pTask, /* Subtask context (for pKeyInfo) */ +** size nKey2 bytes). Argument pKeyInfo supplies the collation functions +** used by the comparison. If an error occurs, return an SQLite error code. +** Otherwise, return SQLITE_OK and set *pRes to a negative, zero or positive +** value, depending on whether key1 is smaller, equal to or larger than key2. +** +** If the bOmitRowid argument is non-zero, assume both keys end in a rowid +** field. For the purposes of the comparison, ignore it. Also, if bOmitRowid +** is true and key1 contains even a single NULL value, it is considered to +** be less than key2. Even if key2 also contains NULL values. +** +** If pKey2 is passed a NULL pointer, then it is assumed that the pCsr->aSpace +** has been allocated and contains an unpacked record that is used as key2. +*/ +static void vdbeSorterCompare( + const VdbeCursor *pCsr, /* Cursor object (for pKeyInfo) */ + int nKeyCol, /* Num of columns. 0 means "all" */ const void *pKey1, int nKey1, /* Left side of comparison */ - const void *pKey2, int nKey2 /* Right side of comparison */ + const void *pKey2, int nKey2, /* Right side of comparison */ + int *pRes /* OUT: Result of comparison */ ){ - UnpackedRecord *r2 = pTask->pUnpacked; + KeyInfo *pKeyInfo = pCsr->pKeyInfo; + VdbeSorter *pSorter = pCsr->pSorter; + UnpackedRecord *r2 = pSorter->pUnpacked; + int i; + if( pKey2 ){ - sqlite3VdbeRecordUnpack(pTask->pSorter->pKeyInfo, nKey2, pKey2, r2); + sqlite3VdbeRecordUnpack(pKeyInfo, nKey2, pKey2, r2); } - return sqlite3VdbeRecordCompare(nKey1, pKey1, r2); + + if( nKeyCol ){ + r2->nField = nKeyCol; + for(i=0; i nTree && iOut>0 ); + + if( iOut>=(pSorter->nTree/2) ){ + i1 = (iOut - pSorter->nTree/2) * 2; + i2 = i1 + 1; + }else{ + i1 = pSorter->aTree[iOut*2]; + i2 = pSorter->aTree[iOut*2+1]; + } + + p1 = &pSorter->aIter[i1]; + p2 = &pSorter->aIter[i2]; + + if( p1->pFile==0 ){ + iRes = i2; + }else if( p2->pFile==0 ){ + iRes = i1; + }else{ + int res; + assert( pCsr->pSorter->pUnpacked!=0 ); /* allocated in vdbeSorterMerge() */ + vdbeSorterCompare( + pCsr, 0, p1->aKey, p1->nKey, p2->aKey, p2->nKey, &res + ); + if( res<=0 ){ + iRes = i1; + }else{ + iRes = i2; + } + } + + pSorter->aTree[iOut] = iRes; + return SQLITE_OK; } /* ** Initialize the temporary index cursor just opened as a sorter cursor. -** -** Usually, the sorter module uses the value of (pCsr->pKeyInfo->nField) -** to determine the number of fields that should be compared from the -** records being sorted. However, if the value passed as argument nField -** is non-zero and the sorter is able to guarantee a stable sort, nField -** is used instead. This is used when sorting records for a CREATE INDEX -** statement. In this case, keys are always delivered to the sorter in -** order of the primary key, which happens to be make up the final part -** of the records being sorted. So if the sort is stable, there is never -** any reason to compare PK fields and they can be ignored for a small -** performance boost. -** -** The sorter can guarantee a stable sort when running in single-threaded -** mode, but not in multi-threaded mode. -** -** SQLITE_OK is returned if successful, or an SQLite error code otherwise. */ -SQLITE_PRIVATE int sqlite3VdbeSorterInit( - sqlite3 *db, /* Database connection (for malloc()) */ - int nField, /* Number of key fields in each record */ - VdbeCursor *pCsr /* Cursor that holds the new sorter */ -){ +SQLITE_PRIVATE int sqlite3VdbeSorterInit(sqlite3 *db, VdbeCursor *pCsr){ int pgsz; /* Page size of main database */ - int i; /* Used to iterate through aTask[] */ int mxCache; /* Cache size */ VdbeSorter *pSorter; /* The new sorter */ - KeyInfo *pKeyInfo; /* Copy of pCsr->pKeyInfo with db==0 */ - int szKeyInfo; /* Size of pCsr->pKeyInfo in bytes */ - int sz; /* Size of pSorter in bytes */ - int rc = SQLITE_OK; -#if SQLITE_MAX_WORKER_THREADS==0 -# define nWorker 0 -#else - int nWorker; -#endif - - /* Initialize the upper limit on the number of worker threads */ -#if SQLITE_MAX_WORKER_THREADS>0 - if( sqlite3TempInMemory(db) || sqlite3GlobalConfig.bCoreMutex==0 ){ - nWorker = 0; - }else{ - nWorker = db->aLimit[SQLITE_LIMIT_WORKER_THREADS]; - } -#endif - - /* Do not allow the total number of threads (main thread + all workers) - ** to exceed the maximum merge count */ -#if SQLITE_MAX_WORKER_THREADS>=SORTER_MAX_MERGE_COUNT - if( nWorker>=SORTER_MAX_MERGE_COUNT ){ - nWorker = SORTER_MAX_MERGE_COUNT-1; - } -#endif + char *d; /* Dummy */ assert( pCsr->pKeyInfo && pCsr->pBt==0 ); - szKeyInfo = sizeof(KeyInfo) + (pCsr->pKeyInfo->nField-1)*sizeof(CollSeq*); - sz = sizeof(VdbeSorter) + nWorker * sizeof(SortSubtask); - - pSorter = (VdbeSorter*)sqlite3DbMallocZero(db, sz + szKeyInfo); - pCsr->pSorter = pSorter; + pCsr->pSorter = pSorter = sqlite3DbMallocZero(db, sizeof(VdbeSorter)); if( pSorter==0 ){ - rc = SQLITE_NOMEM; - }else{ - pSorter->pKeyInfo = pKeyInfo = (KeyInfo*)((u8*)pSorter + sz); - memcpy(pKeyInfo, pCsr->pKeyInfo, szKeyInfo); - pKeyInfo->db = 0; - if( nField && nWorker==0 ) pKeyInfo->nField = nField; - pSorter->pgsz = pgsz = sqlite3BtreeGetPageSize(db->aDb[0].pBt); - pSorter->nTask = nWorker + 1; - pSorter->bUseThreads = (pSorter->nTask>1); - pSorter->db = db; - for(i=0; i nTask; i++){ - SortSubtask *pTask = &pSorter->aTask[i]; - pTask->pSorter = pSorter; - } - - if( !sqlite3TempInMemory(db) ){ - pSorter->mnPmaSize = SORTER_MIN_WORKING * pgsz; - mxCache = db->aDb[0].pSchema->cache_size; - if( mxCache mxPmaSize = mxCache * pgsz; - - /* If the application has not configure scratch memory using - ** SQLITE_CONFIG_SCRATCH then we assume it is OK to do large memory - ** allocations. If scratch memory has been configured, then assume - ** large memory allocations should be avoided to prevent heap - ** fragmentation. - */ - if( sqlite3GlobalConfig.pScratch==0 ){ - assert( pSorter->iMemory==0 ); - pSorter->nMemory = pgsz; - pSorter->list.aMemory = (u8*)sqlite3Malloc(pgsz); - if( !pSorter->list.aMemory ) rc = SQLITE_NOMEM; - } - } - } - - return rc; -} -#undef nWorker /* Defined at the top of this function */ + return SQLITE_NOMEM; + } + + pSorter->pUnpacked = sqlite3VdbeAllocUnpackedRecord(pCsr->pKeyInfo, 0, 0, &d); + if( pSorter->pUnpacked==0 ) return SQLITE_NOMEM; + assert( pSorter->pUnpacked==(UnpackedRecord *)d ); + + if( !sqlite3TempInMemory(db) ){ + pgsz = sqlite3BtreeGetPageSize(db->aDb[0].pBt); + pSorter->mnPmaSize = SORTER_MIN_WORKING * pgsz; + mxCache = db->aDb[0].pSchema->cache_size; + if( mxCache mxPmaSize = mxCache * pgsz; + } + + return SQLITE_OK; +} /* ** Free the list of sorted records starting at pRecord. */ static void vdbeSorterRecordFree(sqlite3 *db, SorterRecord *pRecord){ SorterRecord *p; SorterRecord *pNext; for(p=pRecord; p; p=pNext){ - pNext = p->u.pNext; + pNext = p->pNext; sqlite3DbFree(db, p); } } -/* -** Free all resources owned by the object indicated by argument pTask. All -** fields of *pTask are zeroed before returning. -*/ -static void vdbeSortSubtaskCleanup(sqlite3 *db, SortSubtask *pTask){ - sqlite3DbFree(db, pTask->pUnpacked); - pTask->pUnpacked = 0; -#if SQLITE_MAX_WORKER_THREADS>0 - /* pTask->list.aMemory can only be non-zero if it was handed memory - ** from the main thread. That only occurs SQLITE_MAX_WORKER_THREADS>0 */ - if( pTask->list.aMemory ){ - sqlite3_free(pTask->list.aMemory); - pTask->list.aMemory = 0; - }else -#endif - { - assert( pTask->list.aMemory==0 ); - vdbeSorterRecordFree(0, pTask->list.pList); - } - pTask->list.pList = 0; - if( pTask->file.pFd ){ - sqlite3OsCloseFree(pTask->file.pFd); - pTask->file.pFd = 0; - pTask->file.iEof = 0; - } - if( pTask->file2.pFd ){ - sqlite3OsCloseFree(pTask->file2.pFd); - pTask->file2.pFd = 0; - pTask->file2.iEof = 0; - } -} - -#ifdef SQLITE_DEBUG_SORTER_THREADS -static void vdbeSorterWorkDebug(SortSubtask *pTask, const char *zEvent){ - i64 t; - int iTask = (pTask - pTask->pSorter->aTask); - sqlite3OsCurrentTimeInt64(pTask->pSorter->db->pVfs, &t); - fprintf(stderr, "%lld:%d %s\n", t, iTask, zEvent); -} -static void vdbeSorterRewindDebug(const char *zEvent){ - i64 t; - sqlite3OsCurrentTimeInt64(sqlite3_vfs_find(0), &t); - fprintf(stderr, "%lld:X %s\n", t, zEvent); -} -static void vdbeSorterPopulateDebug( - SortSubtask *pTask, - const char *zEvent -){ - i64 t; - int iTask = (pTask - pTask->pSorter->aTask); - sqlite3OsCurrentTimeInt64(pTask->pSorter->db->pVfs, &t); - fprintf(stderr, "%lld:bg%d %s\n", t, iTask, zEvent); -} -static void vdbeSorterBlockDebug( - SortSubtask *pTask, - int bBlocked, - const char *zEvent -){ - if( bBlocked ){ - i64 t; - sqlite3OsCurrentTimeInt64(pTask->pSorter->db->pVfs, &t); - fprintf(stderr, "%lld:main %s\n", t, zEvent); - } -} -#else -# define vdbeSorterWorkDebug(x,y) -# define vdbeSorterRewindDebug(y) -# define vdbeSorterPopulateDebug(x,y) -# define vdbeSorterBlockDebug(x,y,z) -#endif - -#if SQLITE_MAX_WORKER_THREADS>0 -/* -** Join thread pTask->thread. -*/ -static int vdbeSorterJoinThread(SortSubtask *pTask){ - int rc = SQLITE_OK; - if( pTask->pThread ){ -#ifdef SQLITE_DEBUG_SORTER_THREADS - int bDone = pTask->bDone; -#endif - void *pRet = SQLITE_INT_TO_PTR(SQLITE_ERROR); - vdbeSorterBlockDebug(pTask, !bDone, "enter"); - (void)sqlite3ThreadJoin(pTask->pThread, &pRet); - vdbeSorterBlockDebug(pTask, !bDone, "exit"); - rc = SQLITE_PTR_TO_INT(pRet); - assert( pTask->bDone==1 ); - pTask->bDone = 0; - pTask->pThread = 0; - } - return rc; -} - -/* -** Launch a background thread to run xTask(pIn). -*/ -static int vdbeSorterCreateThread( - SortSubtask *pTask, /* Thread will use this task object */ - void *(*xTask)(void*), /* Routine to run in a separate thread */ - void *pIn /* Argument passed into xTask() */ -){ - assert( pTask->pThread==0 && pTask->bDone==0 ); - return sqlite3ThreadCreate(&pTask->pThread, xTask, pIn); -} - -/* -** Join all outstanding threads launched by SorterWrite() to create -** level-0 PMAs. -*/ -static int vdbeSorterJoinAll(VdbeSorter *pSorter, int rcin){ - int rc = rcin; - int i; - - /* This function is always called by the main user thread. - ** - ** If this function is being called after SorterRewind() has been called, - ** it is possible that thread pSorter->aTask[pSorter->nTask-1].pThread - ** is currently attempt to join one of the other threads. To avoid a race - ** condition where this thread also attempts to join the same object, join - ** thread pSorter->aTask[pSorter->nTask-1].pThread first. */ - for(i=pSorter->nTask-1; i>=0; i--){ - SortSubtask *pTask = &pSorter->aTask[i]; - int rc2 = vdbeSorterJoinThread(pTask); - if( rc==SQLITE_OK ) rc = rc2; - } - return rc; -} -#else -# define vdbeSorterJoinAll(x,rcin) (rcin) -# define vdbeSorterJoinThread(pTask) SQLITE_OK -#endif - -/* -** Allocate a new MergeEngine object capable of handling up to -** nReader PmaReader inputs. -** -** nReader is automatically rounded up to the next power of two. -** nReader may not exceed SORTER_MAX_MERGE_COUNT even after rounding up. -*/ -static MergeEngine *vdbeMergeEngineNew(int nReader){ - int N = 2; /* Smallest power of two >= nReader */ - int nByte; /* Total bytes of space to allocate */ - MergeEngine *pNew; /* Pointer to allocated object to return */ - - assert( nReader<=SORTER_MAX_MERGE_COUNT ); - - while( N nTree = N; - pNew->pTask = 0; - pNew->aReadr = (PmaReader*)&pNew[1]; - pNew->aTree = (int*)&pNew->aReadr[N]; - } - return pNew; -} - -/* -** Free the MergeEngine object passed as the only argument. -*/ -static void vdbeMergeEngineFree(MergeEngine *pMerger){ - int i; - if( pMerger ){ - for(i=0; i nTree; i++){ - vdbePmaReaderClear(&pMerger->aReadr[i]); - } - } - sqlite3_free(pMerger); -} - -/* -** Free all resources associated with the IncrMerger object indicated by -** the first argument. -*/ -static void vdbeIncrFree(IncrMerger *pIncr){ - if( pIncr ){ -#if SQLITE_MAX_WORKER_THREADS>0 - if( pIncr->bUseThread ){ - vdbeSorterJoinThread(pIncr->pTask); - if( pIncr->aFile[0].pFd ) sqlite3OsCloseFree(pIncr->aFile[0].pFd); - if( pIncr->aFile[1].pFd ) sqlite3OsCloseFree(pIncr->aFile[1].pFd); - } -#endif - vdbeMergeEngineFree(pIncr->pMerger); - sqlite3_free(pIncr); - } -} - /* ** Reset a sorting cursor back to its original empty state. */ SQLITE_PRIVATE void sqlite3VdbeSorterReset(sqlite3 *db, VdbeSorter *pSorter){ - int i; - (void)vdbeSorterJoinAll(pSorter, SQLITE_OK); - assert( pSorter->bUseThreads || pSorter->pReader==0 ); -#if SQLITE_MAX_WORKER_THREADS>0 - if( pSorter->pReader ){ - vdbePmaReaderClear(pSorter->pReader); - sqlite3DbFree(db, pSorter->pReader); - pSorter->pReader = 0; - } -#endif - vdbeMergeEngineFree(pSorter->pMerger); - pSorter->pMerger = 0; - for(i=0; i nTask; i++){ - SortSubtask *pTask = &pSorter->aTask[i]; - vdbeSortSubtaskCleanup(db, pTask); - } - if( pSorter->list.aMemory==0 ){ - vdbeSorterRecordFree(0, pSorter->list.pList); - } - pSorter->list.pList = 0; - pSorter->list.szPMA = 0; - pSorter->bUsePMA = 0; - pSorter->iMemory = 0; - pSorter->mxKeysize = 0; - sqlite3DbFree(db, pSorter->pUnpacked); - pSorter->pUnpacked = 0; -} + if( pSorter->aIter ){ + int i; + for(i=0; i nTree; i++){ + vdbeSorterIterZero(db, &pSorter->aIter[i]); + } + sqlite3DbFree(db, pSorter->aIter); + pSorter->aIter = 0; + } + if( pSorter->pTemp1 ){ + sqlite3OsCloseFree(pSorter->pTemp1); + pSorter->pTemp1 = 0; + } + vdbeSorterRecordFree(db, pSorter->pRecord); + pSorter->pRecord = 0; + pSorter->iWriteOff = 0; + pSorter->iReadOff = 0; + pSorter->nInMemory = 0; + pSorter->nTree = 0; + pSorter->nPMA = 0; + pSorter->aTree = 0; +} + /* ** Free any cursor components allocated by sqlite3VdbeSorterXXX routines. */ SQLITE_PRIVATE void sqlite3VdbeSorterClose(sqlite3 *db, VdbeCursor *pCsr){ VdbeSorter *pSorter = pCsr->pSorter; if( pSorter ){ sqlite3VdbeSorterReset(db, pSorter); - sqlite3_free(pSorter->list.aMemory); + sqlite3DbFree(db, pSorter->pUnpacked); sqlite3DbFree(db, pSorter); pCsr->pSorter = 0; } } -#if SQLITE_MAX_MMAP_SIZE>0 -/* -** The first argument is a file-handle open on a temporary file. The file -** is guaranteed to be nByte bytes or smaller in size. This function -** attempts to extend the file to nByte bytes in size and to ensure that -** the VFS has memory mapped it. -** -** Whether or not the file does end up memory mapped of course depends on -** the specific VFS implementation. -*/ -static void vdbeSorterExtendFile(sqlite3 *db, sqlite3_file *pFd, i64 nByte){ - if( nByte<=(i64)(db->nMaxSorterMmap) && pFd->pMethods->iVersion>=3 ){ - int rc = sqlite3OsTruncate(pFd, nByte); - if( rc==SQLITE_OK ){ - void *p = 0; - sqlite3OsFetch(pFd, 0, (int)nByte, &p); - sqlite3OsUnfetch(pFd, 0, p); - } - } -} -#else -# define vdbeSorterExtendFile(x,y,z) -#endif - /* ** Allocate space for a file-handle and open a temporary file. If successful, -** set *ppFd to point to the malloc'd file-handle and return SQLITE_OK. -** Otherwise, set *ppFd to 0 and return an SQLite error code. +** set *ppFile to point to the malloc'd file-handle and return SQLITE_OK. +** Otherwise, set *ppFile to 0 and return an SQLite error code. */ -static int vdbeSorterOpenTempFile( - sqlite3 *db, /* Database handle doing sort */ - i64 nExtend, /* Attempt to extend file to this size */ - sqlite3_file **ppFd -){ - int rc; - rc = sqlite3OsOpenMalloc(db->pVfs, 0, ppFd, +static int vdbeSorterOpenTempFile(sqlite3 *db, sqlite3_file **ppFile){ + int dummy; + return sqlite3OsOpenMalloc(db->pVfs, 0, ppFile, SQLITE_OPEN_TEMP_JOURNAL | SQLITE_OPEN_READWRITE | SQLITE_OPEN_CREATE | - SQLITE_OPEN_EXCLUSIVE | SQLITE_OPEN_DELETEONCLOSE, &rc - ); - if( rc==SQLITE_OK ){ - i64 max = SQLITE_MAX_MMAP_SIZE; - sqlite3OsFileControlHint(*ppFd, SQLITE_FCNTL_MMAP_SIZE, (void*)&max); - if( nExtend>0 ){ - vdbeSorterExtendFile(db, *ppFd, nExtend); - } - } - return rc; -} - -/* -** If it has not already been allocated, allocate the UnpackedRecord -** structure at pTask->pUnpacked. Return SQLITE_OK if successful (or -** if no allocation was required), or SQLITE_NOMEM otherwise. -*/ -static int vdbeSortAllocUnpacked(SortSubtask *pTask){ - if( pTask->pUnpacked==0 ){ - char *pFree; - pTask->pUnpacked = sqlite3VdbeAllocUnpackedRecord( - pTask->pSorter->pKeyInfo, 0, 0, &pFree - ); - assert( pTask->pUnpacked==(UnpackedRecord*)pFree ); - if( pFree==0 ) return SQLITE_NOMEM; - pTask->pUnpacked->nField = pTask->pSorter->pKeyInfo->nField; - pTask->pUnpacked->errCode = 0; - } - return SQLITE_OK; -} - + SQLITE_OPEN_EXCLUSIVE | SQLITE_OPEN_DELETEONCLOSE, &dummy + ); +} /* ** Merge the two sorted lists p1 and p2 into a single list. ** Set *ppOut to the head of the new list. */ static void vdbeSorterMerge( - SortSubtask *pTask, /* Calling thread context */ + const VdbeCursor *pCsr, /* For pKeyInfo */ SorterRecord *p1, /* First list to merge */ SorterRecord *p2, /* Second list to merge */ SorterRecord **ppOut /* OUT: Head of merged list */ ){ SorterRecord *pFinal = 0; SorterRecord **pp = &pFinal; - void *pVal2 = p2 ? SRVAL(p2) : 0; + void *pVal2 = p2 ? p2->pVal : 0; while( p1 && p2 ){ int res; - res = vdbeSorterCompare(pTask, SRVAL(p1), p1->nVal, pVal2, p2->nVal); + vdbeSorterCompare(pCsr, 0, p1->pVal, p1->nVal, pVal2, p2->nVal, &res); if( res<=0 ){ *pp = p1; - pp = &p1->u.pNext; - p1 = p1->u.pNext; + pp = &p1->pNext; + p1 = p1->pNext; pVal2 = 0; }else{ *pp = p2; - pp = &p2->u.pNext; - p2 = p2->u.pNext; + pp = &p2->pNext; + p2 = p2->pNext; if( p2==0 ) break; - pVal2 = SRVAL(p2); + pVal2 = p2->pVal; } } *pp = p1 ? p1 : p2; *ppOut = pFinal; } /* -** Sort the linked list of records headed at pTask->pList. Return -** SQLITE_OK if successful, or an SQLite error code (i.e. SQLITE_NOMEM) if -** an error occurs. +** Sort the linked list of records headed at pCsr->pRecord. Return SQLITE_OK +** if successful, or an SQLite error code (i.e. SQLITE_NOMEM) if an error +** occurs. */ -static int vdbeSorterSort(SortSubtask *pTask, SorterList *pList){ +static int vdbeSorterSort(const VdbeCursor *pCsr){ int i; SorterRecord **aSlot; SorterRecord *p; - int rc; - - rc = vdbeSortAllocUnpacked(pTask); - if( rc!=SQLITE_OK ) return rc; + VdbeSorter *pSorter = pCsr->pSorter; aSlot = (SorterRecord **)sqlite3MallocZero(64 * sizeof(SorterRecord *)); if( !aSlot ){ return SQLITE_NOMEM; } - p = pList->pList; + p = pSorter->pRecord; while( p ){ - SorterRecord *pNext; - if( pList->aMemory ){ - if( (u8*)p==pList->aMemory ){ - pNext = 0; - }else{ - assert( p->u.iNext nTree && iOut>0 ); - - if( iOut>=(pMerger->nTree/2) ){ - i1 = (iOut - pMerger->nTree/2) * 2; - i2 = i1 + 1; - }else{ - i1 = pMerger->aTree[iOut*2]; - i2 = pMerger->aTree[iOut*2+1]; - } - - p1 = &pMerger->aReadr[i1]; - p2 = &pMerger->aReadr[i2]; - - if( p1->pFd==0 ){ - iRes = i2; - }else if( p2->pFd==0 ){ - iRes = i1; - }else{ - int res; - assert( pMerger->pTask->pUnpacked!=0 ); /* from vdbeSortSubtaskMain() */ - res = vdbeSorterCompare( - pMerger->pTask, p1->aKey, p1->nKey, p2->aKey, p2->nKey - ); - if( res<=0 ){ - iRes = i1; - }else{ - iRes = i2; - } - } - - pMerger->aTree[iOut] = iRes; -} - -/* -** Allowed values for the eMode parameter to vdbeMergeEngineInit() -** and vdbePmaReaderIncrMergeInit(). -** -** Only INCRINIT_NORMAL is valid in single-threaded builds (when -** SQLITE_MAX_WORKER_THREADS==0). The other values are only used -** when there exists one or more separate worker threads. -*/ -#define INCRINIT_NORMAL 0 -#define INCRINIT_TASK 1 -#define INCRINIT_ROOT 2 - -/* Forward reference. -** The vdbeIncrMergeInit() and vdbePmaReaderIncrMergeInit() routines call each -** other (when building a merge tree). -*/ -static int vdbePmaReaderIncrMergeInit(PmaReader *pReadr, int eMode); - -/* -** Initialize the MergeEngine object passed as the second argument. Once this -** function returns, the first key of merged data may be read from the -** MergeEngine object in the usual fashion. -** -** If argument eMode is INCRINIT_ROOT, then it is assumed that any IncrMerge -** objects attached to the PmaReader objects that the merger reads from have -** already been populated, but that they have not yet populated aFile[0] and -** set the PmaReader objects up to read from it. In this case all that is -** required is to call vdbePmaReaderNext() on each PmaReader to point it at -** its first key. -** -** Otherwise, if eMode is any value other than INCRINIT_ROOT, then use -** vdbePmaReaderIncrMergeInit() to initialize each PmaReader that feeds data -** to pMerger. -** -** SQLITE_OK is returned if successful, or an SQLite error code otherwise. -*/ -static int vdbeMergeEngineInit( - SortSubtask *pTask, /* Thread that will run pMerger */ - MergeEngine *pMerger, /* MergeEngine to initialize */ - int eMode /* One of the INCRINIT_XXX constants */ -){ + if( rc==SQLITE_OK && pSorter->mxPmaSize>0 && ( + (pSorter->nInMemory>pSorter->mxPmaSize) + || (pSorter->nInMemory>pSorter->mnPmaSize && sqlite3HeapNearlyFull()) + )){ +#ifdef SQLITE_DEBUG + i64 nExpect = pSorter->iWriteOff + + sqlite3VarintLen(pSorter->nInMemory) + + pSorter->nInMemory; +#endif + rc = vdbeSorterListToPMA(db, pCsr); + pSorter->nInMemory = 0; + assert( rc!=SQLITE_OK || (nExpect==pSorter->iWriteOff) ); + } + + return rc; +} + +/* +** Helper function for sqlite3VdbeSorterRewind(). +*/ +static int vdbeSorterInitMerge( + sqlite3 *db, /* Database handle */ + const VdbeCursor *pCsr, /* Cursor handle for this sorter */ + i64 *pnByte /* Sum of bytes in all opened PMAs */ +){ + VdbeSorter *pSorter = pCsr->pSorter; int rc = SQLITE_OK; /* Return code */ - int i; /* For looping over PmaReader objects */ - int nTree = pMerger->nTree; - - /* eMode is always INCRINIT_NORMAL in single-threaded mode */ - assert( SQLITE_MAX_WORKER_THREADS>0 || eMode==INCRINIT_NORMAL ); - - /* Verify that the MergeEngine is assigned to a single thread */ - assert( pMerger->pTask==0 ); - pMerger->pTask = pTask; - - for(i=0; i aReadr[i]; - rc = vdbePmaReaderInit(pTask, &pTask->file, iOff, pReadr, &nDummy); - iOff = pReadr->iEof; - } - - if( rc!=SQLITE_OK ){ - vdbeMergeEngineFree(pNew); - *ppOut = 0; - } - *piOffset = iOff; - return rc; -} - -/* -** Return the depth of a tree comprising nPMA PMAs, assuming a fanout of -** SORTER_MAX_MERGE_COUNT. The returned value does not include leaf nodes. -** -** i.e. -** -** nPMA<=16 -> TreeDepth() == 0 -** nPMA<=256 -> TreeDepth() == 1 -** nPMA<=65536 -> TreeDepth() == 2 -*/ -static int vdbeSorterTreeDepth(int nPMA){ - int nDepth = 0; - i64 nDiv = SORTER_MAX_MERGE_COUNT; - while( nDiv < (i64)nPMA ){ - nDiv = nDiv * SORTER_MAX_MERGE_COUNT; - nDepth++; - } - return nDepth; -} - -/* -** pRoot is the root of an incremental merge-tree with depth nDepth (according -** to vdbeSorterTreeDepth()). pLeaf is the iSeq'th leaf to be added to the -** tree, counting from zero. This function adds pLeaf to the tree. -** -** If successful, SQLITE_OK is returned. If an error occurs, an SQLite error -** code is returned and pLeaf is freed. -*/ -static int vdbeSorterAddToTree( - SortSubtask *pTask, /* Task context */ - int nDepth, /* Depth of tree according to TreeDepth() */ - int iSeq, /* Sequence number of leaf within tree */ - MergeEngine *pRoot, /* Root of tree */ - MergeEngine *pLeaf /* Leaf to add to tree */ -){ - int rc = SQLITE_OK; - int nDiv = 1; - int i; - MergeEngine *p = pRoot; - IncrMerger *pIncr; - - rc = vdbeIncrMergerNew(pTask, pLeaf, &pIncr); - - for(i=1; i nTask; iTask++){ - SortSubtask *pTask = &pSorter->aTask[iTask]; - assert( pTask->nPMA>0 || SQLITE_MAX_WORKER_THREADS>0 ); - if( SQLITE_MAX_WORKER_THREADS==0 || pTask->nPMA ){ - MergeEngine *pRoot = 0; /* Root node of tree for this task */ - int nDepth = vdbeSorterTreeDepth(pTask->nPMA); - i64 iReadOff = 0; - - if( pTask->nPMA<=SORTER_MAX_MERGE_COUNT ){ - rc = vdbeMergeEngineLevel0(pTask, pTask->nPMA, &iReadOff, &pRoot); - }else{ - int i; - int iSeq = 0; - pRoot = vdbeMergeEngineNew(SORTER_MAX_MERGE_COUNT); - if( pRoot==0 ) rc = SQLITE_NOMEM; - for(i=0; i nPMA && rc==SQLITE_OK; i += SORTER_MAX_MERGE_COUNT){ - MergeEngine *pMerger = 0; /* New level-0 PMA merger */ - int nReader; /* Number of level-0 PMAs to merge */ - - nReader = MIN(pTask->nPMA - i, SORTER_MAX_MERGE_COUNT); - rc = vdbeMergeEngineLevel0(pTask, nReader, &iReadOff, &pMerger); - if( rc==SQLITE_OK ){ - rc = vdbeSorterAddToTree(pTask, nDepth, iSeq++, pRoot, pMerger); - } - } - } - - if( rc==SQLITE_OK ){ -#if SQLITE_MAX_WORKER_THREADS>0 - if( pMain!=0 ){ - rc = vdbeIncrMergerNew(pTask, pRoot, &pMain->aReadr[iTask].pIncr); - }else -#endif - { - assert( pMain==0 ); - pMain = pRoot; - } - }else{ - vdbeMergeEngineFree(pRoot); - } - } - } - - if( rc!=SQLITE_OK ){ - vdbeMergeEngineFree(pMain); - pMain = 0; - } - *ppOut = pMain; - return rc; -} - -/* -** This function is called as part of an sqlite3VdbeSorterRewind() operation -** on a sorter that has written two or more PMAs to temporary files. It sets -** up either VdbeSorter.pMerger (for single threaded sorters) or pReader -** (for multi-threaded sorters) so that it can be used to iterate through -** all records stored in the sorter. -** -** SQLITE_OK is returned if successful, or an SQLite error code otherwise. -*/ -static int vdbeSorterSetupMerge(VdbeSorter *pSorter){ - int rc; /* Return code */ - SortSubtask *pTask0 = &pSorter->aTask[0]; - MergeEngine *pMain = 0; -#if SQLITE_MAX_WORKER_THREADS - sqlite3 *db = pTask0->pSorter->db; -#endif - - rc = vdbeSorterMergeTreeBuild(pSorter, &pMain); - if( rc==SQLITE_OK ){ -#if SQLITE_MAX_WORKER_THREADS - assert( pSorter->bUseThreads==0 || pSorter->nTask>1 ); - if( pSorter->bUseThreads ){ - int iTask; - PmaReader *pReadr; - SortSubtask *pLast = &pSorter->aTask[pSorter->nTask-1]; - rc = vdbeSortAllocUnpacked(pLast); - if( rc==SQLITE_OK ){ - pReadr = (PmaReader*)sqlite3DbMallocZero(db, sizeof(PmaReader)); - pSorter->pReader = pReadr; - if( pReadr==0 ) rc = SQLITE_NOMEM; - } - if( rc==SQLITE_OK ){ - rc = vdbeIncrMergerNew(pLast, pMain, &pReadr->pIncr); - if( rc==SQLITE_OK ){ - vdbeIncrMergerSetThreads(pReadr->pIncr); - for(iTask=0; iTask<(pSorter->nTask-1); iTask++){ - IncrMerger *pIncr; - if( (pIncr = pMain->aReadr[iTask].pIncr) ){ - vdbeIncrMergerSetThreads(pIncr); - assert( pIncr->pTask!=pLast ); - } - } - for(iTask=0; rc==SQLITE_OK && iTask nTask; iTask++){ - PmaReader *p = &pMain->aReadr[iTask]; - assert( p->pIncr==0 || p->pIncr->pTask==&pSorter->aTask[iTask] ); - if( p->pIncr ){ - if( iTask==pSorter->nTask-1 ){ - rc = vdbePmaReaderIncrMergeInit(p, INCRINIT_TASK); - }else{ - rc = vdbePmaReaderBgIncrInit(p); - } - } - } - } - pMain = 0; - } - if( rc==SQLITE_OK ){ - rc = vdbePmaReaderIncrMergeInit(pReadr, INCRINIT_ROOT); - } - }else -#endif - { - rc = vdbeMergeEngineInit(pTask0, pMain, INCRINIT_NORMAL); - pSorter->pMerger = pMain; - pMain = 0; - } - } - - if( rc!=SQLITE_OK ){ - vdbeMergeEngineFree(pMain); - } - return rc; -} - - -/* -** Once the sorter has been populated by calls to sqlite3VdbeSorterWrite, -** this function is called to prepare for iterating through the records -** in sorted order. -*/ -SQLITE_PRIVATE int sqlite3VdbeSorterRewind(const VdbeCursor *pCsr, int *pbEof){ + int i; /* Used to iterator through aIter[] */ + i64 nByte = 0; /* Total bytes in all opened PMAs */ + + /* Initialize the iterators. */ + for(i=0; i aIter = (VdbeSorterIter *)sqlite3DbMallocZero(db, nByte); + if( !pSorter->aIter ) return SQLITE_NOMEM; + pSorter->aTree = (int *)&pSorter->aIter[N]; + pSorter->nTree = N; + + do { + int iNew; /* Index of new, merged, PMA */ + + for(iNew=0; + rc==SQLITE_OK && iNew*SORTER_MAX_MERGE_COUNT nPMA; + iNew++ + ){ + int rc2; /* Return code from fileWriterFinish() */ + FileWriter writer; /* Object used to write to disk */ + i64 nWrite; /* Number of bytes in new PMA */ + + memset(&writer, 0, sizeof(FileWriter)); + + /* If there are SORTER_MAX_MERGE_COUNT or less PMAs in file pTemp1, + ** initialize an iterator for each of them and break out of the loop. + ** These iterators will be incrementally merged as the VDBE layer calls + ** sqlite3VdbeSorterNext(). + ** + ** Otherwise, if pTemp1 contains more than SORTER_MAX_MERGE_COUNT PMAs, + ** initialize interators for SORTER_MAX_MERGE_COUNT of them. These PMAs + ** are merged into a single PMA that is written to file pTemp2. + */ + rc = vdbeSorterInitMerge(db, pCsr, &nWrite); + assert( rc!=SQLITE_OK || pSorter->aIter[ pSorter->aTree[1] ].pFile ); + if( rc!=SQLITE_OK || pSorter->nPMA<=SORTER_MAX_MERGE_COUNT ){ + break; + } + + /* Open the second temp file, if it is not already open. */ + if( pTemp2==0 ){ + assert( iWrite2==0 ); + rc = vdbeSorterOpenTempFile(db, &pTemp2); + } + + if( rc==SQLITE_OK ){ + int bEof = 0; + fileWriterInit(db, pTemp2, &writer, iWrite2); + fileWriterWriteVarint(&writer, nWrite); + while( rc==SQLITE_OK && bEof==0 ){ + VdbeSorterIter *pIter = &pSorter->aIter[ pSorter->aTree[1] ]; + assert( pIter->pFile ); + + fileWriterWriteVarint(&writer, pIter->nKey); + fileWriterWrite(&writer, pIter->aKey, pIter->nKey); + rc = sqlite3VdbeSorterNext(db, pCsr, &bEof); + } + rc2 = fileWriterFinish(db, &writer, &iWrite2); + if( rc==SQLITE_OK ) rc = rc2; + } + } + + if( pSorter->nPMA<=SORTER_MAX_MERGE_COUNT ){ + break; + }else{ + sqlite3_file *pTmp = pSorter->pTemp1; + pSorter->nPMA = iNew; + pSorter->pTemp1 = pTemp2; + pTemp2 = pTmp; + pSorter->iWriteOff = iWrite2; + pSorter->iReadOff = 0; + iWrite2 = 0; + } + }while( rc==SQLITE_OK ); + + if( pTemp2 ){ + sqlite3OsCloseFree(pTemp2); + } + *pbEof = (pSorter->aIter[pSorter->aTree[1]].pFile==0); return rc; } /* ** Advance to the next element in the sorter. @@ -78303,31 +75966,67 @@ */ SQLITE_PRIVATE int sqlite3VdbeSorterNext(sqlite3 *db, const VdbeCursor *pCsr, int *pbEof){ VdbeSorter *pSorter = pCsr->pSorter; int rc; /* Return code */ - assert( pSorter->bUsePMA || (pSorter->pReader==0 && pSorter->pMerger==0) ); - if( pSorter->bUsePMA ){ - assert( pSorter->pReader==0 || pSorter->pMerger==0 ); - assert( pSorter->bUseThreads==0 || pSorter->pReader ); - assert( pSorter->bUseThreads==1 || pSorter->pMerger ); -#if SQLITE_MAX_WORKER_THREADS>0 - if( pSorter->bUseThreads ){ - rc = vdbePmaReaderNext(pSorter->pReader); - *pbEof = (pSorter->pReader->pFd==0); - }else -#endif - /*if( !pSorter->bUseThreads )*/ { - assert( pSorter->pMerger->pTask==(&pSorter->aTask[0]) ); - rc = vdbeMergeEngineStep(pSorter->pMerger, pbEof); + if( pSorter->aTree ){ + int iPrev = pSorter->aTree[1];/* Index of iterator to advance */ + rc = vdbeSorterIterNext(db, &pSorter->aIter[iPrev]); + if( rc==SQLITE_OK ){ + int i; /* Index of aTree[] to recalculate */ + VdbeSorterIter *pIter1; /* First iterator to compare */ + VdbeSorterIter *pIter2; /* Second iterator to compare */ + u8 *pKey2; /* To pIter2->aKey, or 0 if record cached */ + + /* Find the first two iterators to compare. The one that was just + ** advanced (iPrev) and the one next to it in the array. */ + pIter1 = &pSorter->aIter[(iPrev & 0xFFFE)]; + pIter2 = &pSorter->aIter[(iPrev | 0x0001)]; + pKey2 = pIter2->aKey; + + for(i=(pSorter->nTree+iPrev)/2; i>0; i=i/2){ + /* Compare pIter1 and pIter2. Store the result in variable iRes. */ + int iRes; + if( pIter1->pFile==0 ){ + iRes = +1; + }else if( pIter2->pFile==0 ){ + iRes = -1; + }else{ + vdbeSorterCompare(pCsr, 0, + pIter1->aKey, pIter1->nKey, pKey2, pIter2->nKey, &iRes + ); + } + + /* If pIter1 contained the smaller value, set aTree[i] to its index. + ** Then set pIter2 to the next iterator to compare to pIter1. In this + ** case there is no cache of pIter2 in pSorter->pUnpacked, so set + ** pKey2 to point to the record belonging to pIter2. + ** + ** Alternatively, if pIter2 contains the smaller of the two values, + ** set aTree[i] to its index and update pIter1. If vdbeSorterCompare() + ** was actually called above, then pSorter->pUnpacked now contains + ** a value equivalent to pIter2. So set pKey2 to NULL to prevent + ** vdbeSorterCompare() from decoding pIter2 again. */ + if( iRes<=0 ){ + pSorter->aTree[i] = (int)(pIter1 - pSorter->aIter); + pIter2 = &pSorter->aIter[ pSorter->aTree[i ^ 0x0001] ]; + pKey2 = pIter2->aKey; + }else{ + if( pIter1->pFile ) pKey2 = 0; + pSorter->aTree[i] = (int)(pIter2 - pSorter->aIter); + pIter1 = &pSorter->aIter[ pSorter->aTree[i ^ 0x0001] ]; + } + + } + *pbEof = (pSorter->aIter[pSorter->aTree[1]].pFile==0); } }else{ - SorterRecord *pFree = pSorter->list.pList; - pSorter->list.pList = pFree->u.pNext; - pFree->u.pNext = 0; - if( pSorter->list.aMemory==0 ) vdbeSorterRecordFree(db, pFree); - *pbEof = !pSorter->list.pList; + SorterRecord *pFree = pSorter->pRecord; + pSorter->pRecord = pFree->pNext; + pFree->pNext = 0; + vdbeSorterRecordFree(db, pFree); + *pbEof = !pSorter->pRecord; rc = SQLITE_OK; } return rc; } @@ -78338,25 +76037,18 @@ static void *vdbeSorterRowkey( const VdbeSorter *pSorter, /* Sorter object */ int *pnKey /* OUT: Size of current key in bytes */ ){ void *pKey; - if( pSorter->bUsePMA ){ - PmaReader *pReader; -#if SQLITE_MAX_WORKER_THREADS>0 - if( pSorter->bUseThreads ){ - pReader = pSorter->pReader; - }else -#endif - /*if( !pSorter->bUseThreads )*/{ - pReader = &pSorter->pMerger->aReadr[pSorter->pMerger->aTree[1]]; - } - *pnKey = pReader->nKey; - pKey = pReader->aKey; + if( pSorter->aTree ){ + VdbeSorterIter *pIter; + pIter = &pSorter->aIter[ pSorter->aTree[1] ]; + *pnKey = pIter->nKey; + pKey = pIter->aKey; }else{ - *pnKey = pSorter->list.pList->nVal; - pKey = SRVAL(pSorter->list.pList); + *pnKey = pSorter->pRecord->nVal; + pKey = pSorter->pRecord->pVal; } return pKey; } /* @@ -78365,11 +76057,11 @@ SQLITE_PRIVATE int sqlite3VdbeSorterRowkey(const VdbeCursor *pCsr, Mem *pOut){ VdbeSorter *pSorter = pCsr->pSorter; void *pKey; int nKey; /* Sorter key to copy into pOut */ pKey = vdbeSorterRowkey(pSorter, &nKey); - if( sqlite3VdbeMemClearAndResize(pOut, nKey) ){ + if( sqlite3VdbeMemGrow(pOut, nKey, 0) ){ return SQLITE_NOMEM; } pOut->n = nKey; MemSetTypeFlag(pOut, MEM_Blob); memcpy(pOut->z, pKey, nKey); @@ -78380,52 +76072,26 @@ /* ** Compare the key in memory cell pVal with the key that the sorter cursor ** passed as the first argument currently points to. For the purposes of ** the comparison, ignore the rowid field at the end of each record. ** -** If the sorter cursor key contains any NULL values, consider it to be -** less than pVal. Even if pVal also contains NULL values. -** ** If an error occurs, return an SQLite error code (i.e. SQLITE_NOMEM). ** Otherwise, set *pRes to a negative, zero or positive value if the ** key in pVal is smaller than, equal to or larger than the current sorter ** key. -** -** This routine forms the core of the OP_SorterCompare opcode, which in -** turn is used to verify uniqueness when constructing a UNIQUE INDEX. */ SQLITE_PRIVATE int sqlite3VdbeSorterCompare( const VdbeCursor *pCsr, /* Sorter cursor */ Mem *pVal, /* Value to compare to current sorter key */ - int nKeyCol, /* Compare this many columns */ + int nKeyCol, /* Only compare this many fields */ int *pRes /* OUT: Result of comparison */ ){ VdbeSorter *pSorter = pCsr->pSorter; - UnpackedRecord *r2 = pSorter->pUnpacked; - KeyInfo *pKeyInfo = pCsr->pKeyInfo; - int i; void *pKey; int nKey; /* Sorter key to compare pVal with */ - if( r2==0 ){ - char *p; - r2 = pSorter->pUnpacked = sqlite3VdbeAllocUnpackedRecord(pKeyInfo,0,0,&p); - assert( pSorter->pUnpacked==(UnpackedRecord*)p ); - if( r2==0 ) return SQLITE_NOMEM; - r2->nField = nKeyCol; - } - assert( r2->nField==nKeyCol ); - pKey = vdbeSorterRowkey(pSorter, &nKey); - sqlite3VdbeRecordUnpack(pKeyInfo, nKey, pKey, r2); - for(i=0; i */ /* ** Walk an expression tree. Invoke the callback once for each node -** of the expression, while descending. (In other words, the callback +** of the expression, while decending. (In other words, the callback ** is invoked before visiting children.) ** ** The return value from the callback should be one of the WRC_* ** constants to specify how to proceed with the walk. ** @@ -79818,11 +77484,13 @@ ** likelihood(X,0.9375). */ /* TUNING: unlikely() probability is 0.0625. likely() is 0.9375 */ pExpr->iTable = pDef->zName[0]=='u' ? 62 : 938; } } + } #ifndef SQLITE_OMIT_AUTHORIZATION + if( pDef ){ auth = sqlite3AuthCheck(pParse, SQLITE_FUNCTION, 0, pDef->zName, 0); if( auth!=SQLITE_OK ){ if( auth==SQLITE_DENY ){ sqlite3ErrorMsg(pParse, "not authorized to use function: %s", pDef->zName); @@ -79829,13 +77497,13 @@ pNC->nErr++; } pExpr->op = TK_NULL; return WRC_Prune; } -#endif if( pDef->funcFlags & SQLITE_FUNC_CONSTANT ) ExprSetProperty(pExpr,EP_Constant); } +#endif if( is_agg && (pNC->ncFlags & NC_AllowAgg)==0 ){ sqlite3ErrorMsg(pParse, "misuse of aggregate function %.*s()", nId,zId); pNC->nErr++; is_agg = 0; }else if( no_such_func && pParse->db->init.busy==0 ){ @@ -79854,17 +77522,11 @@ pExpr->op2 = 0; while( pNC2 && !sqlite3FunctionUsesThisSrc(pExpr, pNC2->pSrcList) ){ pExpr->op2++; pNC2 = pNC2->pNext; } - assert( pDef!=0 ); - if( pNC2 ){ - assert( SQLITE_FUNC_MINMAX==NC_MinMaxAgg ); - testcase( (pDef->funcFlags & SQLITE_FUNC_MINMAX)!=0 ); - pNC2->ncFlags |= NC_HasAgg | (pDef->funcFlags & SQLITE_FUNC_MINMAX); - - } + if( pNC2 ) pNC2->ncFlags |= NC_HasAgg; pNC->ncFlags |= NC_AllowAgg; } /* FIX ME: Compute pExpr->affinity based on the expected return ** type of the function */ @@ -80221,11 +77883,11 @@ } return sqlite3ResolveOrderGroupBy(pParse, pSelect, pOrderBy, zType); } /* -** Resolve names in the SELECT statement p and all of its descendants. +** Resolve names in the SELECT statement p and all of its descendents. */ static int resolveSelectStep(Walker *pWalker, Select *p){ NameContext *pOuterNC; /* Context that contains this SELECT */ NameContext sNC; /* Name context of this SELECT */ int isCompound; /* True if p is a compound select */ @@ -80325,12 +77987,11 @@ ** expression, do not allow aggregates in any of the other expressions. */ assert( (p->selFlags & SF_Aggregate)==0 ); pGroupBy = p->pGroupBy; if( pGroupBy || (sNC.ncFlags & NC_HasAgg)!=0 ){ - assert( NC_MinMaxAgg==SF_MinMaxAgg ); - p->selFlags |= SF_Aggregate | (sNC.ncFlags&NC_MinMaxAgg); + p->selFlags |= SF_Aggregate; }else{ sNC.ncFlags &= ~NC_AllowAgg; } /* If a HAVING clause is present, then there must be a GROUP BY clause. @@ -80454,11 +78115,11 @@ */ SQLITE_PRIVATE int sqlite3ResolveExprNames( NameContext *pNC, /* Namespace to resolve expressions in. */ Expr *pExpr /* The expression to be analyzed. */ ){ - u16 savedHasAgg; + u8 savedHasAgg; Walker w; if( pExpr==0 ) return 0; #if SQLITE_MAX_EXPR_DEPTH>0 { @@ -80467,12 +78128,12 @@ return 1; } pParse->nHeight += pExpr->nHeight; } #endif - savedHasAgg = pNC->ncFlags & (NC_HasAgg|NC_MinMaxAgg); - pNC->ncFlags &= ~(NC_HasAgg|NC_MinMaxAgg); + savedHasAgg = pNC->ncFlags & NC_HasAgg; + pNC->ncFlags &= ~NC_HasAgg; memset(&w, 0, sizeof(w)); w.xExprCallback = resolveExprStep; w.xSelectCallback = resolveSelectStep; w.pParse = pNC->pParse; w.u.pNC = pNC; @@ -80483,12 +78144,13 @@ if( pNC->nErr>0 || w.pParse->nErr>0 ){ ExprSetProperty(pExpr, EP_Error); } if( pNC->ncFlags & NC_HasAgg ){ ExprSetProperty(pExpr, EP_Agg); + }else if( savedHasAgg ){ + pNC->ncFlags |= NC_HasAgg; } - pNC->ncFlags |= savedHasAgg; return ExprHasProperty(pExpr, EP_Error); } /* @@ -80584,11 +78246,11 @@ ** If pExpr is a column, a reference to a column via an 'AS' alias, ** or a sub-select with a column as the return value, then the ** affinity of that column is returned. Otherwise, 0x00 is returned, ** indicating no affinity for the expression. ** -** i.e. the WHERE clause expressions in the following statements all +** i.e. the WHERE clause expresssions in the following statements all ** have an affinity: ** ** CREATE TABLE t1(a); ** SELECT * FROM t1 WHERE a; ** SELECT a AS b FROM t1 WHERE b; @@ -81063,11 +78725,11 @@ exprSetHeight(pRoot); } } /* -** Allocate an Expr node which joins as many as two subtrees. +** Allocate a Expr node which joins as many as two subtrees. ** ** One or both of the subtrees can be NULL. Return a pointer to the new ** Expr node. Or, if an OOM error occurs, set pParse->db->mallocFailed, ** free the subtrees and return NULL. */ @@ -81173,11 +78835,11 @@ ** sure "nnn" is not too be to avoid a denial of service attack when ** the SQL statement comes from an external source. ** ** Wildcards of the form ":aaa", "@aaa", or "$aaa" are assigned the same number ** as the previous instance of the same wildcard. Or if this is the first -** instance of the wildcard, the next sequential variable number is +** instance of the wildcard, the next sequenial variable number is ** assigned. */ SQLITE_PRIVATE void sqlite3ExprAssignVarNumber(Parse *pParse, Expr *pExpr){ sqlite3 *db = pParse->db; const char *z; @@ -81308,11 +78970,11 @@ ** Note that with flags==EXPRDUP_REDUCE, this routines works on full-size ** (unreduced) Expr objects as they or originally constructed by the parser. ** During expression analysis, extra information is computed and moved into ** later parts of teh Expr object and that extra information might get chopped ** off if the expression is reduced. Note also that it does not work to -** make an EXPRDUP_REDUCE copy of a reduced expression. It is only legal +** make a EXPRDUP_REDUCE copy of a reduced expression. It is only legal ** to reduce a pristine expression tree from the parser. The implementation ** of dupedExprStructSize() contain multiple assert() statements that attempt ** to enforce this constraint. */ static int dupedExprStructSize(Expr *p, int flags){ @@ -81377,11 +79039,11 @@ /* ** This function is similar to sqlite3ExprDup(), except that if pzBuffer ** is not NULL then *pzBuffer is assumed to point to a buffer large enough ** to store the copy of expression p, the copies of p->u.zToken ** (if applicable), and the copies of the p->pLeft and p->pRight expressions, -** if any. Before returning, *pzBuffer is set to the first byte past the +** if any. Before returning, *pzBuffer is set to the first byte passed the ** portion of the buffer copied into by this function. */ static Expr *exprDup(sqlite3 *db, Expr *p, int flags, u8 **pzBuffer){ Expr *pNew = 0; /* Value to return */ if( p ){ @@ -81631,11 +79293,10 @@ pNew->selFlags = p->selFlags & ~SF_UsesEphemeral; pNew->addrOpenEphm[0] = -1; pNew->addrOpenEphm[1] = -1; pNew->nSelectRow = p->nSelectRow; pNew->pWith = withDup(db, p->pWith); - sqlite3SelectSetName(pNew, p->zSelName); return pNew; } #else SQLITE_PRIVATE Select *sqlite3SelectDup(sqlite3 *db, Select *p, int flags){ assert( p==0 ); @@ -81774,44 +79435,36 @@ } /* ** These routines are Walker callbacks. Walker.u.pi is a pointer ** to an integer. These routines are checking an expression to see -** if it is a constant. Set *Walker.u.i to 0 if the expression is +** if it is a constant. Set *Walker.u.pi to 0 if the expression is ** not constant. ** ** These callback routines are used to implement the following: ** -** sqlite3ExprIsConstant() pWalker->u.i==1 -** sqlite3ExprIsConstantNotJoin() pWalker->u.i==2 -** sqlite3ExprIsConstantOrFunction() pWalker->u.i==3 or 4 +** sqlite3ExprIsConstant() +** sqlite3ExprIsConstantNotJoin() +** sqlite3ExprIsConstantOrFunction() ** -** The sqlite3ExprIsConstantOrFunction() is used for evaluating expressions -** in a CREATE TABLE statement. The Walker.u.i value is 4 when parsing -** an existing schema and 3 when processing a new statement. A bound -** parameter raises an error for new statements, but is silently converted -** to NULL for existing schemas. This allows sqlite_master tables that -** contain a bound parameter because they were generated by older versions -** of SQLite to be parsed by newer versions of SQLite without raising a -** malformed schema error. */ static int exprNodeIsConstant(Walker *pWalker, Expr *pExpr){ - /* If pWalker->u.i is 2 then any term of the expression that comes from + /* If pWalker->u.i is 3 then any term of the expression that comes from ** the ON or USING clauses of a join disqualifies the expression ** from being considered constant. */ - if( pWalker->u.i==2 && ExprHasProperty(pExpr, EP_FromJoin) ){ + if( pWalker->u.i==3 && ExprHasProperty(pExpr, EP_FromJoin) ){ pWalker->u.i = 0; return WRC_Abort; } switch( pExpr->op ){ /* Consider functions to be constant if all their arguments are constant - ** and either pWalker->u.i==3 or 4 or the function as the SQLITE_FUNC_CONST + ** and either pWalker->u.i==2 or the function as the SQLITE_FUNC_CONST ** flag. */ case TK_FUNCTION: - if( pWalker->u.i>=3 || ExprHasProperty(pExpr,EP_Constant) ){ + if( pWalker->u.i==2 || ExprHasProperty(pExpr,EP_Constant) ){ return WRC_Continue; } /* Fall through */ case TK_ID: case TK_COLUMN: @@ -81821,23 +79474,10 @@ testcase( pExpr->op==TK_COLUMN ); testcase( pExpr->op==TK_AGG_FUNCTION ); testcase( pExpr->op==TK_AGG_COLUMN ); pWalker->u.i = 0; return WRC_Abort; - case TK_VARIABLE: - if( pWalker->u.i==4 ){ - /* Silently convert bound parameters that appear inside of CREATE - ** statements into a NULL when parsing the CREATE statement text out - ** of the sqlite_master table */ - pExpr->op = TK_NULL; - }else if( pWalker->u.i==3 ){ - /* A bound parameter in a CREATE statement that originates from - ** sqlite3_prepare() causes an error */ - pWalker->u.i = 0; - return WRC_Abort; - } - /* Fall through */ default: testcase( pExpr->op==TK_SELECT ); /* selectNodeIsConstant will disallow */ testcase( pExpr->op==TK_EXISTS ); /* selectNodeIsConstant will disallow */ return WRC_Continue; } @@ -81874,11 +79514,11 @@ ** that does no originate from the ON or USING clauses of a join. ** Return 0 if it involves variables or function calls or terms from ** an ON or USING clause. */ SQLITE_PRIVATE int sqlite3ExprIsConstantNotJoin(Expr *p){ - return exprIsConst(p, 2); + return exprIsConst(p, 3); } /* ** Walk an expression tree. Return 1 if the expression is constant ** or a function call with constant arguments. Return and 0 if there @@ -81886,13 +79526,12 @@ ** ** For the purposes of this function, a double-quoted string (ex: "abc") ** is considered a variable but a single-quoted string (ex: 'abc') is ** a constant. */ -SQLITE_PRIVATE int sqlite3ExprIsConstantOrFunction(Expr *p, u8 isInit){ - assert( isInit==0 || isInit==1 ); - return exprIsConst(p, 3+isInit); +SQLITE_PRIVATE int sqlite3ExprIsConstantOrFunction(Expr *p){ + return exprIsConst(p, 2); } /* ** If the expression p codes a constant integer that is small enough ** to fit in a 32-bit integer, return 1 and put the value of the integer @@ -82126,11 +79765,11 @@ ** ** SELECT FROM ** ** If the RHS of the IN operator is a list or a more complex subquery, then ** an ephemeral table might need to be generated from the RHS and then -** pX->iTable made to point to the ephemeral table instead of an +** pX->iTable made to point to the ephermeral table instead of an ** existing table. ** ** The inFlags parameter must contain exactly one of the bits ** IN_INDEX_MEMBERSHIP or IN_INDEX_LOOP. If inFlags contains ** IN_INDEX_MEMBERSHIP, then the generated table will be used for a @@ -82256,11 +79895,11 @@ /* If no preexisting index is available for the IN clause ** and IN_INDEX_NOOP is an allowed reply ** and the RHS of the IN operator is a list, not a subquery ** and the RHS is not contant or has two or fewer terms, - ** then it is not worth creating an ephemeral table to evaluate + ** then it is not worth creating an ephermeral table to evaluate ** the IN operator so return IN_INDEX_NOOP. */ if( eType==0 && (inFlags & IN_INDEX_NOOP_OK) && !ExprHasProperty(pX, EP_xIsSelect) @@ -82497,11 +80136,10 @@ assert( ExprHasProperty(pExpr, EP_xIsSelect) ); pSel = pExpr->x.pSelect; sqlite3SelectDestInit(&dest, 0, ++pParse->nMem); if( pExpr->op==TK_SELECT ){ dest.eDest = SRT_Mem; - dest.iSdst = dest.iSDParm; sqlite3VdbeAddOp2(v, OP_Null, 0, dest.iSDParm); VdbeComment((v, "Init subquery result")); }else{ dest.eDest = SRT_Exists; sqlite3VdbeAddOp2(v, OP_Integer, 0, dest.iSDParm); @@ -83017,13 +80655,20 @@ /* ** Generate code to move content from registers iFrom...iFrom+nReg-1 ** over to iTo..iTo+nReg-1. Keep the column cache up-to-date. */ SQLITE_PRIVATE void sqlite3ExprCodeMove(Parse *pParse, int iFrom, int iTo, int nReg){ + int i; + struct yColCache *p; assert( iFrom>=iTo+nReg || iFrom+nReg<=iTo ); sqlite3VdbeAddOp3(pParse->pVdbe, OP_Move, iFrom, iTo, nReg); - sqlite3ExprCacheRemove(pParse, iFrom, nReg); + for(i=0, p=pParse->aColCache; i
nExpr; i++){ - sqlite3TreeViewExpr(pView, pList->a[i].pExpr, i nExpr-1); -#if 0 - if( pList->a[i].zName ){ + sqlite3ExplainPrintf(pOut, "item[%d] = ", i); + sqlite3ExplainPush(pOut); + sqlite3ExplainExpr(pOut, pList->a[i].pExpr); + sqlite3ExplainPop(pOut); + if( pList->a[i].zName ){ sqlite3ExplainPrintf(pOut, " AS %s", pList->a[i].zName); } if( pList->a[i].bSpanIsTab ){ sqlite3ExplainPrintf(pOut, " (%s)", pList->a[i].zSpan); } -#endif + if( i nExpr-1 ){ + sqlite3ExplainNL(pOut); + } } + sqlite3ExplainPop(pOut); } - sqlite3TreeViewPop(pView); } #endif /* SQLITE_DEBUG */ /* ** Generate code that pushes the value of every element of the given @@ -84120,11 +81788,11 @@ ** The above is equivalent to ** ** x>=y AND x<=z ** ** Code it as such, taking care to do the common subexpression -** elimination of x. +** elementation of x. */ static void exprCodeBetween( Parse *pParse, /* Parsing and code generating context */ Expr *pExpr, /* The BETWEEN expression */ int dest, /* Jump here if the jump is taken */ @@ -84857,11 +82525,11 @@ /* ** Deallocate a register, making available for reuse for some other ** purpose. ** ** If a register is currently being used by the column cache, then -** the deallocation is deferred until the column cache line that uses +** the dallocation is deferred until the column cache line that uses ** the register becomes stale. */ SQLITE_PRIVATE void sqlite3ReleaseTempReg(Parse *pParse, int iReg){ if( iReg && pParse->nTempReg auth.authLevel ){ - rc = SQLITE_AUTH_USER; - } - } -#endif if( rc ){ int iDb = db->nDb - 1; assert( iDb>=2 ); if( db->aDb[iDb].pBt ){ sqlite3BtreeClose(db->aDb[iDb].pBt); @@ -88283,11 +85911,11 @@ sqlite3 *db, int (*xAuth)(void*,int,const char*,const char*,const char*,const char*), void *pArg ){ sqlite3_mutex_enter(db->mutex); - db->xAuth = (sqlite3_xauth)xAuth; + db->xAuth = xAuth; db->pAuthArg = pArg; sqlite3ExpirePreparedStatements(db); sqlite3_mutex_leave(db->mutex); return SQLITE_OK; } @@ -88318,15 +85946,11 @@ ){ sqlite3 *db = pParse->db; /* Database handle */ char *zDb = db->aDb[iDb].zName; /* Name of attached database */ int rc; /* Auth callback return code */ - rc = db->xAuth(db->pAuthArg, SQLITE_READ, zTab,zCol,zDb,pParse->zAuthContext -#ifdef SQLITE_USER_AUTHENTICATION - ,db->auth.zAuthUser -#endif - ); + rc = db->xAuth(db->pAuthArg, SQLITE_READ, zTab,zCol,zDb,pParse->zAuthContext); if( rc==SQLITE_DENY ){ if( db->nDb>2 || iDb!=0 ){ sqlite3ErrorMsg(pParse, "access to %s.%s.%s is prohibited",zDb,zTab,zCol); }else{ sqlite3ErrorMsg(pParse, "access to %s.%s is prohibited", zTab, zCol); @@ -88422,15 +86046,11 @@ } if( db->xAuth==0 ){ return SQLITE_OK; } - rc = db->xAuth(db->pAuthArg, code, zArg1, zArg2, zArg3, pParse->zAuthContext -#ifdef SQLITE_USER_AUTHENTICATION - ,db->auth.zAuthUser -#endif - ); + rc = db->xAuth(db->pAuthArg, code, zArg1, zArg2, zArg3, pParse->zAuthContext); if( rc==SQLITE_DENY ){ sqlite3ErrorMsg(pParse, "not authorized"); pParse->rc = SQLITE_AUTH; }else if( rc!=SQLITE_OK && rc!=SQLITE_IGNORE ){ rc = SQLITE_DENY; @@ -88625,21 +86245,10 @@ || sqlite3VdbeAssertMayAbort(v, pParse->mayAbort)); if( v ){ while( sqlite3VdbeDeletePriorOpcode(v, OP_Close) ){} sqlite3VdbeAddOp0(v, OP_Halt); -#if SQLITE_USER_AUTHENTICATION - if( pParse->nTableLock>0 && db->init.busy==0 ){ - sqlite3UserAuthInit(db); - if( db->auth.authLevel rc = SQLITE_AUTH_USER; - sqlite3ErrorMsg(pParse, "user not authenticated"); - return; - } - } -#endif - /* The cookie mask contains one bit for each database file open. ** (Bit 0 is for main, bit 1 is for temp, and so forth.) Bits are ** set for each database that is used. Generate code to start a ** transaction on each used database and to verify the schema cookie ** on each used database. @@ -88751,20 +86360,10 @@ sqlite3DbFree(db, zSql); memcpy(&pParse->nVar, saveBuf, SAVE_SZ); pParse->nested--; } -#if SQLITE_USER_AUTHENTICATION -/* -** Return TRUE if zTable is the name of the system table that stores the -** list of users and their access credentials. -*/ -SQLITE_PRIVATE int sqlite3UserAuthTable(const char *zTable){ - return sqlite3_stricmp(zTable, "sqlite_user")==0; -} -#endif - /* ** Locate the in-memory structure that describes a particular database ** table given the name of that table and (optionally) the name of the ** database containing the table. Return NULL if not found. ** @@ -88776,25 +86375,20 @@ ** See also sqlite3LocateTable(). */ SQLITE_PRIVATE Table *sqlite3FindTable(sqlite3 *db, const char *zName, const char *zDatabase){ Table *p = 0; int i; + int nName; assert( zName!=0 ); + nName = sqlite3Strlen30(zName); /* All mutexes are required for schema access. Make sure we hold them. */ assert( zDatabase!=0 || sqlite3BtreeHoldsAllMutexes(db) ); -#if SQLITE_USER_AUTHENTICATION - /* Only the admin user is allowed to know that the sqlite_user table - ** exists */ - if( db->auth.authLevel nDb; i++){ int j = (i<2) ? i^1 : i; /* Search TEMP before MAIN */ if( zDatabase!=0 && sqlite3StrICmp(zDatabase, db->aDb[j].zName) ) continue; assert( sqlite3SchemaMutexHeld(db, j, 0) ); - p = sqlite3HashFind(&db->aDb[j].pSchema->tblHash, zName); + p = sqlite3HashFind(&db->aDb[j].pSchema->tblHash, zName, nName); if( p ) break; } return p; } @@ -88830,16 +86424,10 @@ }else{ sqlite3ErrorMsg(pParse, "%s: %s", zMsg, zName); } pParse->checkSchema = 1; } -#if SQLITE_USER_AUTHENICATION - else if( pParse->db->auth.authLevel nDb; i++){ int j = (i<2) ? i^1 : i; /* Search TEMP before MAIN */ Schema *pSchema = db->aDb[j].pSchema; assert( pSchema ); if( zDb && sqlite3StrICmp(zDb, db->aDb[j].zName) ) continue; assert( sqlite3SchemaMutexHeld(db, j, 0) ); - p = sqlite3HashFind(&pSchema->idxHash, zName); + p = sqlite3HashFind(&pSchema->idxHash, zName, nName); if( p ) break; } return p; } @@ -88904,13 +86493,10 @@ #endif if( db==0 || db->pnBytesFreed==0 ) sqlite3KeyInfoUnref(p->pKeyInfo); sqlite3ExprDelete(db, p->pPartIdxWhere); sqlite3DbFree(db, p->zColAff); if( p->isResized ) sqlite3DbFree(db, p->azColl); -#ifdef SQLITE_ENABLE_STAT3_OR_STAT4 - sqlite3_free(p->aiRowEst); -#endif sqlite3DbFree(db, p); } /* ** For the index called zIdxName which is found in the database iDb, @@ -88918,15 +86504,17 @@ ** the index hash table and free all memory structures associated ** with the index. */ SQLITE_PRIVATE void sqlite3UnlinkAndDeleteIndex(sqlite3 *db, int iDb, const char *zIdxName){ Index *pIndex; + int len; Hash *pHash; assert( sqlite3SchemaMutexHeld(db, iDb, 0) ); pHash = &db->aDb[iDb].pSchema->idxHash; - pIndex = sqlite3HashInsert(pHash, zIdxName, 0); + len = sqlite3Strlen30(zIdxName); + pIndex = sqlite3HashInsert(pHash, zIdxName, len, 0); if( ALWAYS(pIndex) ){ if( pIndex->pTable->pIndex==pIndex ){ pIndex->pTable->pIndex = pIndex->pNext; }else{ Index *p; @@ -89082,11 +86670,11 @@ pNext = pIndex->pNext; assert( pIndex->pSchema==pTable->pSchema ); if( !db || db->pnBytesFreed==0 ){ char *zName = pIndex->zName; TESTONLY ( Index *pOld = ) sqlite3HashInsert( - &pIndex->pSchema->idxHash, zName, 0 + &pIndex->pSchema->idxHash, zName, sqlite3Strlen30(zName), 0 ); assert( db==0 || sqlite3SchemaMutexHeld(db, 0, pIndex->pSchema) ); assert( pOld==pIndex || pOld==0 ); } freeIndex(db, pIndex); @@ -89125,11 +86713,12 @@ assert( iDb>=0 && iDb nDb ); assert( zTabName ); assert( sqlite3SchemaMutexHeld(db, iDb, 0) ); testcase( zTabName[0]==0 ); /* Zero-length table names are allowed */ pDb = &db->aDb[iDb]; - p = sqlite3HashInsert(&pDb->pSchema->tblHash, zTabName, 0); + p = sqlite3HashInsert(&pDb->pSchema->tblHash, zTabName, + sqlite3Strlen30(zTabName),0); sqlite3DeleteTable(db, p); db->flags |= SQLITE_InternChanges; } /* @@ -89649,11 +87238,11 @@ /* If pszEst is not NULL, store an estimate of the field size. The ** estimate is scaled so that the size of an integer is 1. */ if( pszEst ){ *pszEst = 1; /* default size is approx 4 bytes */ - if( aff db; p = pParse->pNewTable; if( p!=0 ){ pCol = &(p->aCol[p->nCol-1]); - if( !sqlite3ExprIsConstantOrFunction(pSpan->pExpr, db->init.busy) ){ + if( !sqlite3ExprIsConstantOrFunction(pSpan->pExpr) ){ sqlite3ErrorMsg(pParse, "default value of column [%s] is not constant", pCol->zName); }else{ /* A copy of pExpr is used instead of the original, as pExpr contains ** tokens that point to volatile memory. The 'span' of the expression @@ -90020,12 +87609,12 @@ k = sqlite3Strlen30(zStmt); identPut(zStmt, &k, p->zName); zStmt[k++] = '('; for(pCol=p->aCol, i=0; i nCol; i++, pCol++){ static const char * const azType[] = { - /* SQLITE_AFF_NONE */ "", /* SQLITE_AFF_TEXT */ " TEXT", + /* SQLITE_AFF_NONE */ "", /* SQLITE_AFF_NUMERIC */ " NUM", /* SQLITE_AFF_INTEGER */ " INT", /* SQLITE_AFF_REAL */ " REAL" }; int len; @@ -90033,19 +87622,19 @@ sqlite3_snprintf(n-k, &zStmt[k], zSep); k += sqlite3Strlen30(&zStmt[k]); zSep = zSep2; identPut(zStmt, &k, pCol->zName); - assert( pCol->affinity-SQLITE_AFF_NONE >= 0 ); - assert( pCol->affinity-SQLITE_AFF_NONE < ArraySize(azType) ); - testcase( pCol->affinity==SQLITE_AFF_NONE ); + assert( pCol->affinity-SQLITE_AFF_TEXT >= 0 ); + assert( pCol->affinity-SQLITE_AFF_TEXT < ArraySize(azType) ); testcase( pCol->affinity==SQLITE_AFF_TEXT ); + testcase( pCol->affinity==SQLITE_AFF_NONE ); testcase( pCol->affinity==SQLITE_AFF_NUMERIC ); testcase( pCol->affinity==SQLITE_AFF_INTEGER ); testcase( pCol->affinity==SQLITE_AFF_REAL ); - zType = azType[pCol->affinity - SQLITE_AFF_NONE]; + zType = azType[pCol->affinity - SQLITE_AFF_TEXT]; len = sqlite3Strlen30(zType); assert( pCol->affinity==SQLITE_AFF_NONE || pCol->affinity==sqlite3AffinityType(zType, 0) ); memcpy(&zStmt[k], zType, len); k += len; @@ -90125,11 +87714,11 @@ ** ** (1) Convert the OP_CreateTable into an OP_CreateIndex. There is ** no rowid btree for a WITHOUT ROWID. Instead, the canonical ** data storage is a covering index btree. ** (2) Bypass the creation of the sqlite_master table entry -** for the PRIMARY KEY as the primary key index is now +** for the PRIMARY KEY as the the primary key index is now ** identified by the sqlite_master table entry of the table itself. ** (3) Set the Index.tnum of the PRIMARY KEY Index object in the ** schema to the rootpage from the main table. ** (4) Set all columns of the PRIMARY KEY schema object to be NOT NULL. ** (5) Add all table columns to the PRIMARY KEY Index object @@ -90146,11 +87735,11 @@ int i, j; sqlite3 *db = pParse->db; Vdbe *v = pParse->pVdbe; /* Convert the OP_CreateTable opcode that would normally create the - ** root-page for the table into an OP_CreateIndex opcode. The index + ** root-page for the table into a OP_CreateIndex opcode. The index ** created will become the PRIMARY KEY index. */ if( pParse->addrCrTab ){ assert( v ); sqlite3VdbeGetOp(v, pParse->addrCrTab)->opcode = OP_CreateIndex; @@ -90447,11 +88036,12 @@ */ if( db->init.busy ){ Table *pOld; Schema *pSchema = p->pSchema; assert( sqlite3SchemaMutexHeld(db, iDb, 0) ); - pOld = sqlite3HashInsert(&pSchema->tblHash, p->zName, p); + pOld = sqlite3HashInsert(&pSchema->tblHash, p->zName, + sqlite3Strlen30(p->zName),p); if( pOld ){ assert( p==pOld ); /* Malloc must have failed inside HashInsert() */ db->mallocFailed = 1; return; } @@ -90558,11 +88148,11 @@ Table *pSelTab; /* A fake table from which we get the result set */ Select *pSel; /* Copy of the SELECT that implements the view */ int nErr = 0; /* Number of errors encountered */ int n; /* Temporarily holds the number of cursors assigned */ sqlite3 *db = pParse->db; /* Database connection for malloc errors */ - sqlite3_xauth xAuth; /* Saved xAuth pointer */ + int (*xAuth)(void*,int,const char*,const char*,const char*,const char*); assert( pTable ); #ifndef SQLITE_OMIT_VIRTUALTABLE if( sqlite3VtabCallConnect(pParse, pTable) ){ @@ -91097,11 +88687,11 @@ pFKey->aAction[0] = (u8)(flags & 0xff); /* ON DELETE action */ pFKey->aAction[1] = (u8)((flags >> 8 ) & 0xff); /* ON UPDATE action */ assert( sqlite3SchemaMutexHeld(db, 0, p->pSchema) ); pNextTo = (FKey *)sqlite3HashInsert(&p->pSchema->fkeyHash, - pFKey->zTo, (void *)pFKey + pFKey->zTo, sqlite3Strlen30(pFKey->zTo), (void *)pFKey ); if( pNextTo==pFKey ){ db->mallocFailed = 1; goto fk_end; } @@ -91160,11 +88750,11 @@ int addr2; /* Address to jump to for next iteration */ int tnum; /* Root page of index */ int iPartIdxLabel; /* Jump to this label to skip a row */ Vdbe *v; /* Generate code into this virtual machine */ KeyInfo *pKey; /* KeyInfo for index */ - int regRecord; /* Register holding assembled index record */ + int regRecord; /* Register holding assemblied index record */ sqlite3 *db = pParse->db; /* The database connection */ int iDb = sqlite3SchemaToIndex(db, pIndex->pSchema); #ifndef SQLITE_OMIT_AUTHORIZATION if( sqlite3AuthCheck(pParse, SQLITE_REINDEX, pIndex->zName, 0, @@ -91185,11 +88775,11 @@ } pKey = sqlite3KeyInfoOfIndex(pParse, pIndex); /* Open the sorter cursor if we are to use one. */ iSorter = pParse->nTab++; - sqlite3VdbeAddOp4(v, OP_SorterOpen, iSorter, 0, pIndex->nKeyCol, (char*) + sqlite3VdbeAddOp4(v, OP_SorterOpen, iSorter, 0, 0, (char*) sqlite3KeyInfoRef(pKey), P4_KEYINFO); /* Open the table. Loop through all rows of the table, inserting index ** records into the sorter. */ sqlite3OpenTable(pParse, iTab, iDb, pTab, OP_OpenRead); @@ -91373,14 +88963,10 @@ pDb = &db->aDb[iDb]; assert( pTab!=0 ); assert( pParse->nErr==0 ); if( sqlite3StrNICmp(pTab->zName, "sqlite_", 7)==0 - && db->init.busy==0 -#if SQLITE_USER_AUTHENTICATION - && sqlite3UserAuthTable(pTab->zName)==0 -#endif && sqlite3StrNICmp(&pTab->zName[7],"altertab_",9)!=0 ){ sqlite3ErrorMsg(pParse, "table %s may not be indexed", pTab->zName); goto exit_create_index; } #ifndef SQLITE_OMIT_VIEW @@ -91538,11 +89124,11 @@ sqlite3ErrorMsg(pParse, "table %s has no column named %s", pTab->zName, zColName); pParse->checkSchema = 1; goto exit_create_index; } - assert( j<=0x7fff ); + assert( pTab->nCol<=0x7fff && j<=0x7fff ); pIndex->aiColumn[i] = (i16)j; if( pListItem->pExpr ){ int nColl; assert( pListItem->pExpr->op==TK_COLLATE ); zColl = pListItem->pExpr->u.zToken; @@ -91649,11 +89235,12 @@ */ if( db->init.busy ){ Index *p; assert( sqlite3SchemaMutexHeld(db, 0, pIndex->pSchema) ); p = sqlite3HashInsert(&pIndex->pSchema->idxHash, - pIndex->zName, pIndex); + pIndex->zName, sqlite3Strlen30(pIndex->zName), + pIndex); if( p ){ assert( p==pIndex ); /* Malloc must have failed */ db->mallocFailed = 1; goto exit_create_index; } @@ -91764,11 +89351,11 @@ /* ** Fill the Index.aiRowEst[] array with default information - information ** to be used when we have not run the ANALYZE command. ** -** aiRowEst[0] is supposed to contain the number of elements in the index. +** aiRowEst[0] is suppose to contain the number of elements in the index. ** Since we do not know, guess 1 million. aiRowEst[1] is an estimate of the ** number of rows in the table that match any particular value of the ** first column of the index. aiRowEst[2] is an estimate of the number ** of rows that match any particular combination of the first 2 columns ** of the index. And so forth. It must always be the case that @@ -92143,11 +89730,11 @@ ** end of a growing FROM clause. The "p" parameter is the part of ** the FROM clause that has already been constructed. "p" is NULL ** if this is the first term of the FROM clause. pTable and pDatabase ** are the name of the table and database named in the FROM clause term. ** pDatabase is NULL if the database name qualifier is missing - the -** usual case. If the term has an alias, then pAlias points to the +** usual case. If the term has a alias, then pAlias points to the ** alias token. If the term is a subquery, then pSubquery is the ** SELECT statement that the subquery encodes. The pTable and ** pDatabase parameters are NULL for subqueries. The pOn and pUsing ** parameters are the content of the ON and USING clauses. ** @@ -92906,11 +90493,11 @@ ** specified by zName and nName is not found and parameter 'create' is ** true, then create a new entry. Otherwise return NULL. ** ** Each pointer stored in the sqlite3.aCollSeq hash table contains an ** array of three CollSeq structures. The first is the collation sequence -** preferred for UTF-8, the second UTF-16le, and the third UTF-16be. +** prefferred for UTF-8, the second UTF-16le, and the third UTF-16be. ** ** Stored immediately after the three collation sequences is a copy of ** the collation sequence name. A pointer to this string is stored in ** each collation sequence structure. */ @@ -92918,15 +90505,15 @@ sqlite3 *db, /* Database connection */ const char *zName, /* Name of the collating sequence */ int create /* Create a new entry if true */ ){ CollSeq *pColl; - pColl = sqlite3HashFind(&db->aCollSeq, zName); + int nName = sqlite3Strlen30(zName); + pColl = sqlite3HashFind(&db->aCollSeq, zName, nName); if( 0==pColl && create ){ - int nName = sqlite3Strlen30(zName); - pColl = sqlite3DbMallocZero(db, 3*sizeof(*pColl) + nName + 1); + pColl = sqlite3DbMallocZero(db, 3*sizeof(*pColl) + nName + 1 ); if( pColl ){ CollSeq *pDel = 0; pColl[0].zName = (char*)&pColl[3]; pColl[0].enc = SQLITE_UTF8; pColl[1].zName = (char*)&pColl[3]; @@ -92933,11 +90520,11 @@ pColl[1].enc = SQLITE_UTF16LE; pColl[2].zName = (char*)&pColl[3]; pColl[2].enc = SQLITE_UTF16BE; memcpy(pColl[0].zName, zName, nName); pColl[0].zName[nName] = 0; - pDel = sqlite3HashInsert(&db->aCollSeq, pColl[0].zName, pColl); + pDel = sqlite3HashInsert(&db->aCollSeq, pColl[0].zName, nName, pColl); /* If a malloc() failure occurred in sqlite3HashInsert(), it will ** return the pColl pointer to be deleted (because it wasn't added ** to the hash table). */ @@ -93333,11 +90920,11 @@ */ SQLITE_PRIVATE void sqlite3MaterializeView( Parse *pParse, /* Parsing context */ Table *pView, /* View definition */ Expr *pWhere, /* Optional WHERE clause to be added */ - int iCur /* Cursor number for ephemeral table */ + int iCur /* Cursor number for ephemerial table */ ){ SelectDest dest; Select *pSel; SrcList *pFrom; sqlite3 *db = pParse->db; @@ -93491,11 +91078,11 @@ int iEphCur = 0; /* Ephemeral table holding all primary key values */ int iRowSet = 0; /* Register for rowset of rows to delete */ int addrBypass = 0; /* Address of jump over the delete logic */ int addrLoop = 0; /* Top of the delete loop */ int addrDelete = 0; /* Jump directly to the delete logic */ - int addrEphOpen = 0; /* Instruction to open the Ephemeral table */ + int addrEphOpen = 0; /* Instruction to open the Ephermeral table */ #ifndef SQLITE_OMIT_TRIGGER int isView; /* True if attempting to delete from a view */ Trigger *pTrigger; /* List of table triggers, if required */ #endif @@ -93571,11 +91158,11 @@ } if( pParse->nested==0 ) sqlite3VdbeCountChanges(v); sqlite3BeginWriteOperation(pParse, 1, iDb); /* If we are trying to delete from a view, realize that view into - ** an ephemeral table. + ** a ephemeral table. */ #if !defined(SQLITE_OMIT_VIEW) && !defined(SQLITE_OMIT_TRIGGER) if( isView ){ sqlite3MaterializeView(pParse, pTab, pWhere, iTabCur); iDataCur = iIdxCur = iTabCur; @@ -93625,11 +91212,11 @@ pPk = 0; nPk = 1; iRowSet = ++pParse->nMem; sqlite3VdbeAddOp2(v, OP_Null, 0, iRowSet); }else{ - /* For a WITHOUT ROWID table, create an ephemeral table used to + /* For a WITHOUT ROWID table, create an ephermeral table used to ** hold all primary keys for rows to be deleted. */ pPk = sqlite3PrimaryKeyIndex(pTab); assert( pPk!=0 ); nPk = pPk->nKeyCol; iPk = pParse->nMem+1; @@ -93709,24 +91296,22 @@ ** deleting from and all its indices. If this is a view, then the ** only effect this statement has is to fire the INSTEAD OF ** triggers. */ if( !isView ){ - testcase( IsVirtual(pTab) ); sqlite3OpenTableAndIndices(pParse, pTab, OP_OpenWrite, iTabCur, aToOpen, &iDataCur, &iIdxCur); - assert( pPk || IsVirtual(pTab) || iDataCur==iTabCur ); - assert( pPk || IsVirtual(pTab) || iIdxCur==iDataCur+1 ); + assert( pPk || iDataCur==iTabCur ); + assert( pPk || iIdxCur==iDataCur+1 ); } /* Set up a loop over the rowids/primary-keys that were found in the ** where-clause loop above. */ if( okOnePass ){ /* Just one row. Hence the top-of-loop is a no-op */ - assert( nKey==nPk ); /* OP_Found will use an unpacked key */ - assert( !IsVirtual(pTab) ); + assert( nKey==nPk ); /* OP_Found will use an unpacked key */ if( aToOpen[iDataCur-iTabCur] ){ assert( pPk!=0 ); sqlite3VdbeAddOp4Int(v, OP_NotFound, iDataCur, addrBypass, iKey, nKey); VdbeCoverage(v); } @@ -93800,11 +91385,11 @@ sqlite3ExprDelete(db, pWhere); sqlite3DbFree(db, aToOpen); return; } /* Make sure "isView" and other macros defined above are undefined. Otherwise -** they may interfere with compilation of other functions in this file +** thely may interfere with compilation of other functions in this file ** (or in another file, if this file becomes part of the amalgamation). */ #ifdef isView #undef isView #endif #ifdef pTrigger @@ -94094,11 +91679,11 @@ ** May you do good and not evil. ** May you find forgiveness for yourself and forgive others. ** May you share freely, never taking more than you give. ** ************************************************************************* -** This file contains the C-language implementations for many of the SQL +** This file contains the C-language implementions for many of the SQL ** functions of SQLite. (Some function, and in particular the date and ** time functions, are implemented separately.) */ /* #include */ /* #include */ @@ -94105,14 +91690,11 @@ /* ** Return the collating function associated with a function. */ static CollSeq *sqlite3GetFuncCollSeq(sqlite3_context *context){ - VdbeOp *pOp = &context->pVdbe->aOp[context->iOp-1]; - assert( pOp->opcode==OP_CollSeq ); - assert( pOp->p4type==P4_COLLSEQ ); - return pOp->p4.pColl; + return context->pColl; } /* ** Indicate that the accumulator load should be skipped on this ** iteration of the aggregate loop. @@ -94411,18 +91993,17 @@ p1--; } for(z2=z; *z2 && p2; p2--){ SQLITE_SKIP_UTF8(z2); } - sqlite3_result_text64(context, (char*)z, z2-z, SQLITE_TRANSIENT, - SQLITE_UTF8); + sqlite3_result_text(context, (char*)z, (int)(z2-z), SQLITE_TRANSIENT); }else{ if( p1+p2>len ){ p2 = len-p1; if( p2<0 ) p2 = 0; } - sqlite3_result_blob64(context, (char*)&z[p1], (u64)p2, SQLITE_TRANSIENT); + sqlite3_result_blob(context, (char*)&z[p1], (int)p2, SQLITE_TRANSIENT); } } /* ** Implementation of the round() function @@ -94477,11 +92058,11 @@ testcase( nByte==db->aLimit[SQLITE_LIMIT_LENGTH]+1 ); if( nByte>db->aLimit[SQLITE_LIMIT_LENGTH] ){ sqlite3_result_error_toobig(context); z = 0; }else{ - z = sqlite3Malloc(nByte); + z = sqlite3Malloc((int)nByte); if( !z ){ sqlite3_result_error_nomem(context); } } return z; @@ -94653,16 +92234,14 @@ ** character is exactly one byte in size. Also, all characters are ** able to participate in upper-case-to-lower-case mappings in EBCDIC ** whereas only characters less than 0x80 do in ASCII. */ #if defined(SQLITE_EBCDIC) -# define sqlite3Utf8Read(A) (*((*A)++)) -# define GlobUpperToLower(A) A = sqlite3UpperToLower[A] -# define GlobUpperToLowerAscii(A) A = sqlite3UpperToLower[A] +# define sqlite3Utf8Read(A) (*((*A)++)) +# define GlobUpperToLower(A) A = sqlite3UpperToLower[A] #else -# define GlobUpperToLower(A) if( A<=0x7f ){ A = sqlite3UpperToLower[A]; } -# define GlobUpperToLowerAscii(A) A = sqlite3UpperToLower[A] +# define GlobUpperToLower(A) if( !((A)&~0x7f) ){ A = sqlite3UpperToLower[A]; } #endif static const struct compareInfo globInfo = { '*', '?', '[', 0 }; /* The correct SQL-92 behavior is for the LIKE operator to ignore ** case. Thus 'a' LIKE 'A' would be true. */ @@ -94671,11 +92250,11 @@ ** is case sensitive causing 'a' LIKE 'A' to be false */ static const struct compareInfo likeInfoAlt = { '%', '_', 0, 0 }; /* ** Compare two UTF-8 strings for equality where the first string can -** potentially be a "glob" or "like" expression. Return true (1) if they +** potentially be a "glob" expression. Return true (1) if they ** are the same and false (0) if they are different. ** ** Globbing rules: ** ** '*' Matches any sequence of zero or more characters. @@ -94691,147 +92270,120 @@ ** in the list by making it the first character after '[' or '^'. A ** range of characters can be specified using '-'. Example: ** "[a-z]" matches any single lower-case letter. To match a '-', make ** it the last character in the list. ** -** Like matching rules: -** -** '%' Matches any sequence of zero or more characters -** -*** '_' Matches any one character -** -** Ec Where E is the "esc" character and c is any other -** character, including '%', '_', and esc, match exactly c. -** -** The comments through this routine usually assume glob matching. -** ** This routine is usually quick, but can be N**2 in the worst case. +** +** Hints: to match '*' or '?', put them in "[]". Like this: +** +** abc[*]xyz Matches "abc*xyz" only */ static int patternCompare( const u8 *zPattern, /* The glob pattern */ const u8 *zString, /* The string to compare against the glob */ const struct compareInfo *pInfo, /* Information about how to do the compare */ u32 esc /* The escape character */ ){ - u32 c, c2; /* Next pattern and input string chars */ - u32 matchOne = pInfo->matchOne; /* "?" or "_" */ - u32 matchAll = pInfo->matchAll; /* "*" or "%" */ - u32 matchOther; /* "[" or the escape character */ - u8 noCase = pInfo->noCase; /* True if uppercase==lowercase */ - const u8 *zEscaped = 0; /* One past the last escaped input char */ - - /* The GLOB operator does not have an ESCAPE clause. And LIKE does not - ** have the matchSet operator. So we either have to look for one or - ** the other, never both. Hence the single variable matchOther is used - ** to store the one we have to look for. - */ - matchOther = esc ? esc : pInfo->matchSet; + u32 c, c2; + int invert; + int seen; + u8 matchOne = pInfo->matchOne; + u8 matchAll = pInfo->matchAll; + u8 matchSet = pInfo->matchSet; + u8 noCase = pInfo->noCase; + int prevEscape = 0; /* True if the previous character was 'escape' */ while( (c = sqlite3Utf8Read(&zPattern))!=0 ){ - if( c==matchAll ){ /* Match "*" */ - /* Skip over multiple "*" characters in the pattern. If there - ** are also "?" characters, skip those as well, but consume a - ** single character of the input string for each "?" skipped */ + if( c==matchAll && !prevEscape ){ while( (c=sqlite3Utf8Read(&zPattern)) == matchAll || c == matchOne ){ if( c==matchOne && sqlite3Utf8Read(&zString)==0 ){ return 0; } } if( c==0 ){ - return 1; /* "*" at the end of the pattern matches */ - }else if( c==matchOther ){ - if( esc ){ - c = sqlite3Utf8Read(&zPattern); - if( c==0 ) return 0; - }else{ - /* "[...]" immediately follows the "*". We have to do a slow - ** recursive search in this case, but it is an unusual case. */ - assert( matchOther<0x80 ); /* '[' is a single-byte character */ - while( *zString - && patternCompare(&zPattern[-1],zString,pInfo,esc)==0 ){ - SQLITE_SKIP_UTF8(zString); - } - return *zString!=0; - } - } - - /* At this point variable c contains the first character of the - ** pattern string past the "*". Search in the input string for the - ** first matching character and recursively contine the match from - ** that point. - ** - ** For a case-insensitive search, set variable cx to be the same as - ** c but in the other case and search the input string for either - ** c or cx. - */ - if( c<=0x80 ){ - u32 cx; - if( noCase ){ - cx = sqlite3Toupper(c); - c = sqlite3Tolower(c); - }else{ - cx = c; - } - while( (c2 = *(zString++))!=0 ){ - if( c2!=c && c2!=cx ) continue; - if( patternCompare(zPattern,zString,pInfo,esc) ) return 1; - } - }else{ - while( (c2 = sqlite3Utf8Read(&zString))!=0 ){ - if( c2!=c ) continue; - if( patternCompare(zPattern,zString,pInfo,esc) ) return 1; - } - } - return 0; - } - if( c==matchOther ){ - if( esc ){ - c = sqlite3Utf8Read(&zPattern); - if( c==0 ) return 0; - zEscaped = zPattern; - }else{ - u32 prior_c = 0; - int seen = 0; - int invert = 0; - c = sqlite3Utf8Read(&zString); - if( c==0 ) return 0; - c2 = sqlite3Utf8Read(&zPattern); - if( c2=='^' ){ - invert = 1; - c2 = sqlite3Utf8Read(&zPattern); - } - if( c2==']' ){ - if( c==']' ) seen = 1; - c2 = sqlite3Utf8Read(&zPattern); - } - while( c2 && c2!=']' ){ - if( c2=='-' && zPattern[0]!=']' && zPattern[0]!=0 && prior_c>0 ){ - c2 = sqlite3Utf8Read(&zPattern); - if( c>=prior_c && c<=c2 ) seen = 1; - prior_c = 0; - }else{ - if( c==c2 ){ - seen = 1; - } - prior_c = c2; - } - c2 = sqlite3Utf8Read(&zPattern); - } - if( c2==0 || (seen ^ invert)==0 ){ - return 0; - } - continue; - } - } - c2 = sqlite3Utf8Read(&zString); - if( c==c2 ) continue; - if( noCase && c<0x80 && c2<0x80 && sqlite3Tolower(c)==sqlite3Tolower(c2) ){ - continue; - } - if( c==matchOne && zPattern!=zEscaped && c2!=0 ) continue; - return 0; + return 1; + }else if( c==esc ){ + c = sqlite3Utf8Read(&zPattern); + if( c==0 ){ + return 0; + } + }else if( c==matchSet ){ + assert( esc==0 ); /* This is GLOB, not LIKE */ + assert( matchSet<0x80 ); /* '[' is a single-byte character */ + while( *zString && patternCompare(&zPattern[-1],zString,pInfo,esc)==0 ){ + SQLITE_SKIP_UTF8(zString); + } + return *zString!=0; + } + while( (c2 = sqlite3Utf8Read(&zString))!=0 ){ + if( noCase ){ + GlobUpperToLower(c2); + GlobUpperToLower(c); + while( c2 != 0 && c2 != c ){ + c2 = sqlite3Utf8Read(&zString); + GlobUpperToLower(c2); + } + }else{ + while( c2 != 0 && c2 != c ){ + c2 = sqlite3Utf8Read(&zString); + } + } + if( c2==0 ) return 0; + if( patternCompare(zPattern,zString,pInfo,esc) ) return 1; + } + return 0; + }else if( c==matchOne && !prevEscape ){ + if( sqlite3Utf8Read(&zString)==0 ){ + return 0; + } + }else if( c==matchSet ){ + u32 prior_c = 0; + assert( esc==0 ); /* This only occurs for GLOB, not LIKE */ + seen = 0; + invert = 0; + c = sqlite3Utf8Read(&zString); + if( c==0 ) return 0; + c2 = sqlite3Utf8Read(&zPattern); + if( c2=='^' ){ + invert = 1; + c2 = sqlite3Utf8Read(&zPattern); + } + if( c2==']' ){ + if( c==']' ) seen = 1; + c2 = sqlite3Utf8Read(&zPattern); + } + while( c2 && c2!=']' ){ + if( c2=='-' && zPattern[0]!=']' && zPattern[0]!=0 && prior_c>0 ){ + c2 = sqlite3Utf8Read(&zPattern); + if( c>=prior_c && c<=c2 ) seen = 1; + prior_c = 0; + }else{ + if( c==c2 ){ + seen = 1; + } + prior_c = c2; + } + c2 = sqlite3Utf8Read(&zPattern); + } + if( c2==0 || (seen ^ invert)==0 ){ + return 0; + } + }else if( esc==c && !prevEscape ){ + prevEscape = 1; + }else{ + c2 = sqlite3Utf8Read(&zString); + if( noCase ){ + GlobUpperToLower(c); + GlobUpperToLower(c2); + } + if( c!=c2 ){ + return 0; + } + prevEscape = 0; + } } return *zString==0; } /* @@ -95157,11 +92709,11 @@ *zOut++ = 0x80 + (u8)((c>>12) & 0x3F); *zOut++ = 0x80 + (u8)((c>>6) & 0x3F); *zOut++ = 0x80 + (u8)(c & 0x3F); } \ } - sqlite3_result_text64(context, (char*)z, zOut-z, sqlite3_free, SQLITE_UTF8); + sqlite3_result_text(context, (char*)z, (int)(zOut-z), sqlite3_free); } /* ** The hex() function. Interpret the argument as a blob. Return ** a hexadecimal rendering as text. @@ -95607,11 +93159,10 @@ sqlite3VdbeMemCopy(pBest, pArg); }else{ sqlite3SkipAccumulatorLoad(context); } }else{ - pBest->db = sqlite3_context_db_handle(context); sqlite3VdbeMemCopy(pBest, pArg); } } static void minMaxFinalize(sqlite3_context *context){ sqlite3_value *pRes; @@ -95755,11 +93306,11 @@ *pIsNocase = (pDef->funcFlags & SQLITE_FUNC_CASE)==0; return 1; } /* -** All of the FuncDef structures in the aBuiltinFunc[] array above +** All all of the FuncDef structures in the aBuiltinFunc[] array above ** to the global function hash table. This occurs at start-time (as ** a consequence of calling sqlite3_initialize()). ** ** After this routine runs */ @@ -95779,16 +93330,14 @@ FUNCTION(rtrim, 2, 2, 0, trimFunc ), FUNCTION(trim, 1, 3, 0, trimFunc ), FUNCTION(trim, 2, 3, 0, trimFunc ), FUNCTION(min, -1, 0, 1, minmaxFunc ), FUNCTION(min, 0, 0, 1, 0 ), - AGGREGATE2(min, 1, 0, 1, minmaxStep, minMaxFinalize, - SQLITE_FUNC_MINMAX ), + AGGREGATE(min, 1, 0, 1, minmaxStep, minMaxFinalize ), FUNCTION(max, -1, 1, 1, minmaxFunc ), FUNCTION(max, 0, 1, 1, 0 ), - AGGREGATE2(max, 1, 1, 1, minmaxStep, minMaxFinalize, - SQLITE_FUNC_MINMAX ), + AGGREGATE(max, 1, 1, 1, minmaxStep, minMaxFinalize ), FUNCTION2(typeof, 1, 0, 0, typeofFunc, SQLITE_FUNC_TYPEOF), FUNCTION2(length, 1, 0, 0, lengthFunc, SQLITE_FUNC_LENGTH), FUNCTION(instr, 2, 0, 0, instrFunc ), FUNCTION(substr, 2, 0, 0, substrFunc ), FUNCTION(substr, 3, 0, 0, substrFunc ), @@ -95814,13 +93363,10 @@ VFUNCTION(randomblob, 1, 0, 0, randomBlob ), FUNCTION(nullif, 2, 0, 1, nullifFunc ), FUNCTION(sqlite_version, 0, 0, 0, versionFunc ), FUNCTION(sqlite_source_id, 0, 0, 0, sourceidFunc ), FUNCTION(sqlite_log, 2, 0, 0, errlogFunc ), -#if SQLITE_USER_AUTHENTICATION - FUNCTION(sqlite_crypt, 2, 0, 0, sqlite3CryptFunc ), -#endif #ifndef SQLITE_OMIT_COMPILEOPTION_DIAGS FUNCTION(sqlite_compileoption_used,1, 0, 0, compileoptionusedFunc ), FUNCTION(sqlite_compileoption_get, 1, 0, 0, compileoptiongetFunc ), #endif /* SQLITE_OMIT_COMPILEOPTION_DIAGS */ FUNCTION(quote, 1, 0, 0, quoteFunc ), @@ -95837,12 +93383,12 @@ FUNCTION(load_extension, 2, 0, 0, loadExt ), #endif AGGREGATE(sum, 1, 0, 0, sumStep, sumFinalize ), AGGREGATE(total, 1, 0, 0, sumStep, totalFinalize ), AGGREGATE(avg, 1, 0, 0, sumStep, avgFinalize ), - AGGREGATE2(count, 0, 0, 0, countStep, countFinalize, - SQLITE_FUNC_COUNT ), + /* AGGREGATE(count, 0, 0, 0, countStep, countFinalize ), */ + {0,SQLITE_UTF8|SQLITE_FUNC_COUNT,0,0,0,countStep,countFinalize,"count",0,0}, AGGREGATE(count, 1, 0, 0, countStep, countFinalize ), AGGREGATE(group_concat, 1, 0, 0, groupConcatStep, groupConcatFinalize), AGGREGATE(group_concat, 2, 0, 0, groupConcatStep, groupConcatFinalize), LIKEFUNC(glob, 2, &globInfo, SQLITE_FUNC_LIKE|SQLITE_FUNC_CASE), @@ -96045,11 +93591,11 @@ ** foreign key definition, and the parent table does not have a ** PRIMARY KEY, or ** ** 4) No parent key columns were provided explicitly as part of the ** foreign key definition, and the PRIMARY KEY of the parent table -** consists of a different number of columns to the child key in +** consists of a a different number of columns to the child key in ** the child table. ** ** then non-zero is returned, and a "foreign key mismatch" error loaded ** into pParse. If an OOM error occurs, non-zero is returned and the ** pParse->db->mallocFailed flag is set. @@ -96531,11 +94077,12 @@ ** "t2". Calling this function with "t2" as the argument would return a ** NULL pointer (as there are no FK constraints for which t2 is the parent ** table). */ SQLITE_PRIVATE FKey *sqlite3FkReferences(Table *pTab){ - return (FKey *)sqlite3HashFind(&pTab->pSchema->fkeyHash, pTab->zName); + int nName = sqlite3Strlen30(pTab->zName); + return (FKey *)sqlite3HashFind(&pTab->pSchema->fkeyHash, pTab->zName, nName); } /* ** The second argument is a Trigger structure allocated by the ** fkActionTrigger() routine. This function deletes the Trigger structure @@ -97209,11 +94756,11 @@ if( pFKey->pPrevTo ){ pFKey->pPrevTo->pNextTo = pFKey->pNextTo; }else{ void *p = (void *)pFKey->pNextTo; const char *z = (p ? pFKey->pNextTo->zTo : pFKey->zTo); - sqlite3HashInsert(&pTab->pSchema->fkeyHash, z, p); + sqlite3HashInsert(&pTab->pSchema->fkeyHash, z, sqlite3Strlen30(z), p); } if( pFKey->pNextTo ){ pFKey->pNextTo->pPrevTo = pFKey->pPrevTo; } } @@ -97292,17 +94839,17 @@ ** pIdx. A column affinity string has one character for each column in ** the table, according to the affinity of the column: ** ** Character Column affinity ** ------------------------------ -** 'A' NONE -** 'B' TEXT -** 'C' NUMERIC -** 'D' INTEGER -** 'F' REAL +** 'a' TEXT +** 'b' NONE +** 'c' NUMERIC +** 'd' INTEGER +** 'e' REAL ** -** An extra 'D' is appended to the end of the string to cover the +** An extra 'd' is appended to the end of the string to cover the ** rowid that appears as the last column in every index. ** ** Memory for the buffer containing the column index affinity string ** is managed along with the rest of the Index structure. It will be ** released when sqlite3DeleteIndex() is called. @@ -97347,15 +94894,15 @@ ** ** A column affinity string has one character per column: ** ** Character Column affinity ** ------------------------------ -** 'A' NONE -** 'B' TEXT -** 'C' NUMERIC -** 'D' INTEGER -** 'E' REAL +** 'a' TEXT +** 'b' NONE +** 'c' NUMERIC +** 'd' INTEGER +** 'e' REAL */ SQLITE_PRIVATE void sqlite3TableAffinity(Vdbe *v, Table *pTab, int iReg){ int i; char *zColAff = pTab->zColAff; if( zColAff==0 ){ @@ -97646,11 +95193,11 @@ ** D: cleanup ** ** The 4th template is used if the insert statement takes its ** values from a SELECT but the data is being inserted into a table ** that is also read as part of the SELECT. In the third form, -** we have to use an intermediate table to store the results of +** we have to use a intermediate table to store the results of ** the select. The template is like this: ** ** X <- A ** goto B ** A: setup for the SELECT @@ -97811,11 +95358,11 @@ ** sqlite_sequence table and store it in memory cell regAutoinc. */ regAutoinc = autoIncBegin(pParse, iDb, pTab); /* Allocate registers for holding the rowid of the new row, - ** the content of the new row, and the assembled row record. + ** the content of the new row, and the assemblied row record. */ regRowid = regIns = pParse->nMem+1; pParse->nMem += pTab->nCol + 1; if( IsVirtual(pTab) ){ regRowid++; @@ -98263,11 +95810,11 @@ sqlite3IdListDelete(db, pColumn); sqlite3DbFree(db, aRegIdx); } /* Make sure "isView" and other macros defined above are undefined. Otherwise -** they may interfere with compilation of other functions in this file +** thely may interfere with compilation of other functions in this file ** (or in another file, if this file becomes part of the amalgamation). */ #ifdef isView #undef isView #endif #ifdef pTrigger @@ -98379,11 +95926,11 @@ sqlite3 *db; /* Database connection */ int i; /* loop counter */ int ix; /* Index loop counter */ int nCol; /* Number of columns */ int onError; /* Conflict resolution strategy */ - int j1; /* Address of jump instruction */ + int j1; /* Addresss of jump instruction */ int seenReplace = 0; /* True if REPLACE is used to resolve INT PK conflict */ int nPkField; /* Number of fields in PRIMARY KEY. 1 for ROWID tables */ int ipkTop = 0; /* Top of the rowid change constraint check */ int ipkBottom = 0; /* Bottom of the rowid change constraint check */ u8 isUpdate; /* True if this is an UPDATE operation */ @@ -98783,11 +96330,11 @@ ){ Vdbe *v; /* Prepared statements under construction */ Index *pIdx; /* An index being inserted or updated */ u8 pik_flags; /* flag values passed to the btree insert */ int regData; /* Content registers (after the rowid) */ - int regRec; /* Register holding assembled record for the table */ + int regRec; /* Register holding assemblied record for the table */ int i; /* Loop counter */ u8 bAffinityDone = 0; /* True if OP_Affinity has been run already */ v = sqlite3GetVdbe(pParse); assert( v!=0 ); @@ -98848,13 +96395,10 @@ ** ** For a rowid table, *piDataCur will be exactly one less than *piIdxCur. ** For a WITHOUT ROWID table, *piDataCur will be somewhere in the range ** of *piIdxCurs, depending on where the PRIMARY KEY index appears on the ** pTab->pIndex list. -** -** If pTab is a virtual table, then this routine is a no-op and the -** *piDataCur and *piIdxCur values are left uninitialized. */ SQLITE_PRIVATE int sqlite3OpenTableAndIndices( Parse *pParse, /* Parsing context */ Table *pTab, /* Table to be opened */ int op, /* OP_OpenRead or OP_OpenWrite */ @@ -98869,13 +96413,13 @@ Index *pIdx; Vdbe *v; assert( op==OP_OpenRead || op==OP_OpenWrite ); if( IsVirtual(pTab) ){ - /* This routine is a no-op for virtual tables. Leave the output - ** variables *piDataCur and *piIdxCur uninitialized so that valgrind - ** can detect if they are used by mistake in the caller. */ + assert( aToOpen==0 ); + *piDataCur = 0; + *piIdxCur = 1; return 0; } iDb = sqlite3SchemaToIndex(pParse->db, pTab->pSchema); v = sqlite3GetVdbe(pParse); assert( v!=0 ); @@ -98908,11 +96452,11 @@ #ifdef SQLITE_TEST /* ** The following global variable is incremented whenever the ** transfer optimization is used. This is used for testing ** purposes only - to make sure the transfer optimization really -** is happening when it is supposed to. +** is happening when it is suppose to. */ SQLITE_API int sqlite3_xferopt_count; #endif /* SQLITE_TEST */ @@ -98975,11 +96519,11 @@ ** Attempt the transfer optimization on INSERTs of the form ** ** INSERT INTO tab1 SELECT * FROM tab2; ** ** The xfer optimization transfers raw records from tab2 over to tab1. -** Columns are not decoded and reassembled, which greatly improves +** Columns are not decoded and reassemblied, which greatly improves ** performance. Raw index records are transferred in the same way. ** ** The xfer optimization is only attempted if tab1 and tab2 are compatible. ** There are lots of rules for determining compatibility - see comments ** embedded in the code for details. @@ -99303,11 +96847,11 @@ if( !sqlite3SafetyCheckOk(db) ) return SQLITE_MISUSE_BKPT; if( zSql==0 ) zSql = ""; sqlite3_mutex_enter(db->mutex); - sqlite3Error(db, SQLITE_OK); + sqlite3Error(db, SQLITE_OK, 0); while( rc==SQLITE_OK && zSql[0] ){ int nCol; char **azVals = 0; pStmt = 0; @@ -99361,11 +96905,11 @@ ** sqlite3_exec() returns non-zero, then sqlite3_exec() will ** return SQLITE_ABORT. */ rc = SQLITE_ABORT; sqlite3VdbeFinalize((Vdbe *)pStmt); pStmt = 0; - sqlite3Error(db, SQLITE_ABORT); + sqlite3Error(db, SQLITE_ABORT, 0); goto exec_out; } } if( rc!=SQLITE_ROW ){ @@ -99384,18 +96928,18 @@ exec_out: if( pStmt ) sqlite3VdbeFinalize((Vdbe *)pStmt); sqlite3DbFree(db, azCols); rc = sqlite3ApiExit(db, rc); - if( rc!=SQLITE_OK && pzErrMsg ){ + if( rc!=SQLITE_OK && ALWAYS(rc==sqlite3_errcode(db)) && pzErrMsg ){ int nErrMsg = 1 + sqlite3Strlen30(sqlite3_errmsg(db)); *pzErrMsg = sqlite3Malloc(nErrMsg); if( *pzErrMsg ){ memcpy(*pzErrMsg, sqlite3_errmsg(db), nErrMsg); }else{ rc = SQLITE_NOMEM; - sqlite3Error(db, SQLITE_NOMEM); + sqlite3Error(db, SQLITE_NOMEM, 0); } }else if( pzErrMsg ){ *pzErrMsg = 0; } @@ -99453,11 +96997,11 @@ ** routines. ** ** WARNING: In order to maintain backwards compatibility, add new ** interfaces to the end of this structure only. If you insert new ** interfaces in the middle of this structure, then older different -** versions of SQLite will not be able to load each other's shared +** versions of SQLite will not be able to load each others' shared ** libraries! */ struct sqlite3_api_routines { void * (*aggregate_context)(sqlite3_context*,int nBytes); int (*aggregate_count)(sqlite3_context*); @@ -99675,32 +97219,15 @@ int (*uri_boolean)(const char*,const char*,int); sqlite3_int64 (*uri_int64)(const char*,const char*,sqlite3_int64); const char *(*uri_parameter)(const char*,const char*); char *(*vsnprintf)(int,char*,const char*,va_list); int (*wal_checkpoint_v2)(sqlite3*,const char*,int,int*,int*); - /* Version 3.8.7 and later */ - int (*auto_extension)(void(*)(void)); - int (*bind_blob64)(sqlite3_stmt*,int,const void*,sqlite3_uint64, - void(*)(void*)); - int (*bind_text64)(sqlite3_stmt*,int,const char*,sqlite3_uint64, - void(*)(void*),unsigned char); - int (*cancel_auto_extension)(void(*)(void)); - int (*load_extension)(sqlite3*,const char*,const char*,char**); - void *(*malloc64)(sqlite3_uint64); - sqlite3_uint64 (*msize)(void*); - void *(*realloc64)(void*,sqlite3_uint64); - void (*reset_auto_extension)(void); - void (*result_blob64)(sqlite3_context*,const void*,sqlite3_uint64, - void(*)(void*)); - void (*result_text64)(sqlite3_context*,const char*,sqlite3_uint64, - void(*)(void*), unsigned char); - int (*strglob)(const char*,const char*); }; /* ** The following macros redefine the API routines so that they are -** redirected through the global sqlite3_api structure. +** redirected throught the global sqlite3_api structure. ** ** This header file is also used by the loadext.c source file ** (part of the main SQLite library - not an extension) so that ** it can get access to the sqlite3_api_routines structure ** definition. But the main library does not want to redefine @@ -99909,23 +97436,10 @@ #define sqlite3_uri_boolean sqlite3_api->uri_boolean #define sqlite3_uri_int64 sqlite3_api->uri_int64 #define sqlite3_uri_parameter sqlite3_api->uri_parameter #define sqlite3_uri_vsnprintf sqlite3_api->vsnprintf #define sqlite3_wal_checkpoint_v2 sqlite3_api->wal_checkpoint_v2 -/* Version 3.8.7 and later */ -#define sqlite3_auto_extension sqlite3_api->auto_extension -#define sqlite3_bind_blob64 sqlite3_api->bind_blob64 -#define sqlite3_bind_text64 sqlite3_api->bind_text64 -#define sqlite3_cancel_auto_extension sqlite3_api->cancel_auto_extension -#define sqlite3_load_extension sqlite3_api->load_extension -#define sqlite3_malloc64 sqlite3_api->malloc64 -#define sqlite3_msize sqlite3_api->msize -#define sqlite3_realloc64 sqlite3_api->realloc64 -#define sqlite3_reset_auto_extension sqlite3_api->reset_auto_extension -#define sqlite3_result_blob64 sqlite3_api->result_blob64 -#define sqlite3_result_text64 sqlite3_api->result_text64 -#define sqlite3_strglob sqlite3_api->strglob #endif /* SQLITE_CORE */ #ifndef SQLITE_CORE /* This case when the file really is being compiled as a loadable ** extension */ @@ -100315,24 +97829,11 @@ sqlite3_stricmp, sqlite3_uri_boolean, sqlite3_uri_int64, sqlite3_uri_parameter, sqlite3_vsnprintf, - sqlite3_wal_checkpoint_v2, - /* Version 3.8.7 and later */ - sqlite3_auto_extension, - sqlite3_bind_blob64, - sqlite3_bind_text64, - sqlite3_cancel_auto_extension, - sqlite3_load_extension, - sqlite3_malloc64, - sqlite3_msize, - sqlite3_realloc64, - sqlite3_reset_auto_extension, - sqlite3_result_blob64, - sqlite3_result_text64, - sqlite3_strglob + sqlite3_wal_checkpoint_v2 }; /* ** Attempt to load an SQLite extension library contained in the file ** zFile. The entry point is zProc. zProc may be 0 in which case a @@ -100687,11 +98188,11 @@ wsdAutoext.aExt[i]; } sqlite3_mutex_leave(mutex); zErrmsg = 0; if( xInit && (rc = xInit(db, &zErrmsg, &sqlite3Apis))!=0 ){ - sqlite3ErrorWithMsg(db, rc, + sqlite3Error(db, rc, "automatic extension loading failed: %s", zErrmsg); go = 0; } sqlite3_free(zErrmsg); } @@ -100759,19 +98260,18 @@ #define PragTyp_STATS 28 #define PragTyp_SYNCHRONOUS 29 #define PragTyp_TABLE_INFO 30 #define PragTyp_TEMP_STORE 31 #define PragTyp_TEMP_STORE_DIRECTORY 32 -#define PragTyp_THREADS 33 -#define PragTyp_WAL_AUTOCHECKPOINT 34 -#define PragTyp_WAL_CHECKPOINT 35 -#define PragTyp_ACTIVATE_EXTENSIONS 36 -#define PragTyp_HEXKEY 37 -#define PragTyp_KEY 38 -#define PragTyp_REKEY 39 -#define PragTyp_LOCK_STATUS 40 -#define PragTyp_PARSER_TRACE 41 +#define PragTyp_WAL_AUTOCHECKPOINT 33 +#define PragTyp_WAL_CHECKPOINT 34 +#define PragTyp_ACTIVATE_EXTENSIONS 35 +#define PragTyp_HEXKEY 36 +#define PragTyp_KEY 37 +#define PragTyp_REKEY 38 +#define PragTyp_LOCK_STATUS 39 +#define PragTyp_PARSER_TRACE 40 #define PragFlag_NeedSchema 0x01 static const struct sPragmaNames { const char *const zName; /* Name of pragma */ u8 ePragTyp; /* PragTyp_XXX value */ u8 mPragFlag; /* Zero or more PragFlag_XXX values */ @@ -101117,14 +98617,10 @@ { /* zName: */ "temp_store_directory", /* ePragTyp: */ PragTyp_TEMP_STORE_DIRECTORY, /* ePragFlag: */ 0, /* iArg: */ 0 }, #endif - { /* zName: */ "threads", - /* ePragTyp: */ PragTyp_THREADS, - /* ePragFlag: */ 0, - /* iArg: */ 0 }, #if !defined(SQLITE_OMIT_SCHEMA_VERSION_PRAGMAS) { /* zName: */ "user_version", /* ePragTyp: */ PragTyp_HEADER_VALUE, /* ePragFlag: */ 0, /* iArg: */ 0 }, @@ -101168,11 +98664,11 @@ /* ePragTyp: */ PragTyp_FLAG, /* ePragFlag: */ 0, /* iArg: */ SQLITE_WriteSchema|SQLITE_RecoveryMode }, #endif }; -/* Number of pragmas: 57 on by default, 70 total. */ +/* Number of pragmas: 56 on by default, 69 total. */ /* End of the automatically generated pragma table. ***************************************************************************/ /* ** Interpret the given string as a safety level. Return 0 for OFF, @@ -102095,16 +99591,10 @@ if( db->autoCommit==0 ){ /* Foreign key support may not be enabled or disabled while not ** in auto-commit mode. */ mask &= ~(SQLITE_ForeignKeys); } -#if SQLITE_USER_AUTHENTICATION - if( db->auth.authLevel==UAUTH_User ){ - /* Do not allow non-admin users to modify the schema arbitrarily */ - mask &= ~(SQLITE_WriteSchema); - } -#endif if( sqlite3GetBoolean(zRight, 0) ){ db->flags |= mask; }else{ db->flags &= ~mask; @@ -102982,30 +100472,10 @@ } returnSingleInt(pParse, "soft_heap_limit", sqlite3_soft_heap_limit64(-1)); break; } - /* - ** PRAGMA threads - ** PRAGMA threads = N - ** - ** Configure the maximum number of worker threads. Return the new - ** maximum, which might be less than requested. - */ - case PragTyp_THREADS: { - sqlite3_int64 N; - if( zRight - && sqlite3DecOrHexToI64(zRight, &N)==SQLITE_OK - && N>=0 - ){ - sqlite3_limit(db, SQLITE_LIMIT_WORKER_THREADS, (int)(N&0x7fffffff)); - } - returnSingleInt(pParse, "threads", - sqlite3_limit(db, SQLITE_LIMIT_WORKER_THREADS, -1)); - break; - } - #if defined(SQLITE_DEBUG) || defined(SQLITE_TEST) /* ** Report the current state of file logs for all databases */ case PragTyp_LOCK_STATUS: { @@ -103418,11 +100888,11 @@ zSql = sqlite3MPrintf(db, "SELECT name, rootpage, sql FROM '%q'.%s ORDER BY rowid", db->aDb[iDb].zName, zMasterName); #ifndef SQLITE_OMIT_AUTHORIZATION { - sqlite3_xauth xAuth; + int (*xAuth)(void*,int,const char*,const char*,const char*,const char*); xAuth = db->xAuth; db->xAuth = 0; #endif rc = sqlite3_exec(db, zSql, sqlite3InitCallback, &initData, 0); #ifndef SQLITE_OMIT_AUTHORIZATION @@ -103484,11 +100954,10 @@ SQLITE_PRIVATE int sqlite3Init(sqlite3 *db, char **pzErrMsg){ int i, rc; int commit_internal = !(db->flags&SQLITE_InternChanges); assert( sqlite3_mutex_held(db->mutex) ); - assert( db->init.busy==0 ); rc = SQLITE_OK; db->init.busy = 1; for(i=0; rc==SQLITE_OK && i nDb; i++){ if( DbHasProperty(db, i, DB_SchemaLoaded) || i==1 ) continue; rc = sqlite3InitOne(db, i, pzErrMsg); @@ -103500,12 +100969,12 @@ /* Once all the other databases have been initialized, load the schema ** for the TEMP database. This is loaded last, as the TEMP database ** schema may contain references to objects in other databases. */ #ifndef SQLITE_OMIT_TEMPDB - assert( db->nDb>1 ); - if( rc==SQLITE_OK && !DbHasProperty(db, 1, DB_SchemaLoaded) ){ + if( rc==SQLITE_OK && ALWAYS(db->nDb>1) + && !DbHasProperty(db, 1, DB_SchemaLoaded) ){ rc = sqlite3InitOne(db, 1, pzErrMsg); if( rc ){ sqlite3ResetOneSchema(db, 1); } } @@ -103684,11 +101153,11 @@ if( pBt ){ assert( sqlite3BtreeHoldsMutex(pBt) ); rc = sqlite3BtreeSchemaLocked(pBt); if( rc ){ const char *zDb = db->aDb[i].zName; - sqlite3ErrorWithMsg(db, rc, "database schema is locked: %s", zDb); + sqlite3Error(db, rc, "database schema is locked: %s", zDb); testcase( db->flags & SQLITE_ReadUncommitted ); goto end_prepare; } } } @@ -103701,11 +101170,11 @@ char *zSqlCopy; int mxLen = db->aLimit[SQLITE_LIMIT_SQL_LENGTH]; testcase( nBytes==mxLen ); testcase( nBytes==mxLen+1 ); if( nBytes>mxLen ){ - sqlite3ErrorWithMsg(db, SQLITE_TOOBIG, "statement too long"); + sqlite3Error(db, SQLITE_TOOBIG, "statement too long"); rc = sqlite3ApiExit(db, SQLITE_TOOBIG); goto end_prepare; } zSqlCopy = sqlite3DbStrNDup(db, zSql, nBytes); if( zSqlCopy ){ @@ -103768,14 +101237,14 @@ }else{ *ppStmt = (sqlite3_stmt*)pParse->pVdbe; } if( zErrMsg ){ - sqlite3ErrorWithMsg(db, rc, "%s", zErrMsg); + sqlite3Error(db, rc, "%s", zErrMsg); sqlite3DbFree(db, zErrMsg); }else{ - sqlite3Error(db, rc); + sqlite3Error(db, rc, 0); } /* Delete any TriggerPrg structures allocated while parsing this statement. */ while( pParse->pTriggerPrg ){ TriggerPrg *pT = pParse->pTriggerPrg; @@ -103991,24 +101460,10 @@ ** ************************************************************************* ** This file contains C code routines that are called by the parser ** to handle SELECT statements in SQLite. */ - -/* -** Trace output macros -*/ -#if SELECTTRACE_ENABLED -/***/ int sqlite3SelectTrace = 0; -# define SELECTTRACE(K,P,S,X) \ - if(sqlite3SelectTrace&(K)) \ - sqlite3DebugPrintf("%*s%s.%p: ",(P)->nSelectIndent*2-2,"",(S)->zSelName,(S)),\ - sqlite3DebugPrintf X -#else -# define SELECTTRACE(K,P,S,X) -#endif - /* ** An instance of the following object is used to record information about ** how to process the DISTINCT keyword, to simplify passing that information ** into the selectInnerLoop() routine. @@ -104117,22 +101572,10 @@ assert( pNew->pSrc!=0 || pParse->nErr>0 ); } assert( pNew!=&standin ); return pNew; } - -#if SELECTTRACE_ENABLED -/* -** Set the name of a Select object -*/ -SQLITE_PRIVATE void sqlite3SelectSetName(Select *p, const char *zName){ - if( p && zName ){ - sqlite3_snprintf(sizeof(p->zSelName), p->zSelName, "%s", zName); - } -} -#endif - /* ** Delete the given Select structure and all of its substructures. */ SQLITE_PRIVATE void sqlite3SelectDelete(sqlite3 *db, Select *p){ @@ -104459,47 +101902,32 @@ int iStart, /* Begin with this column of pList */ int nExtra /* Add this many extra columns to the end */ ); /* -** Generate code that will push the record in registers regData -** through regData+nData-1 onto the sorter. +** Insert code into "v" that will push the record in register regData +** into the sorter. */ static void pushOntoSorter( Parse *pParse, /* Parser context */ SortCtx *pSort, /* Information about the ORDER BY clause */ Select *pSelect, /* The whole SELECT statement */ - int regData, /* First register holding data to be sorted */ - int nData, /* Number of elements in the data array */ - int nPrefixReg /* No. of reg prior to regData available for use */ + int regData /* Register holding data to be sorted */ ){ - Vdbe *v = pParse->pVdbe; /* Stmt under construction */ - int bSeq = ((pSort->sortFlags & SORTFLAG_UseSorter)==0); - int nExpr = pSort->pOrderBy->nExpr; /* No. of ORDER BY terms */ - int nBase = nExpr + bSeq + nData; /* Fields in sorter record */ - int regBase; /* Regs for sorter record */ - int regRecord = ++pParse->nMem; /* Assembled sorter record */ - int nOBSat = pSort->nOBSat; /* ORDER BY terms to skip */ - int op; /* Opcode to add sorter record to sorter */ - - assert( bSeq==0 || bSeq==1 ); - if( nPrefixReg ){ - assert( nPrefixReg==nExpr+bSeq ); - regBase = regData - nExpr - bSeq; - }else{ - regBase = pParse->nMem + 1; - pParse->nMem += nBase; - } - sqlite3ExprCodeExprList(pParse, pSort->pOrderBy, regBase, SQLITE_ECEL_DUP); - if( bSeq ){ - sqlite3VdbeAddOp2(v, OP_Sequence, pSort->iECursor, regBase+nExpr); - } - if( nPrefixReg==0 ){ - sqlite3ExprCodeMove(pParse, regData, regBase+nExpr+bSeq, nData); - } - - sqlite3VdbeAddOp3(v, OP_MakeRecord, regBase+nOBSat, nBase-nOBSat, regRecord); + Vdbe *v = pParse->pVdbe; + int nExpr = pSort->pOrderBy->nExpr; + int regRecord = ++pParse->nMem; + int regBase = pParse->nMem+1; + int nOBSat = pSort->nOBSat; + int op; + + pParse->nMem += nExpr+2; /* nExpr+2 registers allocated at regBase */ + sqlite3ExprCacheClear(pParse); + sqlite3ExprCodeExprList(pParse, pSort->pOrderBy, regBase, 0); + sqlite3VdbeAddOp2(v, OP_Sequence, pSort->iECursor, regBase+nExpr); + sqlite3ExprCodeMove(pParse, regData, regBase+nExpr+1, 1); + sqlite3VdbeAddOp3(v, OP_MakeRecord, regBase+nOBSat, nExpr+2-nOBSat,regRecord); if( nOBSat>0 ){ int regPrevKey; /* The first nOBSat columns of the previous row */ int addrFirst; /* Address of the OP_IfNot opcode */ int addrJmp; /* Address of the OP_Jump opcode */ VdbeOp *pOp; /* Opcode that opens the sorter */ @@ -104506,21 +101934,16 @@ int nKey; /* Number of sorting key columns, including OP_Sequence */ KeyInfo *pKI; /* Original KeyInfo on the sorter table */ regPrevKey = pParse->nMem+1; pParse->nMem += pSort->nOBSat; - nKey = nExpr - pSort->nOBSat + bSeq; - if( bSeq ){ - addrFirst = sqlite3VdbeAddOp1(v, OP_IfNot, regBase+nExpr); - }else{ - addrFirst = sqlite3VdbeAddOp1(v, OP_SequenceTest, pSort->iECursor); - } - VdbeCoverage(v); + nKey = nExpr - pSort->nOBSat + 1; + addrFirst = sqlite3VdbeAddOp1(v, OP_IfNot, regBase+nExpr); VdbeCoverage(v); sqlite3VdbeAddOp3(v, OP_Compare, regPrevKey, regBase, pSort->nOBSat); pOp = sqlite3VdbeGetOp(v, pSort->addrSortIndex); if( pParse->db->mallocFailed ) return; - pOp->p2 = nKey + nData; + pOp->p2 = nKey + 1; pKI = pOp->p4.pKeyInfo; memset(pKI->aSortOrder, 0, pKI->nField); /* Makes OP_Jump below testable */ sqlite3VdbeChangeP4(v, -1, (char*)pKI, P4_KEYINFO); pOp->p4.pKeyInfo = keyInfoFromExprList(pParse, pSort->pOrderBy, nOBSat, 1); addrJmp = sqlite3VdbeCurrentAddr(v); @@ -104528,11 +101951,11 @@ pSort->labelBkOut = sqlite3VdbeMakeLabel(v); pSort->regReturn = ++pParse->nMem; sqlite3VdbeAddOp2(v, OP_Gosub, pSort->regReturn, pSort->labelBkOut); sqlite3VdbeAddOp1(v, OP_ResetSorter, pSort->iECursor); sqlite3VdbeJumpHere(v, addrFirst); - sqlite3ExprCodeMove(pParse, regBase, regPrevKey, pSort->nOBSat); + sqlite3VdbeAddOp3(v, OP_Move, regBase, regPrevKey, pSort->nOBSat); sqlite3VdbeJumpHere(v, addrJmp); } if( pSort->sortFlags & SORTFLAG_UseSorter ){ op = OP_SorterInsert; }else{ @@ -104650,11 +102073,10 @@ int hasDistinct; /* True if the DISTINCT keyword is present */ int regResult; /* Start of memory holding result set */ int eDest = pDest->eDest; /* How to dispose of results */ int iParm = pDest->iSDParm; /* First argument to disposal method */ int nResultCol; /* Number of result columns */ - int nPrefixReg = 0; /* Number of extra registers before regResult */ assert( v ); assert( pEList!=0 ); hasDistinct = pDistinct ? pDistinct->eTnctType : WHERE_DISTINCT_NOOP; if( pSort && pSort->pOrderBy==0 ) pSort = 0; @@ -104666,15 +102088,10 @@ /* Pull the requested columns. */ nResultCol = pEList->nExpr; if( pDest->iSdst==0 ){ - if( pSort ){ - nPrefixReg = pSort->pOrderBy->nExpr; - if( !(pSort->sortFlags & SORTFLAG_UseSorter) ) nPrefixReg++; - pParse->nMem += nPrefixReg; - } pDest->iSdst = pParse->nMem+1; pParse->nMem += nResultCol; }else if( pDest->iSdst+nResultCol > pParse->nMem ){ /* This is an error condition that can result, for example, when a SELECT ** on the right-hand side of an INSERT contains more result columns than @@ -104787,14 +102204,14 @@ */ case SRT_Fifo: case SRT_DistFifo: case SRT_Table: case SRT_EphemTab: { - int r1 = sqlite3GetTempRange(pParse, nPrefixReg+1); + int r1 = sqlite3GetTempReg(pParse); testcase( eDest==SRT_Table ); testcase( eDest==SRT_EphemTab ); - sqlite3VdbeAddOp3(v, OP_MakeRecord, regResult, nResultCol, r1+nPrefixReg); + sqlite3VdbeAddOp3(v, OP_MakeRecord, regResult, nResultCol, r1); #ifndef SQLITE_OMIT_CTE if( eDest==SRT_DistFifo ){ /* If the destination is DistFifo, then cursor (iParm+1) is open ** on an ephemeral index. If the current row is already present ** in the index, do not write it to the output. If not, add the @@ -104805,19 +102222,19 @@ sqlite3VdbeAddOp2(v, OP_IdxInsert, iParm+1, r1); assert( pSort==0 ); } #endif if( pSort ){ - pushOntoSorter(pParse, pSort, p, r1+nPrefixReg, 1, nPrefixReg); + pushOntoSorter(pParse, pSort, p, r1); }else{ int r2 = sqlite3GetTempReg(pParse); sqlite3VdbeAddOp2(v, OP_NewRowid, iParm, r2); sqlite3VdbeAddOp3(v, OP_Insert, iParm, r1, r2); sqlite3VdbeChangeP5(v, OPFLAG_APPEND); sqlite3ReleaseTempReg(pParse, r2); } - sqlite3ReleaseTempRange(pParse, r1, nPrefixReg+1); + sqlite3ReleaseTempReg(pParse, r1); break; } #ifndef SQLITE_OMIT_SUBQUERY /* If we are creating a set for an "expr IN (SELECT ...)" construct, @@ -104831,11 +102248,11 @@ if( pSort ){ /* At first glance you would think we could optimize out the ** ORDER BY in this case since the order of entries in the set ** does not matter. But there might be a LIMIT clause, in which ** case the order does matter */ - pushOntoSorter(pParse, pSort, p, regResult, 1, nPrefixReg); + pushOntoSorter(pParse, pSort, p, regResult); }else{ int r1 = sqlite3GetTempReg(pParse); sqlite3VdbeAddOp4(v, OP_MakeRecord, regResult,1,r1, &pDest->affSdst, 1); sqlite3ExprCacheAffinityChange(pParse, regResult, 1); sqlite3VdbeAddOp2(v, OP_IdxInsert, iParm, r1); @@ -104857,13 +102274,13 @@ ** of the scan loop. */ case SRT_Mem: { assert( nResultCol==1 ); if( pSort ){ - pushOntoSorter(pParse, pSort, p, regResult, 1, nPrefixReg); + pushOntoSorter(pParse, pSort, p, regResult); }else{ - assert( regResult==iParm ); + sqlite3ExprCodeMove(pParse, regResult, iParm, 1); /* The LIMIT clause will jump out of the loop for us */ } break; } #endif /* #ifndef SQLITE_OMIT_SUBQUERY */ @@ -104871,11 +102288,14 @@ case SRT_Coroutine: /* Send data to a co-routine */ case SRT_Output: { /* Return the results */ testcase( eDest==SRT_Coroutine ); testcase( eDest==SRT_Output ); if( pSort ){ - pushOntoSorter(pParse, pSort, p, regResult, nResultCol, nPrefixReg); + int r1 = sqlite3GetTempReg(pParse); + sqlite3VdbeAddOp3(v, OP_MakeRecord, regResult, nResultCol, r1); + pushOntoSorter(pParse, pSort, p, r1); + sqlite3ReleaseTempReg(pParse, r1); }else if( eDest==SRT_Coroutine ){ sqlite3VdbeAddOp1(v, OP_Yield, pDest->iSDParm); }else{ sqlite3VdbeAddOp2(v, OP_ResultRow, regResult, nResultCol); sqlite3ExprCacheAffinityChange(pParse, regResult, nResultCol); @@ -105014,11 +102434,11 @@ ** KeyInfo structure is appropriate for initializing a virtual index to ** implement that clause. If the ExprList is the result set of a SELECT ** then the KeyInfo structure is appropriate for initializing a virtual ** index to implement a DISTINCT test. ** -** Space to hold the KeyInfo structure is obtained from malloc. The calling +** Space to hold the KeyInfo structure is obtain from malloc. The calling ** function is responsible for seeing that this structure is eventually ** freed. */ static KeyInfo *keyInfoFromExprList( Parse *pParse, /* Parsing context */ @@ -105151,66 +102571,50 @@ int addrBreak = sqlite3VdbeMakeLabel(v); /* Jump here to exit loop */ int addrContinue = sqlite3VdbeMakeLabel(v); /* Jump here for next cycle */ int addr; int addrOnce = 0; int iTab; + int pseudoTab = 0; ExprList *pOrderBy = pSort->pOrderBy; int eDest = pDest->eDest; int iParm = pDest->iSDParm; int regRow; int regRowid; int nKey; - int iSortTab; /* Sorter cursor to read from */ - int nSortData; /* Trailing values to read from sorter */ - u8 p5; /* p5 parameter for 1st OP_Column */ - int i; - int bSeq; /* True if sorter record includes seq. no. */ -#ifdef SQLITE_ENABLE_EXPLAIN_COMMENTS - struct ExprList_item *aOutEx = p->pEList->a; -#endif if( pSort->labelBkOut ){ sqlite3VdbeAddOp2(v, OP_Gosub, pSort->regReturn, pSort->labelBkOut); sqlite3VdbeAddOp2(v, OP_Goto, 0, addrBreak); sqlite3VdbeResolveLabel(v, pSort->labelBkOut); + addrOnce = sqlite3CodeOnce(pParse); VdbeCoverage(v); } iTab = pSort->iECursor; + regRow = sqlite3GetTempReg(pParse); if( eDest==SRT_Output || eDest==SRT_Coroutine ){ + pseudoTab = pParse->nTab++; + sqlite3VdbeAddOp3(v, OP_OpenPseudo, pseudoTab, regRow, nColumn); regRowid = 0; - regRow = pDest->iSdst; - nSortData = nColumn; }else{ regRowid = sqlite3GetTempReg(pParse); - regRow = sqlite3GetTempReg(pParse); - nSortData = 1; } nKey = pOrderBy->nExpr - pSort->nOBSat; if( pSort->sortFlags & SORTFLAG_UseSorter ){ int regSortOut = ++pParse->nMem; - iSortTab = pParse->nTab++; - if( pSort->labelBkOut ){ - addrOnce = sqlite3CodeOnce(pParse); VdbeCoverage(v); - } - sqlite3VdbeAddOp3(v, OP_OpenPseudo, iSortTab, regSortOut, nKey+1+nSortData); + int ptab2 = pParse->nTab++; + sqlite3VdbeAddOp3(v, OP_OpenPseudo, ptab2, regSortOut, nKey+2); if( addrOnce ) sqlite3VdbeJumpHere(v, addrOnce); addr = 1 + sqlite3VdbeAddOp2(v, OP_SorterSort, iTab, addrBreak); VdbeCoverage(v); codeOffset(v, p->iOffset, addrContinue); sqlite3VdbeAddOp2(v, OP_SorterData, iTab, regSortOut); - p5 = OPFLAG_CLEARCACHE; - bSeq = 0; + sqlite3VdbeAddOp3(v, OP_Column, ptab2, nKey+1, regRow); + sqlite3VdbeChangeP5(v, OPFLAG_CLEARCACHE); }else{ + if( addrOnce ) sqlite3VdbeJumpHere(v, addrOnce); addr = 1 + sqlite3VdbeAddOp2(v, OP_Sort, iTab, addrBreak); VdbeCoverage(v); codeOffset(v, p->iOffset, addrContinue); - iSortTab = iTab; - p5 = 0; - bSeq = 1; - } - for(i=0; i nExpr; i++){ - pOrderBy->a[i].u.x.iOrderByCol = 0; - } assert( pParent->pOrderBy==0 ); - assert( pSub->pPrior==0 ); - pParent->pOrderBy = pOrderBy; + pParent->pOrderBy = pSub->pOrderBy; pSub->pOrderBy = 0; }else if( pParent->pOrderBy ){ substExprList(db, pParent->pOrderBy, iParent, pSub->pEList); } if( pSub->pWhere ){ @@ -107620,17 +104997,10 @@ /* Finially, delete what is left of the subquery and return ** success. */ sqlite3SelectDelete(db, pSub1); -#if SELECTTRACE_ENABLED - if( sqlite3SelectTrace & 0x100 ){ - sqlite3DebugPrintf("After flattening:\n"); - sqlite3TreeViewSelect(0, p, 0); - } -#endif - return 1; } #endif /* !defined(SQLITE_OMIT_SUBQUERY) || !defined(SQLITE_OMIT_VIEW) */ /* @@ -107673,11 +105043,11 @@ return eRet; } /* ** The select statement passed as the first argument is an aggregate query. -** The second argument is the associated aggregate-info object. This +** The second argment is the associated aggregate-info object. This ** function tests if the SELECT is of the form: ** ** SELECT count(*) FROM ** ** where table is a database table, not a sub-select or view. If the query @@ -108003,14 +105373,14 @@ ** (2) Fill in the pTabList->a[].pTab fields in the SrcList that ** defines FROM clause. When views appear in the FROM clause, ** fill pTabList->a[].pSelect with a copy of the SELECT statement ** that implements the view. A copy is made of the view's SELECT ** statement so that we can freely modify or delete that statement -** without worrying about messing up the persistent representation +** without worrying about messing up the presistent representation ** of the view. ** -** (3) Add terms to the WHERE clause to accommodate the NATURAL keyword +** (3) Add terms to the WHERE clause to accomodate the NATURAL keyword ** on joins and the ON and USING clause of joins. ** ** (4) Scan the list of columns in the result set (pEList) looking ** for instances of the "*" operator or the TABLE.* operator. ** If found, expand each "*" to be every column in every table @@ -108097,11 +105467,10 @@ if( pTab->pSelect || IsVirtual(pTab) ){ /* We reach here if the named table is a really a view */ if( sqlite3ViewGetColumnNames(pParse, pTab) ) return WRC_Abort; assert( pFrom->pSelect==0 ); pFrom->pSelect = sqlite3SelectDup(db, pTab->pSelect, 0); - sqlite3SelectSetName(pFrom->pSelect, pTab->zName); sqlite3WalkSelect(pWalker, pFrom->pSelect); } #endif } @@ -108632,17 +106001,10 @@ if( p==0 || db->mallocFailed || pParse->nErr ){ return 1; } if( sqlite3AuthCheck(pParse, SQLITE_SELECT, 0, 0, 0) ) return 1; memset(&sAggInfo, 0, sizeof(sAggInfo)); -#if SELECTTRACE_ENABLED - pParse->nSelectIndent++; - SELECTTRACE(1,pParse,p, ("begin processing:\n")); - if( sqlite3SelectTrace & 0x100 ){ - sqlite3TreeViewSelect(0, p, 0); - } -#endif assert( p->pOrderBy==0 || pDest->eDest!=SRT_DistFifo ); assert( p->pOrderBy==0 || pDest->eDest!=SRT_Fifo ); assert( p->pOrderBy==0 || pDest->eDest!=SRT_DistQueue ); assert( p->pOrderBy==0 || pDest->eDest!=SRT_Queue ); @@ -108795,14 +106157,10 @@ /* If there is are a sequence of queries, do the earlier ones first. */ if( p->pPrior ){ rc = multiSelect(pParse, p, pDest); explainSetInteger(pParse->iSelectId, iRestoreSelectId); -#if SELECTTRACE_ENABLED - SELECTTRACE(1,pParse,p,("end compound-select processing\n")); - pParse->nSelectIndent--; -#endif return rc; } #endif /* If the query is DISTINCT with an ORDER BY but is not an aggregate, and @@ -108844,13 +106202,12 @@ KeyInfo *pKeyInfo; pKeyInfo = keyInfoFromExprList(pParse, sSort.pOrderBy, 0, 0); sSort.iECursor = pParse->nTab++; sSort.addrSortIndex = sqlite3VdbeAddOp4(v, OP_OpenEphemeral, - sSort.iECursor, sSort.pOrderBy->nExpr+1+pEList->nExpr, 0, - (char*)pKeyInfo, P4_KEYINFO - ); + sSort.iECursor, sSort.pOrderBy->nExpr+2, 0, + (char*)pKeyInfo, P4_KEYINFO); }else{ sSort.addrSortIndex = -1; } /* If the output is destined for a temporary table, open that table. @@ -108977,11 +106334,11 @@ memset(&sNC, 0, sizeof(sNC)); sNC.pParse = pParse; sNC.pSrcList = pTabList; sNC.pAggInfo = &sAggInfo; sAggInfo.mnReg = pParse->nMem+1; - sAggInfo.nSortingColumn = pGroupBy ? pGroupBy->nExpr : 0; + sAggInfo.nSortingColumn = pGroupBy ? pGroupBy->nExpr+1 : 0; sAggInfo.pGroupBy = pGroupBy; sqlite3ExprAnalyzeAggList(&sNC, pEList); sqlite3ExprAnalyzeAggList(&sNC, sSort.pOrderBy); if( pHaving ){ sqlite3ExprAnalyzeAggregates(&sNC, pHaving); @@ -109070,22 +106427,23 @@ (sDistinct.isTnct && (p->selFlags&SF_Distinct)==0) ? "DISTINCT" : "GROUP BY"); groupBySort = 1; nGroupBy = pGroupBy->nExpr; - nCol = nGroupBy; - j = nGroupBy; + nCol = nGroupBy + 1; + j = nGroupBy+1; for(i=0; i pSrc->nSrc; i++){ struct SrcList_item *pItem = &p->pSrc->a[i]; - StrAccum x; - char zLine[100]; - sqlite3StrAccumInit(&x, zLine, sizeof(zLine), 0); - sqlite3XPrintf(&x, 0, "{%d,*}", pItem->iCursor); - if( pItem->zDatabase ){ - sqlite3XPrintf(&x, 0, " %s.%s", pItem->zDatabase, pItem->zName); + sqlite3ExplainPrintf(pVdbe, "{%d,*} = ", pItem->iCursor); + if( pItem->pSelect ){ + sqlite3ExplainSelect(pVdbe, pItem->pSelect); + if( pItem->pTab ){ + sqlite3ExplainPrintf(pVdbe, " (tabname=%s)", pItem->pTab->zName); + } }else if( pItem->zName ){ - sqlite3XPrintf(&x, 0, " %s", pItem->zName); - } - if( pItem->pTab ){ - sqlite3XPrintf(&x, 0, " tabname=%Q", pItem->pTab->zName); + sqlite3ExplainPrintf(pVdbe, "%s", pItem->zName); } if( pItem->zAlias ){ - sqlite3XPrintf(&x, 0, " (AS %s)", pItem->zAlias); + sqlite3ExplainPrintf(pVdbe, " (AS %s)", pItem->zAlias); } if( pItem->jointype & JT_LEFT ){ - sqlite3XPrintf(&x, 0, " LEFT-JOIN"); - } - sqlite3StrAccumFinish(&x); - sqlite3TreeViewItem(pView, zLine, i pSrc->nSrc-1); - if( pItem->pSelect ){ - sqlite3TreeViewSelect(pView, pItem->pSelect, 0); - } - sqlite3TreeViewPop(pView); - } - sqlite3TreeViewPop(pView); + sqlite3ExplainPrintf(pVdbe, " LEFT-JOIN"); + } + sqlite3ExplainNL(pVdbe); + } + sqlite3ExplainPop(pVdbe); } if( p->pWhere ){ - sqlite3TreeViewItem(pView, "WHERE", (n--)>0); - sqlite3TreeViewExpr(pView, p->pWhere, 0); - sqlite3TreeViewPop(pView); + sqlite3ExplainPrintf(pVdbe, "WHERE "); + sqlite3ExplainExpr(pVdbe, p->pWhere); + sqlite3ExplainNL(pVdbe); } if( p->pGroupBy ){ - sqlite3TreeViewExprList(pView, p->pGroupBy, (n--)>0, "GROUPBY"); + sqlite3ExplainPrintf(pVdbe, "GROUPBY "); + sqlite3ExplainExprList(pVdbe, p->pGroupBy); + sqlite3ExplainNL(pVdbe); } if( p->pHaving ){ - sqlite3TreeViewItem(pView, "HAVING", (n--)>0); - sqlite3TreeViewExpr(pView, p->pHaving, 0); - sqlite3TreeViewPop(pView); + sqlite3ExplainPrintf(pVdbe, "HAVING "); + sqlite3ExplainExpr(pVdbe, p->pHaving); + sqlite3ExplainNL(pVdbe); } if( p->pOrderBy ){ - sqlite3TreeViewExprList(pView, p->pOrderBy, (n--)>0, "ORDERBY"); + sqlite3ExplainPrintf(pVdbe, "ORDERBY "); + sqlite3ExplainExprList(pVdbe, p->pOrderBy); + sqlite3ExplainNL(pVdbe); } if( p->pLimit ){ - sqlite3TreeViewItem(pView, "LIMIT", (n--)>0); - sqlite3TreeViewExpr(pView, p->pLimit, 0); - sqlite3TreeViewPop(pView); + sqlite3ExplainPrintf(pVdbe, "LIMIT "); + sqlite3ExplainExpr(pVdbe, p->pLimit); + sqlite3ExplainNL(pVdbe); } if( p->pOffset ){ - sqlite3TreeViewItem(pView, "OFFSET", (n--)>0); - sqlite3TreeViewExpr(pView, p->pOffset, 0); - sqlite3TreeViewPop(pView); - } - if( p->pPrior ){ - const char *zOp = "UNION"; - switch( p->op ){ - case TK_ALL: zOp = "UNION ALL"; break; - case TK_INTERSECT: zOp = "INTERSECT"; break; - case TK_EXCEPT: zOp = "EXCEPT"; break; - } - sqlite3TreeViewItem(pView, zOp, (n--)>0); - sqlite3TreeViewSelect(pView, p->pPrior, 0); - sqlite3TreeViewPop(pView); - } - sqlite3TreeViewPop(pView); -} -#endif /* SQLITE_DEBUG */ + sqlite3ExplainPrintf(pVdbe, "OFFSET "); + sqlite3ExplainExpr(pVdbe, p->pOffset); + sqlite3ExplainNL(pVdbe); + } +} +SQLITE_PRIVATE void sqlite3ExplainSelect(Vdbe *pVdbe, Select *p){ + if( p==0 ){ + sqlite3ExplainPrintf(pVdbe, "(null-select)"); + return; + } + sqlite3ExplainPush(pVdbe); + while( p ){ + explainOneSelect(pVdbe, p); + p = p->pNext; + if( p==0 ) break; + sqlite3ExplainNL(pVdbe); + sqlite3ExplainPrintf(pVdbe, "%s\n", selectOpName(p->op)); + } + sqlite3ExplainPrintf(pVdbe, "END"); + sqlite3ExplainPop(pVdbe); +} + +/* End of the structure debug printing code +*****************************************************************************/ +#endif /* defined(SQLITE_ENABLE_TREE_EXPLAIN) */ /************** End of select.c **********************************************/ /************** Begin file table.c *******************************************/ /* ** 2001 September 15 @@ -109531,14 +106886,14 @@ ** to the callback function is uses to build the result. */ typedef struct TabResult { char **azResult; /* Accumulated output */ char *zErrMsg; /* Error message text, if an error occurs */ - u32 nAlloc; /* Slots allocated for azResult[] */ - u32 nRow; /* Number of rows in the result */ - u32 nColumn; /* Number of columns in the result */ - u32 nData; /* Slots used in azResult[]. (nRow+1)*nColumn */ + int nAlloc; /* Slots allocated for azResult[] */ + int nRow; /* Number of rows in the result */ + int nColumn; /* Number of columns in the result */ + int nData; /* Slots used in azResult[]. (nRow+1)*nColumn */ int rc; /* Return code from sqlite3_exec() */ } TabResult; /* ** This routine is called once for each row in the result table. Its job @@ -109560,11 +106915,11 @@ need = nCol; } if( p->nData + need > p->nAlloc ){ char **azNew; p->nAlloc = p->nAlloc*2 + need; - azNew = sqlite3_realloc64( p->azResult, sizeof(char*)*p->nAlloc ); + azNew = sqlite3_realloc( p->azResult, sizeof(char*)*p->nAlloc ); if( azNew==0 ) goto malloc_failed; p->azResult = azNew; } /* If this is the first row, then generate an extra row containing @@ -109575,11 +106930,11 @@ for(i=0; i nDb; i++){ int j = (i<2) ? i^1 : i; /* Search TEMP before MAIN */ if( zDb && sqlite3StrICmp(db->aDb[j].zName, zDb) ) continue; assert( sqlite3SchemaMutexHeld(db, j, 0) ); - pTrigger = sqlite3HashFind(&(db->aDb[j].pSchema->trigHash), zName); + pTrigger = sqlite3HashFind(&(db->aDb[j].pSchema->trigHash), zName, nName); if( pTrigger ) break; } if( !pTrigger ){ if( !noErr ){ sqlite3ErrorMsg(pParse, "no such trigger: %S", pName, 0); @@ -110226,11 +107585,12 @@ /* ** Return a pointer to the Table structure for the table that a trigger ** is set on. */ static Table *tableOfTrigger(Trigger *pTrigger){ - return sqlite3HashFind(&pTrigger->pTabSchema->tblHash, pTrigger->table); + int n = sqlite3Strlen30(pTrigger->table); + return sqlite3HashFind(&pTrigger->pTabSchema->tblHash, pTrigger->table, n); } /* ** Drop a trigger given a pointer to that trigger. @@ -110298,11 +107658,11 @@ Trigger *pTrigger; Hash *pHash; assert( sqlite3SchemaMutexHeld(db, iDb, 0) ); pHash = &(db->aDb[iDb].pSchema->trigHash); - pTrigger = sqlite3HashInsert(pHash, zName, 0); + pTrigger = sqlite3HashInsert(pHash, zName, sqlite3Strlen30(zName), 0); if( ALWAYS(pTrigger) ){ if( pTrigger->pSchema==pTrigger->pTabSchema ){ Table *pTab = tableOfTrigger(pTrigger); Trigger **pp; for(pp=&pTab->pTrigger; *pp!=pTrigger; pp=&((*pp)->pNext)); @@ -111149,11 +108509,11 @@ if( isView ){ sqlite3AuthContextPush(pParse, &sContext, pTab->zName); } /* If we are trying to update a view, realize that view into - ** an ephemeral table. + ** a ephemeral table. */ #if !defined(SQLITE_OMIT_VIEW) && !defined(SQLITE_OMIT_TRIGGER) if( isView ){ sqlite3MaterializeView(pParse, pTab, pWhere, iDataCur); } @@ -111310,11 +108670,11 @@ sqlite3VdbeAddOp2(v, OP_Copy, regOldRowid, regNewRowid); } } /* Populate the array of registers beginning at regNew with the new - ** row data. This array is used to check constants, create the new + ** row data. This array is used to check constaints, create the new ** table and index records, and as the values for any new.* references ** made by triggers. ** ** If there are one or more BEFORE triggers, then do not populate the ** registers associated with columns that are (a) not modified by @@ -111490,11 +108850,11 @@ sqlite3ExprListDelete(db, pChanges); sqlite3ExprDelete(db, pWhere); return; } /* Make sure "isView" and other macros defined above are undefined. Otherwise -** they may interfere with compilation of other functions in this file +** thely may interfere with compilation of other functions in this file ** (or in another file, if this file becomes part of the amalgamation). */ #ifdef isView #undef isView #endif #ifdef pTrigger @@ -111503,19 +108863,19 @@ #ifndef SQLITE_OMIT_VIRTUALTABLE /* ** Generate code for an UPDATE of a virtual table. ** -** The strategy is that we create an ephemeral table that contains +** The strategy is that we create an ephemerial table that contains ** for each row to be changed: ** ** (A) The original rowid of that row. ** (B) The revised rowid for the row. (note1) ** (C) The content of every column in the row. ** ** Then we loop over this ephemeral table and for each row in -** the ephemeral table call VUpdate. +** the ephermeral table call VUpdate. ** ** When finished, drop the ephemeral table. ** ** (note1) Actually, if we know in advance that (A) is always the same ** as (B) we only store (A), then duplicate (A) when pulling @@ -111684,11 +109044,11 @@ ** The transient database requires temporary disk space approximately ** equal to the size of the original database. The copy operation of ** step (3) requires additional temporary disk space approximately equal ** to the size of the original database for the rollback journal. ** Hence, temporary disk space that is approximately 2x the size of the -** original database is required. Every page of the database is written +** orginal database is required. Every page of the database is written ** approximately 3 times: Once for step (2) and twice for step (3). ** Two writes per page are required in step (3) because the original ** database content must be written into the rollback journal prior to ** overwriting the database with the vacuumed content. ** @@ -112011,11 +109371,11 @@ int rc = SQLITE_OK; int nName; sqlite3_mutex_enter(db->mutex); nName = sqlite3Strlen30(zName); - if( sqlite3HashFind(&db->aModule, zName) ){ + if( sqlite3HashFind(&db->aModule, zName, nName) ){ rc = SQLITE_MISUSE_BKPT; }else{ Module *pMod; pMod = (Module *)sqlite3DbMallocRaw(db, sizeof(Module) + nName + 1); if( pMod ){ @@ -112024,11 +109384,11 @@ memcpy(zCopy, zName, nName+1); pMod->zName = zCopy; pMod->pModule = pModule; pMod->pAux = pAux; pMod->xDestroy = xDestroy; - pDel = (Module *)sqlite3HashInsert(&db->aModule,zCopy,(void*)pMod); + pDel = (Module *)sqlite3HashInsert(&db->aModule,zCopy,nName,(void*)pMod); assert( pDel==0 || pDel==pMod ); if( pDel ){ db->mallocFailed = 1; sqlite3DbFree(db, pDel); } @@ -112393,12 +109753,13 @@ ** the required virtual table implementations are registered. */ else { Table *pOld; Schema *pSchema = pTab->pSchema; const char *zName = pTab->zName; + int nName = sqlite3Strlen30(zName); assert( sqlite3SchemaMutexHeld(db, 0, pSchema) ); - pOld = sqlite3HashInsert(&pSchema->tblHash, zName, pTab); + pOld = sqlite3HashInsert(&pSchema->tblHash, zName, nName, pTab); if( pOld ){ db->mallocFailed = 1; assert( pTab==pOld ); /* Malloc must have failed inside HashInsert() */ return; } @@ -112560,11 +109921,11 @@ return SQLITE_OK; } /* Locate the required virtual table module */ zMod = pTab->azModuleArg[0]; - pMod = (Module*)sqlite3HashFind(&db->aModule, zMod); + pMod = (Module*)sqlite3HashFind(&db->aModule, zMod, sqlite3Strlen30(zMod)); if( !pMod ){ const char *zModule = pTab->azModuleArg[0]; sqlite3ErrorMsg(pParse, "no such module: %s", zModule); rc = SQLITE_ERROR; @@ -112628,11 +109989,11 @@ pTab = sqlite3FindTable(db, zTab, db->aDb[iDb].zName); assert( pTab && (pTab->tabFlags & TF_Virtual)!=0 && !pTab->pVTable ); /* Locate the required virtual table module */ zMod = pTab->azModuleArg[0]; - pMod = (Module*)sqlite3HashFind(&db->aModule, zMod); + pMod = (Module*)sqlite3HashFind(&db->aModule, zMod, sqlite3Strlen30(zMod)); /* If the module has been registered and includes a Create method, ** invoke it now. If the module has not been registered, return an ** error. Otherwise, do nothing. */ @@ -112667,11 +110028,11 @@ Table *pTab; char *zErr = 0; sqlite3_mutex_enter(db->mutex); if( !db->pVtabCtx || !(pTab = db->pVtabCtx->pTab) ){ - sqlite3Error(db, SQLITE_MISUSE); + sqlite3Error(db, SQLITE_MISUSE, 0); sqlite3_mutex_leave(db->mutex); return SQLITE_MISUSE_BKPT; } assert( (pTab->tabFlags & TF_Virtual)!=0 ); @@ -112695,11 +110056,11 @@ pParse->pNewTable->nCol = 0; pParse->pNewTable->aCol = 0; } db->pVtabCtx->pTab = 0; }else{ - sqlite3ErrorWithMsg(db, SQLITE_ERROR, (zErr ? "%s" : 0), zErr); + sqlite3Error(db, SQLITE_ERROR, (zErr ? "%s" : 0), zErr); sqlite3DbFree(db, zErr); rc = SQLITE_ERROR; } pParse->declareVtab = 0; @@ -113056,11 +110417,11 @@ rc = SQLITE_MISUSE_BKPT; break; } va_end(ap); - if( rc!=SQLITE_OK ) sqlite3Error(db, rc); + if( rc!=SQLITE_OK ) sqlite3Error(db, rc, 0); sqlite3_mutex_leave(db->mutex); return rc; } #endif /* SQLITE_OMIT_VIRTUALTABLE */ @@ -113263,11 +110624,11 @@ ** ** The "solver" works by creating the N best WherePath objects of length ** 1. Then using those as a basis to compute the N best WherePath objects ** of length 2. And so forth until the length of WherePaths equals the ** number of nodes in the FROM clause. The best (lowest cost) WherePath -** at the end is the chosen query plan. +** at the end is the choosen query plan. */ struct WherePath { Bitmask maskLoop; /* Bitmask of all WhereLoop objects in this path */ Bitmask revLoop; /* aLoop[]s that should be reversed for ORDER BY */ LogEst nRow; /* Estimated number of rows generated by this path */ @@ -113895,10 +111256,15 @@ assert( TK_LE>TK_EQ && TK_LE =TK_EQ && op<=TK_GE) || op==TK_ISNULL; } +/* +** Swap two objects of type TYPE. +*/ +#define SWAP(TYPE,A,B) {TYPE t=A; A=B; B=t;} + /* ** Commute a comparison operator. Expressions of the form "X op Y" ** are converted into "Y op X". ** ** If left/right precedence rules come into play when determining the @@ -114227,11 +111593,11 @@ if( *pisComplete && pRight->u.zToken[1] ){ /* If the rhs of the LIKE expression is a variable, and the current ** value of the variable means there is no need to invoke the LIKE ** function, then no OP_Variable will be added to the program. ** This causes problems for the sqlite3_bind_parameter_name() - ** API. To work around them, add a dummy OP_Variable here. + ** API. To workaround them, add a dummy OP_Variable here. */ int r1 = sqlite3GetTempReg(pParse); sqlite3ExprCodeTarget(pParse, pRight, r1); sqlite3VdbeChangeP3(v, sqlite3VdbeCurrentAddr(v)-1, 0); sqlite3ReleaseTempReg(pParse, r1); @@ -114347,11 +111713,11 @@ ** This analysis does not consider whether or not the index exists; that ** is decided elsewhere. This analysis only looks at whether subterms ** appropriate for indexing exist. ** ** All examples A through E above satisfy case 2. But if a term -** also satisfies case 1 (such as B) we know that the optimizer will +** also statisfies case 1 (such as B) we know that the optimizer will ** always prefer case 1, so in that case we pretend that case 2 is not ** satisfied. ** ** It might be the case that multiple tables are indexable. For example, ** (E) above is indexable on tables P, Q, and R. @@ -114505,11 +111871,11 @@ assert( j==1 ); continue; } if( (chngToIN & getMask(&pWInfo->sMaskSet, pOrTerm->leftCursor))==0 ){ /* This term must be of the form t1.a==t2.b where t2 is in the - ** chngToIN set but t1 is not. This term will be either preceded + ** chngToIN set but t1 is not. This term will be either preceeded ** or follwed by an inverted copy (t2.b==t1.a). Skip this term ** and use its inversion. */ testcase( pOrTerm->wtFlags & TERM_COPIED ); testcase( pOrTerm->wtFlags & TERM_VIRTUAL ); assert( pOrTerm->wtFlags & (TERM_COPIED|TERM_VIRTUAL) ); @@ -114916,11 +112282,11 @@ */ pTerm->prereqRight |= extraRight; } /* -** This function searches pList for an entry that matches the iCol-th column +** This function searches pList for a entry that matches the iCol-th column ** of index pIdx. ** ** If such an expression is found, its index in pList->a[] is returned. If ** no expression is found, -1 is returned. */ @@ -115439,11 +112805,11 @@ iCol = pRec->nField - 1; assert( pIdx->nSample>0 ); assert( pRec->nField>0 && iCol nSampleCol ); do{ iTest = (iMin+i)/2; - res = sqlite3VdbeRecordCompare(aSample[iTest].n, aSample[iTest].p, pRec); + res = sqlite3VdbeRecordCompare(aSample[iTest].n, aSample[iTest].p, pRec, 0); if( res<0 ){ iMin = iTest+1; }else{ i = iTest; } @@ -115454,20 +112820,20 @@ ** above found the right answer. This block serves no purpose other ** than to invoke the asserts. */ if( res==0 ){ /* If (res==0) is true, then sample $i must be equal to pRec */ assert( i nSample ); - assert( 0==sqlite3VdbeRecordCompare(aSample[i].n, aSample[i].p, pRec) + assert( 0==sqlite3VdbeRecordCompare(aSample[i].n, aSample[i].p, pRec, 0) || pParse->db->mallocFailed ); }else{ /* Otherwise, pRec must be smaller than sample $i and larger than ** sample ($i-1). */ assert( i==pIdx->nSample - || sqlite3VdbeRecordCompare(aSample[i].n, aSample[i].p, pRec)>0 + || sqlite3VdbeRecordCompare(aSample[i].n, aSample[i].p, pRec, 0)>0 || pParse->db->mallocFailed ); assert( i==0 - || sqlite3VdbeRecordCompare(aSample[i-1].n, aSample[i-1].p, pRec)<0 + || sqlite3VdbeRecordCompare(aSample[i-1].n, aSample[i-1].p, pRec, 0)<0 || pParse->db->mallocFailed ); } #endif /* ifdef SQLITE_DEBUG */ /* At this point, aSample[i] is the first sample that is greater than @@ -115666,11 +113032,11 @@ ** ** When this function is called, *pnOut is set to the sqlite3LogEst() of the ** number of rows that the index scan is expected to visit without ** considering the range constraints. If nEq is 0, this is the number of ** rows in the index. Assuming no error occurs, *pnOut is adjusted (reduced) -** to account for the range constraints pLower and pUpper. +** to account for the range contraints pLower and pUpper. ** ** In the absence of sqlite_stat4 ANALYZE data, or if such data cannot be ** used, a single range inequality reduces the search space by a factor of 4. ** and a pair of constraints (x>? AND xnField!=pBuilder->nRecValid ); - pRec->nField = pBuilder->nRecValid; - } if( nEq==p->nKeyCol ){ aff = SQLITE_AFF_INTEGER; }else{ aff = p->pTable->aCol[p->aiColumn[nEq]].affinity; } @@ -115750,11 +113112,10 @@ tRowcnt iNew; whereKeyStats(pParse, p, pRec, 0, a); iNew = a[0] + ((pLower->eOperator & WO_GT) ? a[1] : 0); if( iNew>iLower ) iLower = iNew; nOut--; - pLower = 0; } } /* If possible, improve on the iUpper estimate using ($P:$U). */ if( pUpper ){ @@ -115766,11 +113127,10 @@ tRowcnt iNew; whereKeyStats(pParse, p, pRec, 1, a); iNew = a[0] + ((pUpper->eOperator & WO_LE) ? a[1] : 0); if( iNew pRec = pRec; if( rc==SQLITE_OK ){ @@ -115780,12 +113140,14 @@ nNew = 10; assert( 10==sqlite3LogEst(2) ); } if( nNew nOut = (LogEst)nOut; + WHERETRACE(0x10, ("range scan regions: %u..%u est=%d\n", (u32)iLower, (u32)iUpper, nOut)); + return SQLITE_OK; } }else{ int bDone = 0; rc = whereRangeSkipScanEst(pParse, pLower, pUpper, pLoop, &bDone); if( bDone ) return rc; @@ -115792,12 +113154,12 @@ } } #else UNUSED_PARAMETER(pParse); UNUSED_PARAMETER(pBuilder); +#endif assert( pLower || pUpper ); -#endif assert( pUpper==0 || (pUpper->wtFlags & TERM_VNULL)==0 ); nNew = whereRangeAdjust(pLower, nOut); nNew = whereRangeAdjust(pUpper, nNew); /* TUNING: If there is both an upper and lower limit, assume the range is @@ -115808,16 +113170,10 @@ if( pLower && pUpper ) nNew -= 20; nOut -= (pLower!=0) + (pUpper!=0); if( nNew<10 ) nNew = 10; if( nNew nOut>nOut ){ - WHERETRACE(0x10,("Range scan lowers nOut from %d to %d\n", - pLoop->nOut, nOut)); - } -#endif pLoop->nOut = (LogEst)nOut; return rc; } #ifdef SQLITE_ENABLE_STAT3_OR_STAT4 @@ -115926,11 +113282,11 @@ } if( rc==SQLITE_OK ){ if( nRowEst > nRow0 ) nRowEst = nRow0; *pnRow = nRowEst; - WHERETRACE(0x10,("IN row estimate: est=%d\n", nRowEst)); + WHERETRACE(0x10,("IN row estimate: est=%g\n", nRowEst)); } assert( pBuilder->nRecValid==nRecValid ); return rc; } #endif /* SQLITE_ENABLE_STAT3_OR_STAT4 */ @@ -116943,11 +114299,11 @@ ** Return 2 # Jump back to the Gosub ** ** B: ** ** Added 2014-05-26: If the table is a WITHOUT ROWID table, then - ** use an ephemeral index instead of a RowSet to record the primary + ** use an ephermeral index instead of a RowSet to record the primary ** keys of the rows we have already seen. ** */ WhereClause *pOrWc; /* The OR-clause broken out into subterms */ SrcList *pOrTab; /* Shortened table list or OR-clause generation */ @@ -116994,11 +114350,11 @@ }else{ pOrTab = pWInfo->pTabList; } /* Initialize the rowset register to contain NULL. An SQL NULL is - ** equivalent to an empty rowset. Or, create an ephemeral index + ** equivalent to an empty rowset. Or, create an ephermeral index ** capable of holding primary keys in the case of a WITHOUT ROWID. ** ** Also initialize regReturn to contain the address of the instruction ** immediately following the OP_Return at the bottom of the loop. This ** is required in a few obscure LEFT JOIN cases where control jumps @@ -117055,13 +114411,12 @@ /* Run a separate WHERE clause for each term of the OR clause. After ** eliminating duplicates from other WHERE clauses, the action for each ** sub-WHERE clause is to to invoke the main loop body as a subroutine. */ - wctrlFlags = WHERE_OMIT_OPEN_CLOSE - | WHERE_FORCE_TABLE - | WHERE_ONETABLE_ONLY; + wctrlFlags = WHERE_OMIT_OPEN_CLOSE | WHERE_AND_ONLY | + WHERE_FORCE_TABLE | WHERE_ONETABLE_ONLY; for(ii=0; ii nTerm; ii++){ WhereTerm *pOrTerm = &pOrWc->a[ii]; if( pOrTerm->leftCursor==iCur || (pOrTerm->eOperator & WO_AND)!=0 ){ WhereInfo *pSubWInfo; /* Info for single OR-term scan */ Expr *pOrExpr = pOrTerm->pExpr; /* Current OR clause term */ @@ -117069,11 +114424,10 @@ if( pAndExpr && !ExprHasProperty(pOrExpr, EP_FromJoin) ){ pAndExpr->pLeft = pOrExpr; pOrExpr = pAndExpr; } /* Loop through table entries that match term pOrTerm. */ - WHERETRACE(0xffff, ("Subplan for OR-clause:\n")); pSubWInfo = sqlite3WhereBegin(pParse, pOrTab, pOrExpr, 0, 0, wctrlFlags, iCovCur); assert( pSubWInfo || pParse->nErr || db->mallocFailed ); if( pSubWInfo ){ WhereLoop *pSubLoop; @@ -117289,30 +114643,25 @@ } return pLevel->notReady; } -#ifdef WHERETRACE_ENABLED +#if defined(WHERETRACE_ENABLED) && defined(SQLITE_ENABLE_TREE_EXPLAIN) /* -** Print the content of a WhereTerm object +** Generate "Explanation" text for a WhereTerm. */ -static void whereTermPrint(WhereTerm *pTerm, int iTerm){ - if( pTerm==0 ){ - sqlite3DebugPrintf("TERM-%-3d NULL\n", iTerm); - }else{ - char zType[4]; - memcpy(zType, "...", 4); - if( pTerm->wtFlags & TERM_VIRTUAL ) zType[0] = 'V'; - if( pTerm->eOperator & WO_EQUIV ) zType[1] = 'E'; - if( ExprHasProperty(pTerm->pExpr, EP_FromJoin) ) zType[2] = 'L'; - sqlite3DebugPrintf("TERM-%-3d %p %s cursor=%-3d prob=%-3d op=0x%03x\n", - iTerm, pTerm, zType, pTerm->leftCursor, pTerm->truthProb, - pTerm->eOperator); - sqlite3TreeViewExpr(0, pTerm->pExpr, 0); - } -} -#endif +static void whereExplainTerm(Vdbe *v, WhereTerm *pTerm){ + char zType[4]; + memcpy(zType, "...", 4); + if( pTerm->wtFlags & TERM_VIRTUAL ) zType[0] = 'V'; + if( pTerm->eOperator & WO_EQUIV ) zType[1] = 'E'; + if( ExprHasProperty(pTerm->pExpr, EP_FromJoin) ) zType[2] = 'L'; + sqlite3ExplainPrintf(v, "%s ", zType); + sqlite3ExplainExpr(v, pTerm->pExpr); +} +#endif /* WHERETRACE_ENABLED && SQLITE_ENABLE_TREE_EXPLAIN */ + #ifdef WHERETRACE_ENABLED /* ** Print a WhereLoop object for debugging purposes */ @@ -117324,12 +114673,12 @@ sqlite3DebugPrintf("%c%2d.%0*llx.%0*llx", p->cId, p->iTab, nb, p->maskSelf, nb, p->prereq); sqlite3DebugPrintf(" %12s", pItem->zAlias ? pItem->zAlias : pTab->zName); if( (p->wsFlags & WHERE_VIRTUALTABLE)==0 ){ - const char *zName; - if( p->u.btree.pIndex && (zName = p->u.btree.pIndex->zName)!=0 ){ + const char *zName; + if( p->u.btree.pIndex && (zName = p->u.btree.pIndex->zName)!=0 ){ if( strncmp(zName, "sqlite_autoindex_", 17)==0 ){ int i = sqlite3Strlen30(zName) - 1; while( zName[i]!='_' ) i--; zName += i; } @@ -117346,22 +114695,33 @@ z = sqlite3_mprintf("(%d,%x)", p->u.vtab.idxNum, p->u.vtab.omitMask); } sqlite3DebugPrintf(" %-19s", z); sqlite3_free(z); } - if( p->wsFlags & WHERE_SKIPSCAN ){ - sqlite3DebugPrintf(" f %05x %d-%d", p->wsFlags, p->nLTerm,p->u.btree.nSkip); - }else{ - sqlite3DebugPrintf(" f %05x N %d", p->wsFlags, p->nLTerm); - } + sqlite3DebugPrintf(" f %05x N %d", p->wsFlags, p->nLTerm); sqlite3DebugPrintf(" cost %d,%d,%d\n", p->rSetup, p->rRun, p->nOut); - if( p->nLTerm && (sqlite3WhereTrace & 0x100)!=0 ){ +#ifdef SQLITE_ENABLE_TREE_EXPLAIN + /* If the 0x100 bit of wheretracing is set, then show all of the constraint + ** expressions in the WhereLoop.aLTerm[] array. + */ + if( p->nLTerm && (sqlite3WhereTrace & 0x100)!=0 ){ /* WHERETRACE 0x100 */ int i; + Vdbe *v = pWInfo->pParse->pVdbe; + sqlite3ExplainBegin(v); for(i=0; i nLTerm; i++){ - whereTermPrint(p->aLTerm[i], i); + WhereTerm *pTerm = p->aLTerm[i]; + if( pTerm==0 ) continue; + sqlite3ExplainPrintf(v, " (%d) #%-2d ", i+1, (int)(pTerm-pWC->a)); + sqlite3ExplainPush(v); + whereExplainTerm(v, pTerm); + sqlite3ExplainPop(v); + sqlite3ExplainNL(v); } + sqlite3ExplainFinish(v); + sqlite3DebugPrintf("%s", sqlite3VdbeExplanation(v)); } +#endif } #endif /* ** Convert bulk memory into a valid WhereLoop that can be passed @@ -117739,20 +115099,18 @@ ** ** In the current implementation, the first extra WHERE clause term reduces ** the number of output rows by a factor of 10 and each additional term ** reduces the number of output rows by sqrt(2). */ -static void whereLoopOutputAdjust( - WhereClause *pWC, /* The WHERE clause */ - WhereLoop *pLoop, /* The loop to adjust downward */ - LogEst nRow /* Number of rows in the entire table */ -){ +static void whereLoopOutputAdjust(WhereClause *pWC, WhereLoop *pLoop){ WhereTerm *pTerm, *pX; Bitmask notAllowed = ~(pLoop->prereq|pLoop->maskSelf); int i, j; - int nEq = 0; /* Number of = constraints not within likely()/unlikely() */ + if( !OptimizationEnabled(pWC->pWInfo->pParse->db, SQLITE_AdjustOutEst) ){ + return; + } for(i=pWC->nTerm, pTerm=pWC->a; i>0; i--, pTerm++){ if( (pTerm->wtFlags & TERM_VIRTUAL)!=0 ) break; if( (pTerm->prereqAll & pLoop->maskSelf)==0 ) continue; if( (pTerm->prereqAll & notAllowed)!=0 ) continue; for(j=pLoop->nLTerm-1; j>=0; j--){ @@ -117760,24 +115118,12 @@ if( pX==0 ) continue; if( pX==pTerm ) break; if( pX->iParent>=0 && (&pWC->a[pX->iParent])==pTerm ) break; } if( j<0 ){ - if( pTerm->truthProb<=0 ){ - pLoop->nOut += pTerm->truthProb; - }else{ - pLoop->nOut--; - if( pTerm->eOperator&WO_EQ ) nEq++; - } - } - } - /* TUNING: If there is at least one equality constraint in the WHERE - ** clause that does not have a likelihood() explicitly assigned to it - ** then do not let the estimated number of output rows exceed half - ** the number of rows in the table. */ - if( nEq && pLoop->nOut>nRow-10 ){ - pLoop->nOut = nRow - 10; + pLoop->nOut += (pTerm->truthProb<=0 ? pTerm->truthProb : -1); + } } } /* ** Adjust the cost C by the costMult facter T. This only occurs if @@ -117820,11 +115166,10 @@ u16 saved_nSkip; /* Original value of pNew->u.btree.nSkip */ u32 saved_wsFlags; /* Original value of pNew->wsFlags */ LogEst saved_nOut; /* Original value of pNew->nOut */ int iCol; /* Index of the column in the table */ int rc = SQLITE_OK; /* Return code */ - LogEst rSize; /* Number of rows in the table */ LogEst rLogSize; /* Logarithm of table size */ WhereTerm *pTop = 0, *pBtm = 0; /* Top and bottom range constraints */ pNew = pBuilder->pNew; if( db->mallocFailed ) return SQLITE_NOMEM; @@ -117850,12 +115195,11 @@ saved_nLTerm = pNew->nLTerm; saved_wsFlags = pNew->wsFlags; saved_prereq = pNew->prereq; saved_nOut = pNew->nOut; pNew->rSetup = 0; - rSize = pProbe->aiRowLogEst[0]; - rLogSize = estLog(rSize); + rLogSize = estLog(pProbe->aiRowLogEst[0]); /* Consider using a skip-scan if there are no WHERE clause constraints ** available for the left-most terms of the index, and if the average ** number of repeats in the left-most terms is at least 18. ** @@ -117864,11 +115208,12 @@ ** contains fewer than 2^17 rows we assume otherwise in other parts of ** the code). And, even if it is not, it should not be too much slower. ** On the other hand, the extra seeks could end up being significantly ** more expensive. */ assert( 42==sqlite3LogEst(18) ); - if( saved_nEq==saved_nSkip + if( pTerm==0 + && saved_nEq==saved_nSkip && saved_nEq+1 nKeyCol && pProbe->aiRowLogEst[saved_nEq+1]>=42 /* TUNING: Minimum for skip-scan */ && (rc = whereLoopResize(db, pNew, pNew->nLTerm+1))==SQLITE_OK ){ LogEst nIter; @@ -117875,24 +115220,13 @@ pNew->u.btree.nEq++; pNew->u.btree.nSkip++; pNew->aLTerm[pNew->nLTerm++] = 0; pNew->wsFlags |= WHERE_SKIPSCAN; nIter = pProbe->aiRowLogEst[saved_nEq] - pProbe->aiRowLogEst[saved_nEq+1]; - if( pTerm ){ - /* TUNING: When estimating skip-scan for a term that is also indexable, - ** multiply the cost of the skip-scan by 2.0, to make it a little less - ** desirable than the regular index lookup. */ - nIter += 10; assert( 10==sqlite3LogEst(2) ); - } pNew->nOut -= nIter; - /* TUNING: Because uncertainties in the estimates for skip-scan queries, - ** add a 1.375 fudge factor to make skip-scan slightly less likely. */ - nIter += 5; whereLoopAddBtreeIndex(pBuilder, pSrc, pProbe, nIter + nInMul); pNew->nOut = saved_nOut; - pNew->u.btree.nEq = saved_nEq; - pNew->u.btree.nSkip = saved_nSkip; } for(; rc==SQLITE_OK && pTerm!=0; pTerm = whereScanNext(&scan)){ u16 eOp = pTerm->eOperator; /* Shorthand for pTerm->eOperator */ LogEst rCostIdx; LogEst nOutUnadjusted; /* nOut before IN() and WHERE adjustments */ @@ -118031,11 +115365,11 @@ ApplyCostMultiplier(pNew->rRun, pProbe->pTable->costMult); nOutUnadjusted = pNew->nOut; pNew->rRun += nInMul + nIn; pNew->nOut += nInMul + nIn; - whereLoopOutputAdjust(pBuilder->pWC, pNew, rSize); + whereLoopOutputAdjust(pBuilder->pWC, pNew); rc = whereLoopInsert(pBuilder, pNew); if( pNew->wsFlags & WHERE_COLUMN_RANGE ){ pNew->nOut = saved_nOut; }else{ @@ -118081,11 +115415,10 @@ if( (pOB = pBuilder->pWInfo->pOrderBy)==0 ) return 0; for(ii=0; ii nExpr; ii++){ Expr *pExpr = sqlite3ExprSkipCollate(pOB->a[ii].pExpr); if( pExpr->op!=TK_COLUMN ) return 0; if( pExpr->iTable==iCursor ){ - if( pExpr->iColumn<0 ) return 1; for(jj=0; jj nKeyCol; jj++){ if( pExpr->iColumn==pIndex->aiColumn[jj] ) return 1; } } } @@ -118239,25 +115572,17 @@ pNew->u.btree.nSkip = 0; pNew->u.btree.pIndex = 0; pNew->nLTerm = 1; pNew->aLTerm[0] = pTerm; /* TUNING: One-time cost for computing the automatic index is - ** estimated to be X*N*log2(N) where N is the number of rows in - ** the table being indexed and where X is 7 (LogEst=28) for normal - ** tables or 1.375 (LogEst=4) for views and subqueries. The value - ** of X is smaller for views and subqueries so that the query planner - ** will be more aggressive about generating automatic indexes for - ** those objects, since there is no opportunity to add schema - ** indexes on subqueries and views. */ - pNew->rSetup = rLogSize + rSize + 4; - if( pTab->pSelect==0 && (pTab->tabFlags & TF_Ephemeral)==0 ){ - pNew->rSetup += 24; - } + ** approximately 7*N*log2(N) where N is the number of rows in + ** the table being indexed. */ + pNew->rSetup = rLogSize + rSize + 28; assert( 28==sqlite3LogEst(7) ); ApplyCostMultiplier(pNew->rSetup, pTab->costMult); /* TUNING: Each index lookup yields 20 rows in the table. This ** is more than the usual guess of 10 rows, since we have no way - ** of knowing how selective the index will ultimately be. It would + ** of knowning how selective the index will ultimately be. It would ** not be unreasonable to make this value much larger. */ pNew->nOut = 43; assert( 43==sqlite3LogEst(20) ); pNew->rRun = sqlite3LogEstAdd(rLogSize,pNew->nOut); pNew->wsFlags = WHERE_AUTO_INDEX; pNew->prereq = mExtra | pTerm->prereqRight; @@ -118269,12 +115594,11 @@ /* Loop over all indices */ for(; rc==SQLITE_OK && pProbe; pProbe=pProbe->pNext, iSortIdx++){ if( pProbe->pPartIdxWhere!=0 - && !whereUsablePartialIndex(pSrc->iCursor, pWC, pProbe->pPartIdxWhere) ){ - testcase( pNew->iTab!=pSrc->iCursor ); /* See ticket [98d973b8f5] */ + && !whereUsablePartialIndex(pNew->iTab, pWC, pProbe->pPartIdxWhere) ){ continue; /* Partial index inappropriate for this query */ } rSize = pProbe->aiRowLogEst[0]; pNew->u.btree.nEq = 0; pNew->u.btree.nSkip = 0; @@ -118294,11 +115618,11 @@ /* Full table scan */ pNew->iSortIdx = b ? iSortIdx : 0; /* TUNING: Cost of full table scan is (N*3.0). */ pNew->rRun = rSize + 16; ApplyCostMultiplier(pNew->rRun, pTab->costMult); - whereLoopOutputAdjust(pWC, pNew, rSize); + whereLoopOutputAdjust(pWC, pNew); rc = whereLoopInsert(pBuilder, pNew); pNew->nOut = rSize; if( rc ) break; }else{ Bitmask m; @@ -118330,11 +115654,11 @@ pNew->rRun = rSize + 1 + (15*pProbe->szIdxRow)/pTab->szTabRow; if( m!=0 ){ pNew->rRun = sqlite3LogEstAdd(pNew->rRun, rSize+16); } ApplyCostMultiplier(pNew->rRun, pTab->costMult); - whereLoopOutputAdjust(pWC, pNew, rSize); + whereLoopOutputAdjust(pWC, pNew); rc = whereLoopInsert(pBuilder, pNew); pNew->nOut = rSize; if( rc ) break; } } @@ -118537,10 +115861,11 @@ WhereLoopBuilder sSubBuild; WhereOrSet sSum, sCur; struct SrcList_item *pItem; pWC = pBuilder->pWC; + if( pWInfo->wctrlFlags & WHERE_AND_ONLY ) return SQLITE_OK; pWCEnd = pWC->a + pWC->nTerm; pNew = pBuilder->pNew; memset(&sSum, 0, sizeof(sSum)); pItem = pWInfo->pTabList->a + pNew->iTab; iCur = pItem->iCursor; @@ -118557,11 +115882,10 @@ sSubBuild = *pBuilder; sSubBuild.pOrderBy = 0; sSubBuild.pOrSet = &sCur; - WHERETRACE(0x200, ("Begin processing OR-clause %p\n", pTerm)); for(pOrTerm=pOrWC->a; pOrTerm nTerm; i++){ - whereTermPrint(&sSubBuild.pWC->a[i], i); - } - } -#endif #ifndef SQLITE_OMIT_VIRTUALTABLE if( IsVirtual(pItem->pTab) ){ rc = whereLoopAddVirtual(&sSubBuild, mExtra); }else #endif { rc = whereLoopAddBtree(&sSubBuild, mExtra); } - if( rc==SQLITE_OK ){ - rc = whereLoopAddOr(&sSubBuild, mExtra); - } assert( rc==SQLITE_OK || sCur.n==0 ); if( sCur.n==0 ){ sSum.n = 0; break; }else if( once ){ @@ -118636,11 +115948,10 @@ pNew->rRun = sSum.a[i].rRun + 1; pNew->nOut = sSum.a[i].nOut; pNew->prereq = sSum.a[i].prereq; rc = whereLoopInsert(pBuilder, pNew); } - WHERETRACE(0x200, ("End processing OR-clause %p\n", pTerm)); } } return rc; } @@ -118696,11 +116007,11 @@ ** ** Note that processing for WHERE_GROUPBY and WHERE_DISTINCTBY is not as ** strict. With GROUP BY and DISTINCT the only requirement is that ** equivalent rows appear immediately adjacent to one another. GROUP BY ** and DISTINCT do not require rows to appear in any particular order as long -** as equivalent rows are grouped together. Thus for GROUP BY and DISTINCT +** as equivelent rows are grouped together. Thus for GROUP BY and DISTINCT ** the pOrderBy terms can be matched in any order. With ORDER BY, the ** pOrderBy terms must be matched in strict left-to-right order. */ static i8 wherePathSatisfiesOrderBy( WhereInfo *pWInfo, /* The WHERE clause */ @@ -119703,20 +117014,27 @@ } } /* Construct the WhereLoop objects */ WHERETRACE(0xffff,("*** Optimizer Start ***\n")); -#if defined(WHERETRACE_ENABLED) /* Display all terms of the WHERE clause */ +#if defined(WHERETRACE_ENABLED) && defined(SQLITE_ENABLE_TREE_EXPLAIN) if( sqlite3WhereTrace & 0x100 ){ int i; + Vdbe *v = pParse->pVdbe; + sqlite3ExplainBegin(v); for(i=0; i nTerm; i++){ - whereTermPrint(&sWLB.pWC->a[i], i); + sqlite3ExplainPrintf(v, "#%-2d ", i); + sqlite3ExplainPush(v); + whereExplainTerm(v, &sWLB.pWC->a[i]); + sqlite3ExplainPop(v); + sqlite3ExplainNL(v); } + sqlite3ExplainFinish(v); + sqlite3DebugPrintf("%s", sqlite3VdbeExplanation(v)); } #endif - if( nTabList!=1 || whereShortCut(&sWLB)==0 ){ rc = whereLoopAddAll(&sWLB); if( rc ) goto whereBeginError; /* Display all of the WhereLoop objects if wheretrace is enabled */ @@ -120239,11 +117557,11 @@ /* A routine to convert a binary TK_IS or TK_ISNOT expression into a ** unary TK_ISNULL or TK_NOTNULL expression. */ static void binaryToUnaryIfNull(Parse *pParse, Expr *pY, Expr *pA, int op){ sqlite3 *db = pParse->db; - if( pY && pA && pY->op==TK_NULL ){ + if( db->mallocFailed==0 && pY->op==TK_NULL ){ pA->op = (u8)op; sqlite3ExprDelete(db, pA->pRight); pA->pRight = 0; } } @@ -121466,13 +118784,13 @@ ** ** Outputs: ** A pointer to a parser. This pointer is used in subsequent calls ** to sqlite3Parser and sqlite3ParserFree. */ -SQLITE_PRIVATE void *sqlite3ParserAlloc(void *(*mallocProc)(u64)){ +SQLITE_PRIVATE void *sqlite3ParserAlloc(void *(*mallocProc)(size_t)){ yyParser *pParser; - pParser = (yyParser*)(*mallocProc)( (u64)sizeof(yyParser) ); + pParser = (yyParser*)(*mallocProc)( (size_t)sizeof(yyParser) ); if( pParser ){ pParser->yyidx = -1; #ifdef YYTRACKMAXSTACKDEPTH pParser->yyidxMax = 0; #endif @@ -122498,10 +119816,13 @@ break; case 111: /* cmd ::= select */ { SelectDest dest = {SRT_Output, 0, 0, 0, 0, 0}; sqlite3Select(pParse, yymsp[0].minor.yy3, &dest); + sqlite3ExplainBegin(pParse->pVdbe); + sqlite3ExplainSelect(pParse->pVdbe, yymsp[0].minor.yy3); + sqlite3ExplainFinish(pParse->pVdbe); sqlite3SelectDelete(pParse->db, yymsp[0].minor.yy3); } break; case 112: /* select ::= with selectnowith */ { @@ -122554,34 +119875,10 @@ {yygotominor.yy328 = TK_ALL;} break; case 118: /* oneselect ::= SELECT distinct selcollist from where_opt groupby_opt having_opt orderby_opt limit_opt */ { yygotominor.yy3 = sqlite3SelectNew(pParse,yymsp[-6].minor.yy14,yymsp[-5].minor.yy65,yymsp[-4].minor.yy132,yymsp[-3].minor.yy14,yymsp[-2].minor.yy132,yymsp[-1].minor.yy14,yymsp[-7].minor.yy381,yymsp[0].minor.yy476.pLimit,yymsp[0].minor.yy476.pOffset); -#if SELECTTRACE_ENABLED - /* Populate the Select.zSelName[] string that is used to help with - ** query planner debugging, to differentiate between multiple Select - ** objects in a complex query. - ** - ** If the SELECT keyword is immediately followed by a C-style comment - ** then extract the first few alphanumeric characters from within that - ** comment to be the zSelName value. Otherwise, the label is #N where - ** is an integer that is incremented with each SELECT statement seen. - */ - if( yygotominor.yy3!=0 ){ - const char *z = yymsp[-8].minor.yy0.z+6; - int i; - sqlite3_snprintf(sizeof(yygotominor.yy3->zSelName), yygotominor.yy3->zSelName, "#%d", - ++pParse->nSelect); - while( z[0]==' ' ) z++; - if( z[0]=='/' && z[1]=='*' ){ - z += 2; - while( z[0]==' ' ) z++; - for(i=0; sqlite3Isalnum(z[i]); i++){} - sqlite3_snprintf(sizeof(yygotominor.yy3->zSelName), yygotominor.yy3->zSelName, "%.*s", i, z); - } - } -#endif /* SELECTRACE_ENABLED */ } break; case 120: /* values ::= VALUES LP nexprlist RP */ { yygotominor.yy3 = sqlite3SelectNew(pParse,yymsp[-1].minor.yy14,0,0,0,0,0,SF_Values,0,0); @@ -124044,11 +121341,11 @@ ** ** For EBCDIC, the rules are more complex but have the same ** end result. ** ** Ticket #1066. the SQL standard does not allow '$' in the -** middle of identifiers. But many SQL implementations do. +** middle of identfiers. But many SQL implementations do. ** SQLite will allow '$' in identifiers for compatibility. ** But the feature is undocumented. */ #ifdef SQLITE_ASCII #define IdChar(C) ((sqlite3CtypeMap[(unsigned char)C]&0x46)!=0) @@ -124069,11 +121366,10 @@ 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, /* Ex */ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 0, /* Fx */ }; #define IdChar(C) (((c=C)>=0x42 && sqlite3IsEbcdicIdChar[c-0x40])) #endif -SQLITE_PRIVATE int sqlite3IsIdChar(u8 c){ return IdChar(c); } /* ** Return the length of the token that begins at z[0]. ** Store the token type in *tokenType before returning. @@ -124366,11 +121662,11 @@ } pParse->rc = SQLITE_OK; pParse->zTail = zSql; i = 0; assert( pzErrMsg!=0 ); - pEngine = sqlite3ParserAlloc(sqlite3Malloc); + pEngine = sqlite3ParserAlloc((void*(*)(size_t))sqlite3Malloc); if( pEngine==0 ){ db->mallocFailed = 1; return SQLITE_NOMEM; } assert( pParse->pNewTable==0 ); @@ -124561,21 +121857,21 @@ ** ** (3) EXPLAIN The keyword EXPLAIN has been seen at the beginning of ** a statement. ** ** (4) CREATE The keyword CREATE has been seen at the beginning of a -** statement, possibly preceded by EXPLAIN and/or followed by +** statement, possibly preceeded by EXPLAIN and/or followed by ** TEMP or TEMPORARY ** ** (5) TRIGGER We are in the middle of a trigger definition that must be ** ended by a semicolon, the keyword END, and another semicolon. ** ** (6) SEMI We've seen the first semicolon in the ";END;" that occurs at ** the end of a trigger definition. ** ** (7) END We've seen the ";END" of the ";END;" that occurs at the end -** of a trigger definition. +** of a trigger difinition. ** ** Transitions between states above are determined by tokens extracted ** from the input. The following tokens are significant: ** ** (0) tkSEMI A semicolon. @@ -124614,11 +121910,11 @@ /* 6 SEMI: */ { 6, 6, 5, 5, 5, 5, 5, 7, }, /* 7 END: */ { 1, 7, 5, 5, 5, 5, 5, 5, }, }; #else /* If triggers are not supported by this compile then the statement machine - ** used to detect the end of a statement is much simpler + ** used to detect the end of a statement is much simplier */ static const u8 trans[3][3] = { /* Token: */ /* State: ** SEMI WS OTHER */ /* 0 INVALID: */ { 1, 0, 2, }, @@ -125340,15 +122636,10 @@ sqlite3GlobalConfig.xLog = va_arg(ap, LOGFUNC_t); sqlite3GlobalConfig.pLogArg = va_arg(ap, void*); break; } - /* EVIDENCE-OF: R-55548-33817 The compile-time setting for URI filenames - ** can be changed at start-time using the - ** sqlite3_config(SQLITE_CONFIG_URI,1) or - ** sqlite3_config(SQLITE_CONFIG_URI,0) configuration calls. - */ case SQLITE_CONFIG_URI: { sqlite3GlobalConfig.bOpenUri = va_arg(ap, int); break; } @@ -125721,11 +123012,11 @@ /* Legacy behavior (sqlite3_close() behavior) is to return ** SQLITE_BUSY if the connection can not be closed immediately. */ if( !forceZombie && connectionIsBusy(db) ){ - sqlite3ErrorWithMsg(db, SQLITE_BUSY, "unable to close due to unfinalized " + sqlite3Error(db, SQLITE_BUSY, "unable to close due to unfinalized " "statements or unfinished backups"); sqlite3_mutex_leave(db->mutex); return SQLITE_BUSY; } @@ -125851,17 +123142,13 @@ sqlite3DbFree(db, pMod); } sqlite3HashClear(&db->aModule); #endif - sqlite3Error(db, SQLITE_OK); /* Deallocates any cached error strings. */ + sqlite3Error(db, SQLITE_OK, 0); /* Deallocates any cached error strings. */ sqlite3ValueFree(db->pErr); sqlite3CloseExtensions(db); -#if SQLITE_USER_AUTHENTICATION - sqlite3_free(db->auth.zAuthUser); - sqlite3_free(db->auth.zAuthPW); -#endif db->magic = SQLITE_MAGIC_ERROR; /* The temp-database schema is allocated differently from the other schema ** objects (using sqliteMalloc() directly, instead of sqlite3BtreeSchema()). @@ -126288,11 +123575,11 @@ ** operation to continue but invalidate all precompiled statements. */ p = sqlite3FindFunction(db, zFunctionName, nName, nArg, (u8)enc, 0); if( p && (p->funcFlags & SQLITE_FUNC_ENCMASK)==enc && p->nArg==nArg ){ if( db->nVdbeActive ){ - sqlite3ErrorWithMsg(db, SQLITE_BUSY, + sqlite3Error(db, SQLITE_BUSY, "unable to delete/modify user-function due to active statements"); assert( !db->mallocFailed ); return SQLITE_BUSY; }else{ sqlite3ExpirePreparedStatements(db); @@ -126626,14 +123913,14 @@ if( zDb && zDb[0] ){ iDb = sqlite3FindDbName(db, zDb); } if( iDb<0 ){ rc = SQLITE_ERROR; - sqlite3ErrorWithMsg(db, SQLITE_ERROR, "unknown database: %s", zDb); + sqlite3Error(db, SQLITE_ERROR, "unknown database: %s", zDb); }else{ rc = sqlite3Checkpoint(db, iDb, eMode, pnLog, pnCkpt); - sqlite3Error(db, rc); + sqlite3Error(db, rc, 0); } rc = sqlite3ApiExit(db, rc); sqlite3_mutex_leave(db->mutex); return rc; #endif @@ -126784,11 +124071,11 @@ if( db->mallocFailed ){ z = (void *)outOfMem; }else{ z = sqlite3_value_text16(db->pErr); if( z==0 ){ - sqlite3ErrorWithMsg(db, db->errCode, sqlite3ErrStr(db->errCode)); + sqlite3Error(db, db->errCode, sqlite3ErrStr(db->errCode)); z = sqlite3_value_text16(db->pErr); } /* A malloc() may have failed within the call to sqlite3_value_text16() ** above. If this is the case, then the db->mallocFailed flag needs to ** be cleared before returning. Do this directly, instead of via @@ -126871,10 +124158,11 @@ int(*xCompare)(void*,int,const void*,int,const void*), void(*xDel)(void*) ){ CollSeq *pColl; int enc2; + int nName = sqlite3Strlen30(zName); assert( sqlite3_mutex_held(db->mutex) ); /* If SQLITE_UTF16 is specified as the encoding type, transform this ** to one of SQLITE_UTF16LE or SQLITE_UTF16BE using the @@ -126895,11 +124183,11 @@ ** are no active VMs, invalidate any pre-compiled statements. */ pColl = sqlite3FindCollSeq(db, (u8)enc2, zName, 0); if( pColl && pColl->xCmp ){ if( db->nVdbeActive ){ - sqlite3ErrorWithMsg(db, SQLITE_BUSY, + sqlite3Error(db, SQLITE_BUSY, "unable to delete/modify collation sequence due to active statements"); return SQLITE_BUSY; } sqlite3ExpirePreparedStatements(db); invalidateCachedKeyInfo(db); @@ -126909,11 +124197,11 @@ ** then any copies made by synthCollSeq() need to be invalidated. ** Also, collation destructor - CollSeq.xDel() - function may need ** to be called. */ if( (pColl->enc & ~SQLITE_UTF16_ALIGNED)==enc2 ){ - CollSeq *aColl = sqlite3HashFind(&db->aCollSeq, zName); + CollSeq *aColl = sqlite3HashFind(&db->aCollSeq, zName, nName); int j; for(j=0; j<3; j++){ CollSeq *p = &aColl[j]; if( p->enc==pColl->enc ){ if( p->xDel ){ @@ -126929,11 +124217,11 @@ if( pColl==0 ) return SQLITE_NOMEM; pColl->xCmp = xCompare; pColl->pUser = pCtx; pColl->xDel = xDel; pColl->enc = (u8)(enc2 | (enc & SQLITE_UTF16_ALIGNED)); - sqlite3Error(db, SQLITE_OK); + sqlite3Error(db, SQLITE_OK, 0); return SQLITE_OK; } /* @@ -126951,11 +124239,10 @@ SQLITE_MAX_FUNCTION_ARG, SQLITE_MAX_ATTACHED, SQLITE_MAX_LIKE_PATTERN_LENGTH, SQLITE_MAX_VARIABLE_NUMBER, /* IMP: R-38091-32352 */ SQLITE_MAX_TRIGGER_DEPTH, - SQLITE_MAX_WORKER_THREADS, }; /* ** Make sure the hard limits are set to reasonable values */ @@ -126987,13 +124274,10 @@ # error SQLITE_MAX_COLUMN must not exceed 32767 #endif #if SQLITE_MAX_TRIGGER_DEPTH<1 # error SQLITE_MAX_TRIGGER_DEPTH must be at least 1 #endif -#if SQLITE_MAX_WORKER_THREADS<0 || SQLITE_MAX_WORKER_THREADS>50 -# error SQLITE_MAX_WORKER_THREADS must be between 0 and 50 -#endif /* ** Change the value of a limit. Report the old value. ** If an invalid limit index is supplied, report -1. @@ -127023,12 +124307,11 @@ assert( aHardLimit[SQLITE_LIMIT_ATTACHED]==SQLITE_MAX_ATTACHED ); assert( aHardLimit[SQLITE_LIMIT_LIKE_PATTERN_LENGTH]== SQLITE_MAX_LIKE_PATTERN_LENGTH ); assert( aHardLimit[SQLITE_LIMIT_VARIABLE_NUMBER]==SQLITE_MAX_VARIABLE_NUMBER); assert( aHardLimit[SQLITE_LIMIT_TRIGGER_DEPTH]==SQLITE_MAX_TRIGGER_DEPTH ); - assert( aHardLimit[SQLITE_LIMIT_WORKER_THREADS]==SQLITE_MAX_WORKER_THREADS ); - assert( SQLITE_LIMIT_WORKER_THREADS==(SQLITE_N_LIMIT-1) ); + assert( SQLITE_LIMIT_TRIGGER_DEPTH==(SQLITE_N_LIMIT-1) ); if( limitId<0 || limitId>=SQLITE_N_LIMIT ){ return -1; } @@ -127082,11 +124365,11 @@ int nUri = sqlite3Strlen30(zUri); assert( *pzErrMsg==0 ); if( ((flags & SQLITE_OPEN_URI) || sqlite3GlobalConfig.bOpenUri) - && nUri>=5 && memcmp(zUri, "file:", 5)==0 /* IMP: R-57884-37496 */ + && nUri>=5 && memcmp(zUri, "file:", 5)==0 ){ char *zOpt; int eState; /* Parser state when parsing URI */ int iIn; /* Input character index */ int iOut = 0; /* Output character index */ @@ -127312,13 +124595,11 @@ assert( SQLITE_OPEN_READWRITE == 0x02 ); assert( SQLITE_OPEN_CREATE == 0x04 ); testcase( (1<<(flags&7))==0x02 ); /* READONLY */ testcase( (1<<(flags&7))==0x04 ); /* READWRITE */ testcase( (1<<(flags&7))==0x40 ); /* READWRITE | CREATE */ - if( ((1<<(flags&7)) & 0x46)==0 ){ - return SQLITE_MISUSE_BKPT; /* IMP: R-65497-44594 */ - } + if( ((1<<(flags&7)) & 0x46)==0 ) return SQLITE_MISUSE_BKPT; if( sqlite3GlobalConfig.bCoreMutex==0 ){ isThreadsafe = 0; }else if( flags & SQLITE_OPEN_NOMUTEX ){ isThreadsafe = 0; @@ -127373,16 +124654,14 @@ db->magic = SQLITE_MAGIC_BUSY; db->aDb = db->aDbStatic; assert( sizeof(db->aLimit)==sizeof(aHardLimit) ); memcpy(db->aLimit, aHardLimit, sizeof(db->aLimit)); - db->aLimit[SQLITE_LIMIT_WORKER_THREADS] = SQLITE_DEFAULT_WORKER_THREADS; db->autoCommit = 1; db->nextAutovac = -1; db->szMmap = sqlite3GlobalConfig.szMmap; db->nextPagesize = 0; - db->nMaxSorterMmap = 0x7FFFFFFF; db->flags |= SQLITE_ShortColNames | SQLITE_EnableTrigger | SQLITE_CacheSpill #if !defined(SQLITE_DEFAULT_AUTOMATIC_INDEX) || SQLITE_DEFAULT_AUTOMATIC_INDEX | SQLITE_AutoIndex #endif #if SQLITE_DEFAULT_FILE_FORMAT<4 @@ -127423,11 +124702,11 @@ /* Parse the filename/URI argument. */ db->openFlags = flags; rc = sqlite3ParseUri(zVfs, zFilename, &flags, &db->pVfs, &zOpen, &zErrMsg); if( rc!=SQLITE_OK ){ if( rc==SQLITE_NOMEM ) db->mallocFailed = 1; - sqlite3ErrorWithMsg(db, rc, zErrMsg ? "%s" : 0, zErrMsg); + sqlite3Error(db, rc, zErrMsg ? "%s" : 0, zErrMsg); sqlite3_free(zErrMsg); goto opendb_out; } /* Open the backend database driver */ @@ -127435,15 +124714,16 @@ flags | SQLITE_OPEN_MAIN_DB); if( rc!=SQLITE_OK ){ if( rc==SQLITE_IOERR_NOMEM ){ rc = SQLITE_NOMEM; } - sqlite3Error(db, rc); + sqlite3Error(db, rc, 0); goto opendb_out; } db->aDb[0].pSchema = sqlite3SchemaGet(db, db->aDb[0].pBt); db->aDb[1].pSchema = sqlite3SchemaGet(db, 0); + /* The default safety_level for the main database is 'full'; for the temp ** database it is 'NONE'. This matches the pager layer defaults. */ db->aDb[0].zName = "main"; @@ -127458,11 +124738,11 @@ /* Register all built-in functions, but do not attempt to read the ** database schema yet. This is delayed until the first time the database ** is accessed. */ - sqlite3Error(db, SQLITE_OK); + sqlite3Error(db, SQLITE_OK, 0); sqlite3RegisterBuiltinFunctions(db); /* Load automatic extensions - extensions that have been registered ** using the sqlite3_automatic_extension() API. */ @@ -127515,11 +124795,11 @@ db->dfltLockMode = SQLITE_DEFAULT_LOCKING_MODE; sqlite3PagerLockingMode(sqlite3BtreePager(db->aDb[0].pBt), SQLITE_DEFAULT_LOCKING_MODE); #endif - if( rc ) sqlite3Error(db, rc); + if( rc ) sqlite3Error(db, rc, 0); /* Enable the lookaside-malloc subsystem */ setupLookaside(db, 0, sqlite3GlobalConfig.szLookaside, sqlite3GlobalConfig.nLookaside); @@ -127727,13 +125007,13 @@ SQLITE_API int sqlite3_get_autocommit(sqlite3 *db){ return db->autoCommit; } /* -** The following routines are substitutes for constants SQLITE_CORRUPT, +** The following routines are subtitutes for constants SQLITE_CORRUPT, ** SQLITE_MISUSE, SQLITE_CANTOPEN, SQLITE_IOERR and possibly other error -** constants. They serve two purposes: +** constants. They server two purposes: ** ** 1. Serve as a convenient place to set a breakpoint in a debugger ** to detect when version error conditions occurs. ** ** 2. Invoke sqlite3_log() to provide the source code location where @@ -127877,11 +125157,11 @@ sqlite3DbFree(db, zErrMsg); zErrMsg = sqlite3MPrintf(db, "no such table column: %s.%s", zTableName, zColumnName); rc = SQLITE_ERROR; } - sqlite3ErrorWithMsg(db, rc, (zErrMsg?"%s":0), zErrMsg); + sqlite3Error(db, rc, (zErrMsg?"%s":0), zErrMsg); sqlite3DbFree(db, zErrMsg); rc = sqlite3ApiExit(db, rc); sqlite3_mutex_leave(db->mutex); return rc; } @@ -128043,11 +125323,11 @@ ** as it existing before this routine was called. ** ** IMPORTANT: Changing the PENDING byte from 0x40000000 results in ** an incompatible database file format. Changing the PENDING byte ** while any database connection is open results in undefined and - ** deleterious behavior. + ** dileterious behavior. */ case SQLITE_TESTCTRL_PENDING_BYTE: { rc = PENDING_BYTE; #ifndef SQLITE_OMIT_WSD { @@ -128197,10 +125477,26 @@ */ case SQLITE_TESTCTRL_LOCALTIME_FAULT: { sqlite3GlobalConfig.bLocaltimeFault = va_arg(ap, int); break; } + +#if defined(SQLITE_ENABLE_TREE_EXPLAIN) + /* sqlite3_test_control(SQLITE_TESTCTRL_EXPLAIN_STMT, + ** sqlite3_stmt*,const char**); + ** + ** If compiled with SQLITE_ENABLE_TREE_EXPLAIN, each sqlite3_stmt holds + ** a string that describes the optimized parse tree. This test-control + ** returns a pointer to that string. + */ + case SQLITE_TESTCTRL_EXPLAIN_STMT: { + sqlite3_stmt *pStmt = va_arg(ap, sqlite3_stmt*); + const char **pzRet = va_arg(ap, const char**); + *pzRet = sqlite3VdbeExplanation((Vdbe*)pStmt); + break; + } +#endif /* sqlite3_test_control(SQLITE_TESTCTRL_NEVER_CORRUPT, int); ** ** Set or clear a flag that indicates that the database file is always well- ** formed and never corrupt. This flag is clear by default, indicating that @@ -128226,26 +125522,20 @@ sqlite3GlobalConfig.pVdbeBranchArg = va_arg(ap,void*); #endif break; } - /* sqlite3_test_control(SQLITE_TESTCTRL_SORTER_MMAP, db, nMax); */ - case SQLITE_TESTCTRL_SORTER_MMAP: { - sqlite3 *db = va_arg(ap, sqlite3*); - db->nMaxSorterMmap = va_arg(ap, int); - break; - } - /* sqlite3_test_control(SQLITE_TESTCTRL_ISINIT); ** ** Return SQLITE_OK if SQLite has been initialized and SQLITE_ERROR if ** not. */ case SQLITE_TESTCTRL_ISINIT: { if( sqlite3GlobalConfig.isInit==0 ) rc = SQLITE_ERROR; break; } + } va_end(ap); #endif /* SQLITE_OMIT_BUILTIN_TEST */ return rc; } @@ -128515,11 +125805,11 @@ } } leaveMutex(); assert( !db->mallocFailed ); - sqlite3ErrorWithMsg(db, rc, (rc?"database is deadlocked":0)); + sqlite3Error(db, rc, (rc?"database is deadlocked":0)); sqlite3_mutex_leave(db->mutex); return rc; } /* @@ -133967,11 +131257,11 @@ if( a[i].bIgnore==0 && (bMaxSet==0 || DOCID_CMP(iMax, a[i].iDocid)<0) ){ iMax = a[i].iDocid; bMaxSet = 1; } } - assert( rc!=SQLITE_OK || (p->nToken>=1 && a[p->nToken-1].bIgnore==0) ); + assert( rc!=SQLITE_OK || a[p->nToken-1].bIgnore==0 ); assert( rc!=SQLITE_OK || bMaxSet ); /* Keep advancing iterators until they all point to the same document */ for(i=0; i nToken; i++){ while( rc==SQLITE_OK && bEof==0 Index: SQLite.Interop/src/core/sqlite3.h ================================================================== --- SQLite.Interop/src/core/sqlite3.h +++ SQLite.Interop/src/core/sqlite3.h @@ -105,13 +105,13 @@ ** ** See also: [sqlite3_libversion()], ** [sqlite3_libversion_number()], [sqlite3_sourceid()], ** [sqlite_version()] and [sqlite_source_id()]. */ -#define SQLITE_VERSION "3.8.7" -#define SQLITE_VERSION_NUMBER 3008007 -#define SQLITE_SOURCE_ID "2014-10-07 23:02:24 f672a380e2e52bede95ff11a533fd9f7d412d494" +#define SQLITE_VERSION "3.8.6" +#define SQLITE_VERSION_NUMBER 3008006 +#define SQLITE_SOURCE_ID "2014-08-15 11:46:33 9491ba7d738528f168657adb43a198238abde19e" /* ** CAPI3REF: Run-Time Library Version Numbers ** KEYWORDS: sqlite3_version, sqlite3_sourceid ** @@ -495,11 +495,10 @@ #define SQLITE_CONSTRAINT_VTAB (SQLITE_CONSTRAINT | (9<<8)) #define SQLITE_CONSTRAINT_ROWID (SQLITE_CONSTRAINT |(10<<8)) #define SQLITE_NOTICE_RECOVER_WAL (SQLITE_NOTICE | (1<<8)) #define SQLITE_NOTICE_RECOVER_ROLLBACK (SQLITE_NOTICE | (2<<8)) #define SQLITE_WARNING_AUTOINDEX (SQLITE_WARNING | (1<<8)) -#define SQLITE_AUTH_USER (SQLITE_AUTH | (1<<8)) /* ** CAPI3REF: Flags For File Open Operations ** ** These bit values are intended for use in the @@ -2098,11 +2097,11 @@ ** ** ^Calling this routine with an argument less than or equal to zero ** turns off all busy handlers. ** ** ^(There can only be a single busy handler for a particular -** [database connection] at any given moment. If another busy handler +** [database connection] any any given moment. If another busy handler ** was defined (using [sqlite3_busy_handler()]) prior to calling ** this routine, that other busy handler is cleared.)^ ** ** See also: [PRAGMA busy_timeout] */ @@ -2302,14 +2301,10 @@ ** ^If sqlite3_malloc() is unable to obtain sufficient free ** memory, it returns a NULL pointer. ^If the parameter N to ** sqlite3_malloc() is zero or negative then sqlite3_malloc() returns ** a NULL pointer. ** -** ^The sqlite3_malloc64(N) routine works just like -** sqlite3_malloc(N) except that N is an unsigned 64-bit integer instead -** of a signed 32-bit integer. -** ** ^Calling sqlite3_free() with a pointer previously returned ** by sqlite3_malloc() or sqlite3_realloc() releases that memory so ** that it might be reused. ^The sqlite3_free() routine is ** a no-op if is called with a NULL pointer. Passing a NULL pointer ** to sqlite3_free() is harmless. After being freed, memory @@ -2317,42 +2312,28 @@ ** memory might result in a segmentation fault or other severe error. ** Memory corruption, a segmentation fault, or other severe error ** might result if sqlite3_free() is called with a non-NULL pointer that ** was not obtained from sqlite3_malloc() or sqlite3_realloc(). ** -** ^The sqlite3_realloc(X,N) interface attempts to resize a -** prior memory allocation X to be at least N bytes. -** ^If the X parameter to sqlite3_realloc(X,N) +** ^(The sqlite3_realloc() interface attempts to resize a +** prior memory allocation to be at least N bytes, where N is the +** second parameter. The memory allocation to be resized is the first +** parameter.)^ ^ If the first parameter to sqlite3_realloc() ** is a NULL pointer then its behavior is identical to calling -** sqlite3_malloc(N). -** ^If the N parameter to sqlite3_realloc(X,N) is zero or +** sqlite3_malloc(N) where N is the second parameter to sqlite3_realloc(). +** ^If the second parameter to sqlite3_realloc() is zero or ** negative then the behavior is exactly the same as calling -** sqlite3_free(X). -** ^sqlite3_realloc(X,N) returns a pointer to a memory allocation -** of at least N bytes in size or NULL if insufficient memory is available. +** sqlite3_free(P) where P is the first parameter to sqlite3_realloc(). +** ^sqlite3_realloc() returns a pointer to a memory allocation +** of at least N bytes in size or NULL if sufficient memory is unavailable. ** ^If M is the size of the prior allocation, then min(N,M) bytes ** of the prior allocation are copied into the beginning of buffer returned -** by sqlite3_realloc(X,N) and the prior allocation is freed. -** ^If sqlite3_realloc(X,N) returns NULL and N is positive, then the -** prior allocation is not freed. -** -** ^The sqlite3_realloc64(X,N) interfaces works the same as -** sqlite3_realloc(X,N) except that N is a 64-bit unsigned integer instead -** of a 32-bit signed integer. -** -** ^If X is a memory allocation previously obtained from sqlite3_malloc(), -** sqlite3_malloc64(), sqlite3_realloc(), or sqlite3_realloc64(), then -** sqlite3_msize(X) returns the size of that memory allocation in bytes. -** ^The value returned by sqlite3_msize(X) might be larger than the number -** of bytes requested when X was allocated. ^If X is a NULL pointer then -** sqlite3_msize(X) returns zero. If X points to something that is not -** the beginning of memory allocation, or if it points to a formerly -** valid memory allocation that has now been freed, then the behavior -** of sqlite3_msize(X) is undefined and possibly harmful. -** -** ^The memory returned by sqlite3_malloc(), sqlite3_realloc(), -** sqlite3_malloc64(), and sqlite3_realloc64() +** by sqlite3_realloc() and the prior allocation is freed. +** ^If sqlite3_realloc() returns NULL, then the prior allocation +** is not freed. +** +** ^The memory returned by sqlite3_malloc() and sqlite3_realloc() ** is always aligned to at least an 8 byte boundary, or to a ** 4 byte boundary if the [SQLITE_4_BYTE_ALIGNED_MALLOC] compile-time ** option is used. ** ** In SQLite version 3.5.0 and 3.5.1, it was possible to define @@ -2376,15 +2357,12 @@ ** The application must not read or write any part of ** a block of memory after it has been released using ** [sqlite3_free()] or [sqlite3_realloc()]. */ SQLITE_API void *sqlite3_malloc(int); -SQLITE_API void *sqlite3_malloc64(sqlite3_uint64); SQLITE_API void *sqlite3_realloc(void*, int); -SQLITE_API void *sqlite3_realloc64(void*, sqlite3_uint64); SQLITE_API void sqlite3_free(void*); -SQLITE_API sqlite3_uint64 sqlite3_msize(void*); /* ** CAPI3REF: Memory Allocator Statistics ** ** SQLite provides these two interfaces for reporting on the status @@ -2667,13 +2645,13 @@ ** [SQLITE_OK] is returned. Otherwise an [error code] is returned.)^ ^The ** [sqlite3_errmsg()] or [sqlite3_errmsg16()] routines can be used to obtain ** an English language description of the error following a failure of any ** of the sqlite3_open() routines. ** -** ^The default encoding will be UTF-8 for databases created using -** sqlite3_open() or sqlite3_open_v2(). ^The default encoding for databases -** created using sqlite3_open16() will be UTF-16 in the native byte order. +** ^The default encoding for the database will be UTF-8 if +** sqlite3_open() or sqlite3_open_v2() is called and +** UTF-16 in the native byte order if sqlite3_open16() is used. ** ** Whether or not an error occurs when it is opened, resources ** associated with the [database connection] handle should be released by ** passing it to [sqlite3_close()] when it is no longer required. ** @@ -2757,18 +2735,17 @@ ** ^SQLite uses the path component of the URI as the name of the disk file ** which contains the database. ^If the path begins with a '/' character, ** then it is interpreted as an absolute path. ^If the path does not begin ** with a '/' (meaning that the authority section is omitted from the URI) ** then the path is interpreted as a relative path. -** ^(On windows, the first component of an absolute path -** is a drive specification (e.g. "C:").)^ +** ^On windows, the first component of an absolute path +** is a drive specification (e.g. "C:"). ** ** [[core URI query parameters]] ** The query component of a URI may contain parameters that are interpreted ** either by SQLite itself, or by a [VFS | custom VFS implementation]. -** SQLite and its built-in [VFSes] interpret the -** following query parameters: +** SQLite interprets the following three query parameters: ** ** -
**
- vfs: ^The "vfs" parameter may be used to specify the name of ** a VFS object that provides the operating system interface that should ** be used to access the database file on disk. ^If this option is set to @@ -2799,13 +2776,15 @@ ** equivalent to setting the SQLITE_OPEN_PRIVATECACHE bit. ** ^If sqlite3_open_v2() is used and the "cache" parameter is present in ** a URI filename, its value overrides any behavior requested by setting ** SQLITE_OPEN_PRIVATECACHE or SQLITE_OPEN_SHAREDCACHE flag. ** -**
- psow: ^The psow parameter indicates whether or not the +**
- psow: ^The psow parameter may be "true" (or "on" or "yes" or +** "1") or "false" (or "off" or "no" or "0") to indicate that the ** [powersafe overwrite] property does or does not apply to the -** storage media on which the database file resides. +** storage media on which the database file resides. ^The psow query +** parameter only works for the built-in unix and Windows VFSes. ** **
- nolock: ^The nolock parameter is a boolean query parameter
** which if set disables file locking in rollback journal modes. This
** is useful for accessing a database on a filesystem that does not
** support locking. Caution: Database corruption might result if two
@@ -3097,14 +3076,10 @@
** ^(
- SQLITE_LIMIT_VARIABLE_NUMBER
**- The maximum index number of any [parameter] in an SQL statement.)^ ** ** [[SQLITE_LIMIT_TRIGGER_DEPTH]] ^(
- SQLITE_LIMIT_TRIGGER_DEPTH
**- The maximum depth of recursion for triggers.
)^ -** -** [[SQLITE_LIMIT_WORKER_THREADS]] ^(- SQLITE_LIMIT_WORKER_THREADS
-**- The maximum number of auxiliary worker threads that a single -** [prepared statement] may start.
)^ ** */ #define SQLITE_LIMIT_LENGTH 0 #define SQLITE_LIMIT_SQL_LENGTH 1 #define SQLITE_LIMIT_COLUMN 2 @@ -3114,11 +3089,10 @@ #define SQLITE_LIMIT_FUNCTION_ARG 6 #define SQLITE_LIMIT_ATTACHED 7 #define SQLITE_LIMIT_LIKE_PATTERN_LENGTH 8 #define SQLITE_LIMIT_VARIABLE_NUMBER 9 #define SQLITE_LIMIT_TRIGGER_DEPTH 10 -#define SQLITE_LIMIT_WORKER_THREADS 11 /* ** CAPI3REF: Compiling An SQL Statement ** KEYWORDS: {SQL statement compiler} ** @@ -3388,37 +3362,29 @@ ** is negative, then the length of the string is ** the number of bytes up to the first zero terminator. ** If the fourth parameter to sqlite3_bind_blob() is negative, then ** the behavior is undefined. ** If a non-negative fourth parameter is provided to sqlite3_bind_text() -** or sqlite3_bind_text16() or sqlite3_bind_text64() then -** that parameter must be the byte offset +** or sqlite3_bind_text16() then that parameter must be the byte offset ** where the NUL terminator would occur assuming the string were NUL ** terminated. If any NUL characters occur at byte offsets less than ** the value of the fourth parameter then the resulting string value will ** contain embedded NULs. The result of expressions involving strings ** with embedded NULs is undefined. ** -** ^The fifth argument to the BLOB and string binding interfaces -** is a destructor used to dispose of the BLOB or +** ^The fifth argument to sqlite3_bind_blob(), sqlite3_bind_text(), and +** sqlite3_bind_text16() is a destructor used to dispose of the BLOB or ** string after SQLite has finished with it. ^The destructor is called -** to dispose of the BLOB or string even if the call to bind API fails. +** to dispose of the BLOB or string even if the call to sqlite3_bind_blob(), +** sqlite3_bind_text(), or sqlite3_bind_text16() fails. ** ^If the fifth argument is ** the special value [SQLITE_STATIC], then SQLite assumes that the ** information is in static, unmanaged space and does not need to be freed. ** ^If the fifth argument has the value [SQLITE_TRANSIENT], then ** SQLite makes its own private copy of the data immediately, before ** the sqlite3_bind_*() routine returns. ** -** ^The sixth argument to sqlite3_bind_text64() must be one of -** [SQLITE_UTF8], [SQLITE_UTF16], [SQLITE_UTF16BE], or [SQLITE_UTF16LE] -** to specify the encoding of the text in the third parameter. If -** the sixth argument to sqlite3_bind_text64() is not one of the -** allowed values shown above, or if the text encoding is different -** from the encoding specified by the sixth parameter, then the behavior -** is undefined. -** ** ^The sqlite3_bind_zeroblob() routine binds a BLOB of length N that ** is filled with zeroes. ^A zeroblob uses a fixed amount of memory ** (just an integer to hold its size) while it is being processed. ** Zeroblobs are intended to serve as placeholders for BLOBs whose ** content is later written using @@ -3435,30 +3401,23 @@ ** ^Bindings are not cleared by the [sqlite3_reset()] routine. ** ^Unbound parameters are interpreted as NULL. ** ** ^The sqlite3_bind_* routines return [SQLITE_OK] on success or an ** [error code] if anything goes wrong. -** ^[SQLITE_TOOBIG] might be returned if the size of a string or BLOB -** exceeds limits imposed by [sqlite3_limit]([SQLITE_LIMIT_LENGTH]) or -** [SQLITE_MAX_LENGTH]. ** ^[SQLITE_RANGE] is returned if the parameter ** index is out of range. ^[SQLITE_NOMEM] is returned if malloc() fails. ** ** See also: [sqlite3_bind_parameter_count()], ** [sqlite3_bind_parameter_name()], and [sqlite3_bind_parameter_index()]. */ SQLITE_API int sqlite3_bind_blob(sqlite3_stmt*, int, const void*, int n, void(*)(void*)); -SQLITE_API int sqlite3_bind_blob64(sqlite3_stmt*, int, const void*, sqlite3_uint64, - void(*)(void*)); SQLITE_API int sqlite3_bind_double(sqlite3_stmt*, int, double); SQLITE_API int sqlite3_bind_int(sqlite3_stmt*, int, int); SQLITE_API int sqlite3_bind_int64(sqlite3_stmt*, int, sqlite3_int64); SQLITE_API int sqlite3_bind_null(sqlite3_stmt*, int); -SQLITE_API int sqlite3_bind_text(sqlite3_stmt*,int,const char*,int,void(*)(void*)); +SQLITE_API int sqlite3_bind_text(sqlite3_stmt*, int, const char*, int n, void(*)(void*)); SQLITE_API int sqlite3_bind_text16(sqlite3_stmt*, int, const void*, int, void(*)(void*)); -SQLITE_API int sqlite3_bind_text64(sqlite3_stmt*, int, const char*, sqlite3_uint64, - void(*)(void*), unsigned char encoding); SQLITE_API int sqlite3_bind_value(sqlite3_stmt*, int, const sqlite3_value*); SQLITE_API int sqlite3_bind_zeroblob(sqlite3_stmt*, int, int n); /* ** CAPI3REF: Number Of SQL Parameters @@ -4203,11 +4162,11 @@ ** These routines work only with [protected sqlite3_value] objects. ** Any attempt to use these routines on an [unprotected sqlite3_value] ** object results in undefined behavior. ** ** ^These routines work just like the corresponding [column access functions] -** except that these routines take a single [protected sqlite3_value] object +** except that these routines take a single [protected sqlite3_value] object ** pointer instead of a [sqlite3_stmt*] pointer and an integer column number. ** ** ^The sqlite3_value_text16() interface extracts a UTF-16 string ** in the native byte-order of the host machine. ^The ** sqlite3_value_text16be() and sqlite3_value_text16le() interfaces @@ -4450,14 +4409,10 @@ ** ^The sqlite3_result_text(), sqlite3_result_text16(), ** sqlite3_result_text16le(), and sqlite3_result_text16be() interfaces ** set the return value of the application-defined function to be ** a text string which is represented as UTF-8, UTF-16 native byte order, ** UTF-16 little endian, or UTF-16 big endian, respectively. -** ^The sqlite3_result_text64() interface sets the return value of an -** application-defined function to be a text string in an encoding -** specified by the fifth (and last) parameter, which must be one -** of [SQLITE_UTF8], [SQLITE_UTF16], [SQLITE_UTF16BE], or [SQLITE_UTF16LE]. ** ^SQLite takes the text result from the application from ** the 2nd parameter of the sqlite3_result_text* interfaces. ** ^If the 3rd parameter to the sqlite3_result_text* interfaces ** is negative, then SQLite takes result text from the 2nd parameter ** through the first zero character. @@ -4497,11 +4452,10 @@ ** If these routines are called from within the different thread ** than the one containing the application-defined function that received ** the [sqlite3_context] pointer, the results are undefined. */ SQLITE_API void sqlite3_result_blob(sqlite3_context*, const void*, int, void(*)(void*)); -SQLITE_API void sqlite3_result_blob64(sqlite3_context*,const void*,sqlite3_uint64,void(*)(void*)); SQLITE_API void sqlite3_result_double(sqlite3_context*, double); SQLITE_API void sqlite3_result_error(sqlite3_context*, const char*, int); SQLITE_API void sqlite3_result_error16(sqlite3_context*, const void*, int); SQLITE_API void sqlite3_result_error_toobig(sqlite3_context*); SQLITE_API void sqlite3_result_error_nomem(sqlite3_context*); @@ -4508,12 +4462,10 @@ SQLITE_API void sqlite3_result_error_code(sqlite3_context*, int); SQLITE_API void sqlite3_result_int(sqlite3_context*, int); SQLITE_API void sqlite3_result_int64(sqlite3_context*, sqlite3_int64); SQLITE_API void sqlite3_result_null(sqlite3_context*); SQLITE_API void sqlite3_result_text(sqlite3_context*, const char*, int, void(*)(void*)); -SQLITE_API void sqlite3_result_text64(sqlite3_context*, const char*,sqlite3_uint64, - void(*)(void*), unsigned char encoding); SQLITE_API void sqlite3_result_text16(sqlite3_context*, const void*, int, void(*)(void*)); SQLITE_API void sqlite3_result_text16le(sqlite3_context*, const void*, int,void(*)(void*)); SQLITE_API void sqlite3_result_text16be(sqlite3_context*, const void*, int,void(*)(void*)); SQLITE_API void sqlite3_result_value(sqlite3_context*, sqlite3_value*); SQLITE_API void sqlite3_result_zeroblob(sqlite3_context*, int n); @@ -6206,17 +6158,16 @@ #define SQLITE_TESTCTRL_RESERVE 14 #define SQLITE_TESTCTRL_OPTIMIZATIONS 15 #define SQLITE_TESTCTRL_ISKEYWORD 16 #define SQLITE_TESTCTRL_SCRATCHMALLOC 17 #define SQLITE_TESTCTRL_LOCALTIME_FAULT 18 -#define SQLITE_TESTCTRL_EXPLAIN_STMT 19 /* NOT USED */ +#define SQLITE_TESTCTRL_EXPLAIN_STMT 19 #define SQLITE_TESTCTRL_NEVER_CORRUPT 20 #define SQLITE_TESTCTRL_VDBE_COVERAGE 21 #define SQLITE_TESTCTRL_BYTEORDER 22 #define SQLITE_TESTCTRL_ISINIT 23 -#define SQLITE_TESTCTRL_SORTER_MMAP 24 -#define SQLITE_TESTCTRL_LAST 24 +#define SQLITE_TESTCTRL_LAST 23 /* ** CAPI3REF: SQLite Runtime Status ** ** ^This interface is used to retrieve runtime status information @@ -6403,25 +6354,25 @@ ** memory already being in use. ** Only the high-water value is meaningful; ** the current value is always zero.)^ ** ** [[SQLITE_DBSTATUS_CACHE_USED]] ^(- SQLITE_DBSTATUS_CACHE_USED
-**- This parameter returns the approximate number of bytes of heap +**
- This parameter returns the approximate number of of bytes of heap ** memory used by all pager caches associated with the database connection.)^ ** ^The highwater mark associated with SQLITE_DBSTATUS_CACHE_USED is always 0. ** ** [[SQLITE_DBSTATUS_SCHEMA_USED]] ^(
- SQLITE_DBSTATUS_SCHEMA_USED
-**- This parameter returns the approximate number of bytes of heap +**
- This parameter returns the approximate number of of bytes of heap ** memory used to store the schema for all databases associated ** with the connection - main, temp, and any [ATTACH]-ed databases.)^ ** ^The full amount of memory used by the schemas is reported, even if the ** schema memory is shared with other database connections due to ** [shared cache mode] being enabled. ** ^The highwater mark associated with SQLITE_DBSTATUS_SCHEMA_USED is always 0. ** ** [[SQLITE_DBSTATUS_STMT_USED]] ^(
- SQLITE_DBSTATUS_STMT_USED
-**- This parameter returns the approximate number of bytes of heap +**
- This parameter returns the approximate number of of bytes of heap ** and lookaside memory used by all prepared statements associated with ** the database connection.)^ ** ^The highwater mark associated with SQLITE_DBSTATUS_STMT_USED is always 0. **
** Index: SQLite.Interop/src/core/sqlite3ext.h ================================================================== --- SQLite.Interop/src/core/sqlite3ext.h +++ SQLite.Interop/src/core/sqlite3ext.h @@ -26,11 +26,11 @@ ** routines. ** ** WARNING: In order to maintain backwards compatibility, add new ** interfaces to the end of this structure only. If you insert new ** interfaces in the middle of this structure, then older different -** versions of SQLite will not be able to load each other's shared +** versions of SQLite will not be able to load each others' shared ** libraries! */ struct sqlite3_api_routines { void * (*aggregate_context)(sqlite3_context*,int nBytes); int (*aggregate_count)(sqlite3_context*); @@ -248,32 +248,15 @@ int (*uri_boolean)(const char*,const char*,int); sqlite3_int64 (*uri_int64)(const char*,const char*,sqlite3_int64); const char *(*uri_parameter)(const char*,const char*); char *(*vsnprintf)(int,char*,const char*,va_list); int (*wal_checkpoint_v2)(sqlite3*,const char*,int,int*,int*); - /* Version 3.8.7 and later */ - int (*auto_extension)(void(*)(void)); - int (*bind_blob64)(sqlite3_stmt*,int,const void*,sqlite3_uint64, - void(*)(void*)); - int (*bind_text64)(sqlite3_stmt*,int,const char*,sqlite3_uint64, - void(*)(void*),unsigned char); - int (*cancel_auto_extension)(void(*)(void)); - int (*load_extension)(sqlite3*,const char*,const char*,char**); - void *(*malloc64)(sqlite3_uint64); - sqlite3_uint64 (*msize)(void*); - void *(*realloc64)(void*,sqlite3_uint64); - void (*reset_auto_extension)(void); - void (*result_blob64)(sqlite3_context*,const void*,sqlite3_uint64, - void(*)(void*)); - void (*result_text64)(sqlite3_context*,const char*,sqlite3_uint64, - void(*)(void*), unsigned char); - int (*strglob)(const char*,const char*); }; /* ** The following macros redefine the API routines so that they are -** redirected through the global sqlite3_api structure. +** redirected throught the global sqlite3_api structure. ** ** This header file is also used by the loadext.c source file ** (part of the main SQLite library - not an extension) so that ** it can get access to the sqlite3_api_routines structure ** definition. But the main library does not want to redefine @@ -482,23 +465,10 @@ #define sqlite3_uri_boolean sqlite3_api->uri_boolean #define sqlite3_uri_int64 sqlite3_api->uri_int64 #define sqlite3_uri_parameter sqlite3_api->uri_parameter #define sqlite3_uri_vsnprintf sqlite3_api->vsnprintf #define sqlite3_wal_checkpoint_v2 sqlite3_api->wal_checkpoint_v2 -/* Version 3.8.7 and later */ -#define sqlite3_auto_extension sqlite3_api->auto_extension -#define sqlite3_bind_blob64 sqlite3_api->bind_blob64 -#define sqlite3_bind_text64 sqlite3_api->bind_text64 -#define sqlite3_cancel_auto_extension sqlite3_api->cancel_auto_extension -#define sqlite3_load_extension sqlite3_api->load_extension -#define sqlite3_malloc64 sqlite3_api->malloc64 -#define sqlite3_msize sqlite3_api->msize -#define sqlite3_realloc64 sqlite3_api->realloc64 -#define sqlite3_reset_auto_extension sqlite3_api->reset_auto_extension -#define sqlite3_result_blob64 sqlite3_api->result_blob64 -#define sqlite3_result_text64 sqlite3_api->result_text64 -#define sqlite3_strglob sqlite3_api->strglob #endif /* SQLITE_CORE */ #ifndef SQLITE_CORE /* This case when the file really is being compiled as a loadable ** extension */ Index: SQLite.Interop/src/win/crypt.c ================================================================== --- SQLite.Interop/src/win/crypt.c +++ SQLite.Interop/src/win/crypt.c @@ -288,15 +288,11 @@ else /* User-supplied passphrase, so create a cryptographic key out of it */ { hKey = DeriveKey(pKey, nKeyLen); if (hKey == MAXDWORD) { -#if SQLITE_VERSION_NUMBER >= 3008007 - sqlite3ErrorWithMsg(db, rc, SQLITECRYPTERROR_PROVIDER); -#else sqlite3Error(db, rc, SQLITECRYPTERROR_PROVIDER); -#endif return rc; } } /* Create a new encryption block and assign the codec to the new attached database */ @@ -346,15 +342,11 @@ HCRYPTKEY hKey = DeriveKey(pKey, nKey); int rc = SQLITE_ERROR; if (hKey == MAXDWORD) { -#if SQLITE_VERSION_NUMBER >= 3008007 - sqlite3ErrorWithMsg(db, rc, SQLITECRYPTERROR_PROVIDER); -#else sqlite3Error(db, rc, SQLITECRYPTERROR_PROVIDER); -#endif return rc; } if (!pBlock && !hKey) return SQLITE_OK; /* Wasn't encrypted to begin with */ Index: SQLite.Interop/src/win/interop.c ================================================================== --- SQLite.Interop/src/win/interop.c +++ SQLite.Interop/src/win/interop.c @@ -742,30 +742,21 @@ sqlite3_result_int64(pctx, *val); } SQLITE_API int WINAPI sqlite3_context_collcompare_interop(sqlite3_context *ctx, const void *p1, int p1len, const void *p2, int p2len) { -#if SQLITE_VERSION_NUMBER >= 3008007 - CollSeq *pColl = sqlite3GetFuncCollSeq(ctx); -#else - CollSeq *pColl = ctx->pColl; -#endif #if SQLITE_VERSION_NUMBER >= 3008001 if ((ctx->pFunc->funcFlags & SQLITE_FUNC_NEEDCOLL) == 0) return 2; #else if ((ctx->pFunc->flags & SQLITE_FUNC_NEEDCOLL) == 0) return 2; #endif - return pColl->xCmp(pColl->pUser, p1len, p1, p2len, p2); + return ctx->pColl->xCmp(ctx->pColl->pUser, p1len, p1, p2len, p2); } SQLITE_API const char * WINAPI sqlite3_context_collseq_interop(sqlite3_context *ctx, int *ptype, int *enc, int *plen) { -#if SQLITE_VERSION_NUMBER >= 3008007 - CollSeq *pColl = sqlite3GetFuncCollSeq(ctx); -#else CollSeq *pColl = ctx->pColl; -#endif *ptype = 0; *plen = 0; *enc = 0; #if SQLITE_VERSION_NUMBER >= 3008001 Index: SQLite.Interop/src/win/interop.h ================================================================== --- SQLite.Interop/src/win/interop.h +++ SQLite.Interop/src/win/interop.h @@ -4,11 +4,11 @@ * Written by Joe Mistachkin. * Released to the public domain, use at your own risk! */ #ifndef INTEROP_VERSION -#define INTEROP_VERSION "1.0.95.0" +#define INTEROP_VERSION "1.0.94.0" #endif #ifndef INTEROP_SOURCE_ID #define INTEROP_SOURCE_ID "0000000000000000000000000000000000000000" #endif Index: Setup/data/SQLite.iss ================================================================== --- Setup/data/SQLite.iss +++ Setup/data/SQLite.iss @@ -161,11 +161,11 @@ Components: {#InstallerCondition}; Tasks: gac\vs2005; Filename: {app}\bin\Installer.exe; Parameters: "-install false -wow64 true -installFlags AllExceptGlobalAssemblyCache -tracePriority Lowest -verbose true -noCompact true -noNetFx35 true -noNetFx40 true -noNetFx45 true -noNetFx451 true -noVs2008 true -noVs2010 true -noVs2012 true -noVs2013 true -whatIf false -confirm true"; Flags: skipifdoesntexist; Check: CheckIsNetFx2Setup() #endif #endif #if Year != "2005" && Year != "2008" -Components: Application\EF6; Tasks: ngen; Filename: {code:GetNetFx4InstallRoot|Ngen.exe}; Parameters: "uninstall ""{app}\bin\System.Data.SQLite.EF6.dll"" /nologo"; Flags: skipifdoesntexist; Check: CheckIsNetFx4Setup() +Components: Application\LINQ; Tasks: ngen; Filename: {code:GetNetFx4InstallRoot|Ngen.exe}; Parameters: "uninstall ""{app}\bin\System.Data.SQLite.EF6.dll"" /nologo"; Flags: skipifdoesntexist; Check: CheckIsNetFx4Setup() #endif #if Year != "2005" Components: Application\LINQ; Tasks: ngen; Filename: {code:GetNetFx4InstallRoot|Ngen.exe}; Parameters: "uninstall ""{app}\bin\System.Data.SQLite.Linq.dll"" /nologo"; Flags: skipifdoesntexist; Check: CheckIsNetFx4Setup() Components: Application\LINQ; Tasks: ngen; Filename: {code:GetNetFx2InstallRoot|Ngen.exe}; Parameters: "uninstall ""{app}\bin\System.Data.SQLite.Linq.dll"" /nologo"; Flags: skipifdoesntexist; Check: CheckIsNetFx2Setup() and CheckForNetFx35(1) Index: Setup/data/verify.lst ================================================================== --- Setup/data/verify.lst +++ Setup/data/verify.lst @@ -649,11 +649,10 @@ Tests/data/Uninstaller_Test_Vs2012.log Tests/data/Uninstaller_Test_Vs2013.log Tests/data/wal.db Tests/empty.eagle Tests/installer.eagle - Tests/linq.eagle Tests/pkgIndex.eagle Tests/speed.eagle Tests/stress.eagle Tests/thread.eagle Tests/tkt-00f86f9739.eagle Index: System.Data.SQLite.Linq/AssemblyInfo.cs ================================================================== --- System.Data.SQLite.Linq/AssemblyInfo.cs +++ System.Data.SQLite.Linq/AssemblyInfo.cs @@ -51,7 +51,7 @@ // Revision // // You can specify all the values or you can default the Build and Revision Numbers // by using the '*' as shown below: // [assembly: AssemblyVersion("1.0.*")] -[assembly: AssemblyVersion("1.0.95.0")] -[assembly: AssemblyFileVersion("1.0.95.0")] +[assembly: AssemblyVersion("1.0.94.0")] +[assembly: AssemblyFileVersion("1.0.94.0")] Index: System.Data.SQLite.Linq/SQL Generation/DmlSqlGenerator.cs ================================================================== --- System.Data.SQLite.Linq/SQL Generation/DmlSqlGenerator.cs +++ System.Data.SQLite.Linq/SQL Generation/DmlSqlGenerator.cs @@ -78,11 +78,11 @@ commandText.AppendLine(); // where c1 = ..., c2 = ... commandText.Append("WHERE "); tree.Predicate.Accept(translator); - commandText.AppendLine(";"); + commandText.AppendLine(); // generate returning sql GenerateReturningSql(commandText, tree, translator, tree.Returning); parameters = translator.Parameters; @@ -225,11 +225,10 @@ // SQL) DbParameter value; if (translator.MemberValues.TryGetValue(keyMember, out value)) { commandText.Append(value.ParameterName); - commandText.AppendLine(";"); } else { // if no value is registered for the key member, it means it is an identity // which can be retrieved using the scope_identity() function Index: System.Data.SQLite/AssemblyInfo.cs ================================================================== --- System.Data.SQLite/AssemblyInfo.cs +++ System.Data.SQLite/AssemblyInfo.cs @@ -70,9 +70,9 @@ // Build Number // Revision // // You can specify all the values or you can default the Revision and Build Numbers // by using the '*' as shown below: -[assembly: AssemblyVersion("1.0.95.0")] +[assembly: AssemblyVersion("1.0.94.0")] #if !PLATFORM_COMPACTFRAMEWORK -[assembly: AssemblyFileVersion("1.0.95.0")] +[assembly: AssemblyFileVersion("1.0.94.0")] #endif Index: System.Data.SQLite/SQLite3.cs ================================================================== --- System.Data.SQLite/SQLite3.cs +++ System.Data.SQLite/SQLite3.cs @@ -55,11 +55,11 @@ "b621ddff5d844727418956997f475eb829429e411aff3e93f97b70de698b972640925bdd44280df0" + "a25a843266973704137cbb0e7441c1fe7cae4e2440ae91ab8cde3933febcb1ac48dd33b40e13c421" + "d8215c18a4349a436dd499e3c385cc683015f886f6c10bd90115eb2bd61b67750839e3a19941dc9c"; #if !PLATFORM_COMPACTFRAMEWORK - internal const string DesignerVersion = "1.0.95.0"; + internal const string DesignerVersion = "1.0.94.0"; #endif ////// The opaque pointer returned to us by the sqlite provider /// Index: System.Data.SQLite/SQLiteDataReader.cs ================================================================== --- System.Data.SQLite/SQLiteDataReader.cs +++ System.Data.SQLite/SQLiteDataReader.cs @@ -385,11 +385,10 @@ case TypeAffinity.Int64: if (typ == DbType.Int16) return affinity; if (typ == DbType.Int32) return affinity; if (typ == DbType.Int64) return affinity; if (typ == DbType.Boolean) return affinity; - if (typ == DbType.SByte) return affinity; if (typ == DbType.Byte) return affinity; if (typ == DbType.DateTime) return affinity; if (typ == DbType.Single) return affinity; if (typ == DbType.Double) return affinity; if (typ == DbType.Decimal) return affinity; @@ -399,11 +398,13 @@ if (typ == DbType.Double) return affinity; if (typ == DbType.Decimal) return affinity; if (typ == DbType.DateTime) return affinity; break; case TypeAffinity.Text: + if (typ == DbType.SByte) return affinity; if (typ == DbType.String) return affinity; + if (typ == DbType.SByte) return affinity; if (typ == DbType.Guid) return affinity; if (typ == DbType.DateTime) return affinity; if (typ == DbType.Decimal) return affinity; break; case TypeAffinity.Blob: Index: System.Data.SQLite/UnsafeNativeMethods.cs ================================================================== --- System.Data.SQLite/UnsafeNativeMethods.cs +++ System.Data.SQLite/UnsafeNativeMethods.cs @@ -1243,11 +1243,11 @@ // be used because it provides several workarounds to .NET Compact // Framework limitations important for proper operation of the core // System.Data.SQLite functionality (e.g. being able to bind // parameters and handle column values of types Int64 and Double). // - internal const string SQLITE_DLL = "SQLite.Interop.095.dll"; + internal const string SQLITE_DLL = "SQLite.Interop.094.dll"; #elif SQLITE_STANDARD // // NOTE: Otherwise, if the standard SQLite library is enabled, use it. // internal const string SQLITE_DLL = "sqlite3"; Index: Tests/basic.eagle ================================================================== --- Tests/basic.eagle +++ Tests/basic.eagle @@ -15,11 +15,78 @@ ############################################################################### package require System.Data.SQLite.Test runSQLiteTestPrologue -runSQLiteTestFilesPrologue + +############################################################################### + +# +# NOTE: Setup the variables that refer to the various non-data files required +# by the tests in this file. +# +set entityFrameworkDllFile [getBuildFileName EntityFramework.dll] +set systemDataSQLiteDllFile [getBuildFileName System.Data.SQLite.dll] +set systemDataSQLiteLinqDllFile [getBuildFileName System.Data.SQLite.Linq.dll] +set systemDataSQLiteEf6DllFile [getBuildFileName System.Data.SQLite.EF6.dll] +set testExeFile [getBuildFileName test.exe] +set testLinqExeFile [getBuildFileName testlinq.exe] +set testEf6ExeFile [getBuildFileName testef6.exe] + +# +# NOTE: Setup the variables that refer to the various data files required by +# the tests in this file. +# +set testLinqOutFile [file nativename [file join \ + [getSQLiteTestDataPath] testlinq.out]] + +set northwindEfDbFile [file nativename [file join \ + [file dirname [file dirname [getSQLiteTestDataPath]]] testlinq \ + northwindEF.db]] + +# +# NOTE: Setup the test constraints specific to the tests in this file. +# +if {![haveConstraint [appendArgs file_ \ + [file tail $entityFrameworkDllFile]]]} then { + checkForFile $test_channel $entityFrameworkDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteLinqDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteLinqDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteEf6DllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteEf6DllFile +} + +if {![haveConstraint [appendArgs file_ [file tail $testExeFile]]]} then { + checkForFile $test_channel $testExeFile test.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $testLinqExeFile]]]} then { + checkForFile $test_channel $testLinqExeFile testlinq.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $testEf6ExeFile]]]} then { + checkForFile $test_channel $testEf6ExeFile testef6.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $northwindEfDbFile]]]} then { + checkForFile $test_channel $northwindEfDbFile +} + +if {![haveConstraint [appendArgs file_ [file tail $testLinqOutFile]]]} then { + checkForFile $test_channel $testLinqOutFile +} ############################################################################### runTest {test data-1.1 {unit tests from the 'test' project} -setup { cleanupFile [file join [file dirname $testExeFile] Test.db3] @@ -3112,8 +3179,14 @@ reportSQLiteResources $test_channel ############################################################################### -runSQLiteTestFilesEpilogue +unset -nocomplain northwindEfDbFile testLinqOutFile testEf6ExeFile \ + testLinqExeFile testExeFile sqliteDesignerDllFile \ + systemDataSQLiteEf6DllFile systemDataSQLiteLinqDllFile \ + systemDataSQLiteDllFile entityFrameworkDllFile + +############################################################################### + runSQLiteTestEpilogue runTestEpilogue Index: Tests/common.eagle ================================================================== --- Tests/common.eagle +++ Tests/common.eagle @@ -2227,53 +2227,10 @@ # the SQLiteLog class being able to setup its logging callback). # Normally, this should only be performed if SQLite is loaded and # ready for use by the test suite. # if {$force || [isSQLiteReady]} then { - # - # BUGFIX: Make sure that any "leaked" transactions and/or connections - # are cleaned up before calling the native shutdown function. - # - foreach transaction [info transactions] { - if {[string match \ - System#Data#SQLite#SQLiteTransaction#* $transaction]} then { - if {[catch { - sql transaction rollback $transaction - } error] == 0} then { - if {!$quiet} then { - tputs $channel [appendArgs \ - "---- rolled back leaked transaction \"" \ - $transaction \"\n] - } - } else { - if {!$quiet} then { - tputs $channel [appendArgs \ - "==== WARNING: failed to rollback leaked transaction \"" \ - $transaction "\", error: " \n\t $error \n] - } - } - } - } - - foreach db [info connections] { - if {[string match \ - System#Data#SQLite#SQLiteConnection#* $db]} then { - if {[catch {sql close $db} error] == 0} then { - if {!$quiet} then { - tputs $channel [appendArgs \ - "---- closed leaked database \"" $db \"\n] - } - } else { - if {!$quiet} then { - tputs $channel [appendArgs \ - "==== WARNING: failed to close leaked database \"" \ - $db "\", error: " \n\t $error \n] - } - } - } - } - # # BUGFIX: Before calling the native shutdown function, make sure both # of the PRAGMA related directory names are freed. # checkForSQLiteDirectories $channel true @@ -3267,227 +3224,10 @@ "---- System.Data.SQLite tests began at " \ [clock format [clock seconds]] \n] } } - proc runSQLiteTestFilesPrologue {} { - uplevel 1 { - # - # NOTE: Setup the variables that refer to the various non-data files - # required by the tests in this file. - # - set entityFrameworkDllFile \ - [getBuildFileName EntityFramework.dll] - - set installerExeFile [getBuildFileName Installer.exe] - set sqliteDesignerDllFile [getBuildFileName SQLite.Designer.dll] - - set systemDataSQLiteDllFile \ - [getBuildFileName System.Data.SQLite.dll] - - set systemDataSQLiteLinqDllFile \ - [getBuildFileName System.Data.SQLite.Linq.dll] - - set systemDataSQLiteEf6DllFile \ - [getBuildFileName System.Data.SQLite.EF6.dll] - - set testExeFile [getBuildFileName test.exe] - set testLinqExeFile [getBuildFileName testlinq.exe] - set testEf6ExeFile [getBuildFileName testef6.exe] - - # - # NOTE: Setup the variables that refer to the various data files - # required by the tests in this file. - # - set testLinqOutFile [file nativename [file join \ - [getSQLiteTestDataPath] testlinq.out]] - - set northwindEfDbFile [file nativename [file join \ - [file dirname [file dirname [getSQLiteTestDataPath]]] \ - testlinq northwindEF.db]] - - set nonWalDbFile [file nativename [file join \ - [getSQLiteTestDataPath] nonWal.db]] - - set walDbFile [file nativename [file join \ - [getSQLiteTestDataPath] wal.db]] - - # - # NOTE: The various install/uninstall log files used to test the - # design-time component installer. - # - set testInstallVs2005LogFile [file nativename [file join \ - [getSQLiteTestDataPath] Installer_Test_Vs2005.log]] - - set testInstallVs2008LogFile [file nativename [file join \ - [getSQLiteTestDataPath] Installer_Test_Vs2008.log]] - - set testInstallVs2010LogFile [file nativename [file join \ - [getSQLiteTestDataPath] Installer_Test_Vs2010.log]] - - set testInstallVs2012LogFile [file nativename [file join \ - [getSQLiteTestDataPath] Installer_Test_Vs2012.log]] - - set testInstallVs2013LogFile [file nativename [file join \ - [getSQLiteTestDataPath] Installer_Test_Vs2013.log]] - - set testUninstallVs2005LogFile [file nativename [file join \ - [getSQLiteTestDataPath] Uninstaller_Test_Vs2005.log]] - - set testUninstallVs2008LogFile [file nativename [file join \ - [getSQLiteTestDataPath] Uninstaller_Test_Vs2008.log]] - - set testUninstallVs2010LogFile [file nativename [file join \ - [getSQLiteTestDataPath] Uninstaller_Test_Vs2010.log]] - - set testUninstallVs2012LogFile [file nativename [file join \ - [getSQLiteTestDataPath] Uninstaller_Test_Vs2012.log]] - - set testUninstallVs2013LogFile [file nativename [file join \ - [getSQLiteTestDataPath] Uninstaller_Test_Vs2013.log]] - - ####################################################################### - - if {![haveConstraint [appendArgs file_ \ - [file tail $entityFrameworkDllFile]]]} then { - checkForFile $test_channel $entityFrameworkDllFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $installerExeFile]]]} then { - checkForFile $test_channel $installerExeFile Installer.exe - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $sqliteDesignerDllFile]]]} then { - checkForFile $test_channel $sqliteDesignerDllFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $systemDataSQLiteDllFile]]]} then { - checkForFile $test_channel $systemDataSQLiteDllFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $systemDataSQLiteLinqDllFile]]]} then { - checkForFile $test_channel $systemDataSQLiteLinqDllFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $systemDataSQLiteEf6DllFile]]]} then { - checkForFile $test_channel $systemDataSQLiteEf6DllFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testExeFile]]]} then { - checkForFile $test_channel $testExeFile test.exe - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testLinqExeFile]]]} then { - checkForFile $test_channel $testLinqExeFile testlinq.exe - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testEf6ExeFile]]]} then { - checkForFile $test_channel $testEf6ExeFile testef6.exe - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testLinqOutFile]]]} then { - checkForFile $test_channel $testLinqOutFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $northwindEfDbFile]]]} then { - checkForFile $test_channel $northwindEfDbFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $nonWalDbFile]]]} then { - checkForFile $test_channel $nonWalDbFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $walDbFile]]]} then { - checkForFile $test_channel $walDbFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testInstallVs2005LogFile]]]} then { - checkForFile $test_channel $testInstallVs2005LogFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testInstallVs2008LogFile]]]} then { - checkForFile $test_channel $testInstallVs2008LogFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testInstallVs2010LogFile]]]} then { - checkForFile $test_channel $testInstallVs2010LogFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testInstallVs2012LogFile]]]} then { - checkForFile $test_channel $testInstallVs2012LogFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testInstallVs2013LogFile]]]} then { - checkForFile $test_channel $testInstallVs2013LogFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testUninstallVs2005LogFile]]]} then { - checkForFile $test_channel $testUninstallVs2005LogFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testUninstallVs2008LogFile]]]} then { - checkForFile $test_channel $testUninstallVs2008LogFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testUninstallVs2010LogFile]]]} then { - checkForFile $test_channel $testUninstallVs2010LogFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testUninstallVs2012LogFile]]]} then { - checkForFile $test_channel $testUninstallVs2012LogFile - } - - if {![haveConstraint [appendArgs file_ \ - [file tail $testUninstallVs2013LogFile]]]} then { - checkForFile $test_channel $testUninstallVs2013LogFile - } - } - } - - proc runSQLiteTestFilesEpilogue {} { - uplevel 1 { - unset -nocomplain \ - testUninstallVs2013LogFile testUninstallVs2012LogFile \ - testUninstallVs2010LogFile testUninstallVs2008LogFile \ - testUninstallVs2005LogFile testInstallVs2013LogFile \ - testInstallVs2012LogFile testInstallVs2010LogFile \ - testInstallVs2008LogFile testInstallVs2005LogFile - - unset -nocomplain \ - northwindEfDbFile testLinqOutFile \ - walDbFile nonWalDbFile - - unset -nocomplain \ - testEf6ExeFile testLinqExeFile \ - testExeFile systemDataSQLiteEf6DllFile \ - systemDataSQLiteLinqDllFile systemDataSQLiteDllFile \ - sqliteDesignerDllFile installerExeFile \ - entityFrameworkDllFile - } - } - proc runSQLiteTestEpilogue {} { # # NOTE: Skip running our custom epilogue if the main one has been # skipped. # Index: Tests/installer.eagle ================================================================== --- Tests/installer.eagle +++ Tests/installer.eagle @@ -15,11 +15,134 @@ ############################################################################### package require System.Data.SQLite.Test runSQLiteTestPrologue -runSQLiteTestFilesPrologue + +############################################################################### + +# +# NOTE: Setup the variables that refer to the various non-data files required +# by the tests in this file. +# +set systemDataSQLiteDllFile [getBuildFileName System.Data.SQLite.dll] +set systemDataSQLiteLinqDllFile [getBuildFileName System.Data.SQLite.Linq.dll] +set systemDataSQLiteEf6DllFile [getBuildFileName System.Data.SQLite.EF6.dll] +set sqliteDesignerDllFile [getBuildFileName SQLite.Designer.dll] +set installerExeFile [getBuildFileName Installer.exe] + +# +# NOTE: The various install/uninstall log files used to test the design-time +# component installer. +# +set testInstallVs2005LogFile [file nativename [file join \ + [getSQLiteTestDataPath] Installer_Test_Vs2005.log]] + +set testInstallVs2008LogFile [file nativename [file join \ + [getSQLiteTestDataPath] Installer_Test_Vs2008.log]] + +set testInstallVs2010LogFile [file nativename [file join \ + [getSQLiteTestDataPath] Installer_Test_Vs2010.log]] + +set testInstallVs2012LogFile [file nativename [file join \ + [getSQLiteTestDataPath] Installer_Test_Vs2012.log]] + +set testInstallVs2013LogFile [file nativename [file join \ + [getSQLiteTestDataPath] Installer_Test_Vs2013.log]] + +set testUninstallVs2005LogFile [file nativename [file join \ + [getSQLiteTestDataPath] Uninstaller_Test_Vs2005.log]] + +set testUninstallVs2008LogFile [file nativename [file join \ + [getSQLiteTestDataPath] Uninstaller_Test_Vs2008.log]] + +set testUninstallVs2010LogFile [file nativename [file join \ + [getSQLiteTestDataPath] Uninstaller_Test_Vs2010.log]] + +set testUninstallVs2012LogFile [file nativename [file join \ + [getSQLiteTestDataPath] Uninstaller_Test_Vs2012.log]] + +set testUninstallVs2013LogFile [file nativename [file join \ + [getSQLiteTestDataPath] Uninstaller_Test_Vs2013.log]] + +# +# NOTE: Setup the test constraints specific to the tests in this file. +# +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteLinqDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteLinqDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteEf6DllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteEf6DllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $sqliteDesignerDllFile]]]} then { + checkForFile $test_channel $sqliteDesignerDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $installerExeFile]]]} then { + checkForFile $test_channel $installerExeFile Installer.exe +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $testInstallVs2005LogFile]]]} then { + checkForFile $test_channel $testInstallVs2005LogFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $testInstallVs2008LogFile]]]} then { + checkForFile $test_channel $testInstallVs2008LogFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $testInstallVs2010LogFile]]]} then { + checkForFile $test_channel $testInstallVs2010LogFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $testInstallVs2012LogFile]]]} then { + checkForFile $test_channel $testInstallVs2012LogFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $testInstallVs2013LogFile]]]} then { + checkForFile $test_channel $testInstallVs2013LogFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $testUninstallVs2005LogFile]]]} then { + checkForFile $test_channel $testUninstallVs2005LogFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $testUninstallVs2008LogFile]]]} then { + checkForFile $test_channel $testUninstallVs2008LogFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $testUninstallVs2010LogFile]]]} then { + checkForFile $test_channel $testUninstallVs2010LogFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $testUninstallVs2012LogFile]]]} then { + checkForFile $test_channel $testUninstallVs2012LogFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $testUninstallVs2013LogFile]]]} then { + checkForFile $test_channel $testUninstallVs2013LogFile +} ############################################################################### runTest {test installer-1.1 {installer tool / Visual Studio 2005} -setup { set fileName [file join [getTemporaryPath] [file tail [string map [list \ @@ -419,8 +542,17 @@ file_System.Data.SQLite.EF6.dll file_SQLite.Designer.dll\ file_Uninstaller_Test_Vs2013.log}] -result {0 True}} ############################################################################### -runSQLiteTestFilesEpilogue +unset -nocomplain testUninstallVs2013LogFile testUninstallVs2012LogFile \ + testUninstallVs2010LogFile testUninstallVs2008LogFile \ + testUninstallVs2005LogFile testInstallVs2013LogFile \ + testInstallVs2012LogFile testInstallVs2010LogFile \ + testInstallVs2008LogFile testInstallVs2005LogFile sqliteDesignerDllFile \ + installerExeFile systemDataSQLiteEf6DllFile systemDataSQLiteLinqDllFile \ + systemDataSQLiteDllFile + +############################################################################### + runSQLiteTestEpilogue runTestEpilogue DELETED Tests/linq.eagle Index: Tests/linq.eagle ================================================================== --- Tests/linq.eagle +++ /dev/null @@ -1,64 +0,0 @@ -############################################################################### -# -# linq.eagle -- -# -# Written by Joe Mistachkin. -# Released to the public domain, use at your own risk! -# -############################################################################### - -package require Eagle -package require Eagle.Library -package require Eagle.Test - -runTestPrologue - -############################################################################### - -package require System.Data.SQLite.Test -runSQLiteTestPrologue -runSQLiteTestFilesPrologue - -############################################################################### - -runTest {test linq-1.1 {DbUpdateCommandTree with UPDATE and SELECT} -body { - # - # NOTE: Re-copy the reference database file used for this unit test to the - # build directory in case it has been changed by a previous test run. - # - file copy -force $northwindEfDbFile \ - [file join [getBuildDirectory] [file tail $northwindEfDbFile]] - - set result [list] - set output "" - - set code [catch { - testClrExec $testLinqExeFile [list -eventflags Wait -directory \ - [file dirname $testLinqExeFile] -nocarriagereturns -stdout output \ - -success 0] -update - } error] - - tlog "---- BEGIN STDOUT OUTPUT\n" - tlog $output - tlog "\n---- END STDOUT OUTPUT\n" - - lappend result $code - - if {$code == 0} then { - lappend result [string trim $output] - } else { - lappend result [string trim $error] - } - - set result -} -cleanup { - unset -nocomplain code output error result -} -constraints {eagle monoToDo SQLite file_System.Data.SQLite.dll testExec\ -file_System.Data.SQLite.Linq.dll file_testlinq.exe file_northwindEF.db} \ --result {0 {inserted 1 updated 1}}} - -############################################################################### - -runSQLiteTestFilesEpilogue -runSQLiteTestEpilogue -runTestEpilogue Index: Tests/tkt-00f86f9739.eagle ================================================================== --- Tests/tkt-00f86f9739.eagle +++ Tests/tkt-00f86f9739.eagle @@ -15,11 +15,66 @@ ############################################################################### package require System.Data.SQLite.Test runSQLiteTestPrologue -runSQLiteTestFilesPrologue + +############################################################################### + +# +# NOTE: Setup the variables that refer to the various non-data files required +# by the tests in this file. +# +set entityFrameworkDllFile [getBuildFileName EntityFramework.dll] +set systemDataSQLiteDllFile [getBuildFileName System.Data.SQLite.dll] +set systemDataSQLiteLinqDllFile [getBuildFileName System.Data.SQLite.Linq.dll] +set systemDataSQLiteEf6DllFile [getBuildFileName System.Data.SQLite.EF6.dll] +set testLinqExeFile [getBuildFileName testlinq.exe] +set testEf6ExeFile [getBuildFileName testef6.exe] + +# +# NOTE: Setup the variables that refer to the various data files required by +# the tests in this file. +# +set northwindEfDbFile [file nativename [file join \ + [file dirname [file dirname [getSQLiteTestDataPath]]] testlinq \ + northwindEF.db]] + +# +# NOTE: Setup the test constraints specific to the tests in this file. +# +if {![haveConstraint [appendArgs file_ \ + [file tail $entityFrameworkDllFile]]]} then { + checkForFile $test_channel $entityFrameworkDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteLinqDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteLinqDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteEf6DllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteEf6DllFile +} + +if {![haveConstraint [appendArgs file_ [file tail $testLinqExeFile]]]} then { + checkForFile $test_channel $testLinqExeFile testlinq.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $testEf6ExeFile]]]} then { + checkForFile $test_channel $testEf6ExeFile testef6.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $northwindEfDbFile]]]} then { + checkForFile $test_channel $northwindEfDbFile +} ############################################################################### runTest {test tkt-00f86f9739-1.1 {LINQ with StartsWith} -body { # @@ -109,8 +164,13 @@ {0 {} 0 {DRACD OLDWO RATTC} 0 {ALFKI CACTU CHOPS FOLKO GALED KOENE LILAS MAGAA\ MAISD OCEAN RANCH SAVEA THECR} 0 {} 0 {} 0 {} 0 {}}} ############################################################################### -runSQLiteTestFilesEpilogue +unset -nocomplain northwindEfDbFile testEf6ExeFile testLinqExeFile \ + systemDataSQLiteEf6DllFile systemDataSQLiteLinqDllFile \ + systemDataSQLiteDllFile entityFrameworkDllFile + +############################################################################### + runSQLiteTestEpilogue runTestEpilogue Index: Tests/tkt-448d663d11.eagle ================================================================== --- Tests/tkt-448d663d11.eagle +++ Tests/tkt-448d663d11.eagle @@ -15,11 +15,30 @@ ############################################################################### package require System.Data.SQLite.Test runSQLiteTestPrologue -runSQLiteTestFilesPrologue + +############################################################################### + +# +# NOTE: Setup the variables that refer to the various data files required by +# the tests in this file. +# +set walDbFile [file nativename [file join [getSQLiteTestDataPath] wal.db]] +set nonWalDbFile [file nativename [file join [getSQLiteTestDataPath] nonWal.db]] + +# +# NOTE: Setup the test constraints specific to the tests in this file. +# +if {![haveConstraint [appendArgs file_ [file tail $walDbFile]]]} then { + checkForFile $test_channel $walDbFile +} + +if {![haveConstraint [appendArgs file_ [file tail $nonWalDbFile]]]} then { + checkForFile $test_channel $nonWalDbFile +} ############################################################################### runTest {test tkt-448d663d11-1.1 {missing journal mode, new db} -body { setupDb [set fileName tkt-448d663d11-1.1.db] @@ -276,8 +295,11 @@ } -constraints {eagle monoBug28 command.sql compile.DATA SQLite\ System.Data.SQLite file_wal.db} -result {wal}} ############################################################################### -runSQLiteTestFilesEpilogue +unset -nocomplain nonWalDbFile walDbFile + +############################################################################### + runSQLiteTestEpilogue runTestEpilogue Index: Tests/tkt-59edc1018b.eagle ================================================================== --- Tests/tkt-59edc1018b.eagle +++ Tests/tkt-59edc1018b.eagle @@ -15,11 +15,60 @@ ############################################################################### package require System.Data.SQLite.Test runSQLiteTestPrologue -runSQLiteTestFilesPrologue + +############################################################################### + +# +# NOTE: Setup the variables that refer to the various non-data files required +# by the tests in this file. +# +set systemDataSQLiteDllFile [getBuildFileName System.Data.SQLite.dll] +set systemDataSQLiteLinqDllFile [getBuildFileName System.Data.SQLite.Linq.dll] +set systemDataSQLiteEf6DllFile [getBuildFileName System.Data.SQLite.EF6.dll] +set testLinqExeFile [getBuildFileName testlinq.exe] +set testEf6ExeFile [getBuildFileName testef6.exe] + +# +# NOTE: Setup the variables that refer to the various data files required by +# the tests in this file. +# +set northwindEfDbFile [file nativename [file join \ + [file dirname [file dirname [getSQLiteTestDataPath]]] testlinq \ + northwindEF.db]] + +# +# NOTE: Setup the test constraints specific to the tests in this file. +# +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteLinqDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteLinqDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteEf6DllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteEf6DllFile +} + +if {![haveConstraint [appendArgs file_ [file tail $testLinqExeFile]]]} then { + checkForFile $test_channel $testLinqExeFile testlinq.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $testEf6ExeFile]]]} then { + checkForFile $test_channel $testEf6ExeFile testef6.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $northwindEfDbFile]]]} then { + checkForFile $test_channel $northwindEfDbFile +} ############################################################################### runTest {test tkt-59edc1018b-1.1 {LINQ with EndsWith} -body { # @@ -107,8 +156,13 @@ file_northwindEF.db} -result {0 {} 0 {FURIB GALED GODOS LAZYK LINOD PRINI REGGC\ WOLZA} 0 {} 0 ERNSH 0 {} 0 {AROUT BSBEV CONSH EASTC NORTS SEVES} 0 {}}} ############################################################################### -runSQLiteTestFilesEpilogue +unset -nocomplain northwindEfDbFile testEf6ExeFile testLinqExeFile \ + systemDataSQLiteEf6DllFile systemDataSQLiteLinqDllFile \ + systemDataSQLiteDllFile + +############################################################################### + runSQLiteTestEpilogue runTestEpilogue Index: Tests/tkt-8b7d179c3c.eagle ================================================================== --- Tests/tkt-8b7d179c3c.eagle +++ Tests/tkt-8b7d179c3c.eagle @@ -15,11 +15,60 @@ ############################################################################### package require System.Data.SQLite.Test runSQLiteTestPrologue -runSQLiteTestFilesPrologue + +############################################################################### + +# +# NOTE: Setup the variables that refer to the various non-data files required +# by the tests in this file. +# +set systemDataSQLiteDllFile [getBuildFileName System.Data.SQLite.dll] +set systemDataSQLiteLinqDllFile [getBuildFileName System.Data.SQLite.Linq.dll] +set systemDataSQLiteEf6DllFile [getBuildFileName System.Data.SQLite.EF6.dll] +set testLinqExeFile [getBuildFileName testlinq.exe] +set testEf6ExeFile [getBuildFileName testef6.exe] + +# +# NOTE: Setup the variables that refer to the various data files required by +# the tests in this file. +# +set northwindEfDbFile [file nativename [file join \ + [file dirname [file dirname [getSQLiteTestDataPath]]] testlinq \ + northwindEF.db]] + +# +# NOTE: Setup the test constraints specific to the tests in this file. +# +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteLinqDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteLinqDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteEf6DllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteEf6DllFile +} + +if {![haveConstraint [appendArgs file_ [file tail $testLinqExeFile]]]} then { + checkForFile $test_channel $testLinqExeFile testlinq.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $testEf6ExeFile]]]} then { + checkForFile $test_channel $testEf6ExeFile testef6.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $northwindEfDbFile]]]} then { + checkForFile $test_channel $northwindEfDbFile +} ############################################################################### runTest {test tkt-8b7d179c3c-1.1 {LINQ with Skip and Take} -body { # @@ -129,8 +178,13 @@ COMMI FAMIA QUEEN TRADH WHITC GODOS SANTG BLONP WANDK FRANS LAMAI BOTTM LAUGB\ LACOR LAZYK WOLZA VAFFE}}} ############################################################################### -runSQLiteTestFilesEpilogue +unset -nocomplain northwindEfDbFile testEf6ExeFile testLinqExeFile \ + systemDataSQLiteEf6DllFile systemDataSQLiteLinqDllFile \ + systemDataSQLiteDllFile + +############################################################################### + runSQLiteTestEpilogue runTestEpilogue Index: Tests/tkt-ccfa69fc32.eagle ================================================================== --- Tests/tkt-ccfa69fc32.eagle +++ Tests/tkt-ccfa69fc32.eagle @@ -15,11 +15,60 @@ ############################################################################### package require System.Data.SQLite.Test runSQLiteTestPrologue -runSQLiteTestFilesPrologue + +############################################################################### + +# +# NOTE: Setup the variables that refer to the various non-data files required +# by the tests in this file. +# +set systemDataSQLiteDllFile [getBuildFileName System.Data.SQLite.dll] +set systemDataSQLiteLinqDllFile [getBuildFileName System.Data.SQLite.Linq.dll] +set systemDataSQLiteEf6DllFile [getBuildFileName System.Data.SQLite.EF6.dll] +set testLinqExeFile [getBuildFileName testlinq.exe] +set testEf6ExeFile [getBuildFileName testef6.exe] + +# +# NOTE: Setup the variables that refer to the various data files required by +# the tests in this file. +# +set northwindEfDbFile [file nativename [file join \ + [file dirname [file dirname [getSQLiteTestDataPath]]] testlinq \ + northwindEF.db]] + +# +# NOTE: Setup the test constraints specific to the tests in this file. +# +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteLinqDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteLinqDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteEf6DllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteEf6DllFile +} + +if {![haveConstraint [appendArgs file_ [file tail $testLinqExeFile]]]} then { + checkForFile $test_channel $testLinqExeFile testlinq.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $testEf6ExeFile]]]} then { + checkForFile $test_channel $testEf6ExeFile testef6.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $northwindEfDbFile]]]} then { + checkForFile $test_channel $northwindEfDbFile +} ############################################################################### runTest {test tkt-ccfa69fc32-1.1 {LINQ Transaction Scope} -body { # @@ -109,8 +158,13 @@ UNIQUE constraint failed: Territories.TerritoryID *} 0 {1 2 3 4 5 6 7 8 9 10 1576 1577 1578 1579 1580 1581 1730 1833 2116 2139}}} ############################################################################### -runSQLiteTestFilesEpilogue +unset -nocomplain northwindEfDbFile testEf6ExeFile testLinqExeFile \ + systemDataSQLiteEf6DllFile systemDataSQLiteLinqDllFile \ + systemDataSQLiteDllFile + +############################################################################### + runSQLiteTestEpilogue runTestEpilogue Index: Tests/tkt-da9f18d039.eagle ================================================================== --- Tests/tkt-da9f18d039.eagle +++ Tests/tkt-da9f18d039.eagle @@ -15,11 +15,60 @@ ############################################################################### package require System.Data.SQLite.Test runSQLiteTestPrologue -runSQLiteTestFilesPrologue + +############################################################################### + +# +# NOTE: Setup the variables that refer to the various non-data files required +# by the tests in this file. +# +set systemDataSQLiteDllFile [getBuildFileName System.Data.SQLite.dll] +set systemDataSQLiteLinqDllFile [getBuildFileName System.Data.SQLite.Linq.dll] +set systemDataSQLiteEf6DllFile [getBuildFileName System.Data.SQLite.EF6.dll] +set testLinqExeFile [getBuildFileName testlinq.exe] +set testEf6ExeFile [getBuildFileName testef6.exe] + +# +# NOTE: Setup the variables that refer to the various data files required by +# the tests in this file. +# +set northwindEfDbFile [file nativename [file join \ + [file dirname [file dirname [getSQLiteTestDataPath]]] testlinq \ + northwindEF.db]] + +# +# NOTE: Setup the test constraints specific to the tests in this file. +# +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteLinqDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteLinqDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteEf6DllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteEf6DllFile +} + +if {![haveConstraint [appendArgs file_ [file tail $testLinqExeFile]]]} then { + checkForFile $test_channel $testLinqExeFile testlinq.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $testEf6ExeFile]]]} then { + checkForFile $test_channel $testEf6ExeFile testef6.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $northwindEfDbFile]]]} then { + checkForFile $test_channel $northwindEfDbFile +} ############################################################################### runTest {test tkt-da9f18d039-1.1 {LINQ DateTime handling} -body { # @@ -95,8 +144,13 @@ file_System.Data.SQLite.EF6.dll file_testef6.exe file_northwindEF.db} \ -result {0 {}}} ############################################################################### -runSQLiteTestFilesEpilogue +unset -nocomplain northwindEfDbFile testEf6ExeFile testLinqExeFile \ + systemDataSQLiteEf6DllFile systemDataSQLiteLinqDllFile \ + systemDataSQLiteDllFile + +############################################################################### + runSQLiteTestEpilogue runTestEpilogue Index: Tests/version.eagle ================================================================== --- Tests/version.eagle +++ Tests/version.eagle @@ -15,11 +15,10 @@ ############################################################################### package require System.Data.SQLite.Test runSQLiteTestPrologue -runSQLiteTestFilesPrologue ############################################################################### # ******************** BEGIN VOLATILE VERSION INFORMATION ********************* ############################################################################### @@ -29,11 +28,11 @@ # the source code files, the built binaries, and the release packages. # The revision number is reserved for NuGet package versioning. # set version(major) 1 set version(minor) 0 -set version(build) 95; # NOTE: Incremented with each release. +set version(build) 94; # NOTE: Incremented with each release. set version(revision) 0 ############################################################################### # ********************* END VOLATILE VERSION INFORMATION ********************** ############################################################################### @@ -52,10 +51,59 @@ # be the same as the above version number, with the exception of the # revision, which may differ. # set version(nuget) [appendArgs $version(major) . $version(minor) . \ $version(build) . \\d+] + +############################################################################### + +# +# NOTE: Setup the variables that refer to the various non-data files required +# by the tests in this file. +# +set systemDataSQLiteDllFile [getBuildFileName System.Data.SQLite.dll] +set systemDataSQLiteLinqDllFile [getBuildFileName System.Data.SQLite.Linq.dll] +set systemDataSQLiteEf6DllFile [getBuildFileName System.Data.SQLite.EF6.dll] +set sqliteDesignerDllFile [getBuildFileName SQLite.Designer.dll] +set testExeFile [getBuildFileName test.exe] +set testLinqExeFile [getBuildFileName testlinq.exe] +set testEf6ExeFile [getBuildFileName testef6.exe] + +# +# NOTE: Setup the test constraints specific to the tests in this file. +# +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteLinqDllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteLinqDllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $systemDataSQLiteEf6DllFile]]]} then { + checkForFile $test_channel $systemDataSQLiteEf6DllFile +} + +if {![haveConstraint [appendArgs file_ \ + [file tail $sqliteDesignerDllFile]]]} then { + checkForFile $test_channel $sqliteDesignerDllFile +} + +if {![haveConstraint [appendArgs file_ [file tail $testExeFile]]]} then { + checkForFile $test_channel $testExeFile test.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $testLinqExeFile]]]} then { + checkForFile $test_channel $testLinqExeFile testlinq.exe +} + +if {![haveConstraint [appendArgs file_ [file tail $testEf6ExeFile]]]} then { + checkForFile $test_channel $testEf6ExeFile testef6.exe +} ############################################################################### runTest {test version-1.1 {'System.Data.SQLite' binary version} -body { file version $systemDataSQLiteDllFile @@ -929,14 +977,14 @@ for {set i 1} {$i <= [llength $fileNames]} {incr i} { set pattern [lindex $patterns [expr {$i - 1}]] set fileName [lindex $fileNames [expr {$i - 1}]] set constraint [string map [list / _ \\ _] $fileName] - set fileName [file nativename [file join $root_path $fileName]] + set fileName [file join $root_path $fileName] if {![haveConstraint [appendArgs file_ $constraint]]} then { - checkForFile $test_channel $fileName $constraint; # EXEMPT: Dynamic. + checkForFile $test_channel $fileName $constraint } runTest {test [appendArgs version-1.21. $i] \ [appendArgs "pattern {" $pattern "} in file \"" $fileName \"] -body { if {[hasRuntimeOption verbose]} then { @@ -952,8 +1000,13 @@ unset -nocomplain constraint fileName pattern fileNames patterns i version ############################################################################### -runSQLiteTestFilesEpilogue +unset -nocomplain testEf6ExeFile testLinqExeFile testExeFile \ + sqliteDesignerDllFile systemDataSQLiteEf6DllFile \ + systemDataSQLiteLinqDllFile systemDataSQLiteDllFile + +############################################################################### + runSQLiteTestEpilogue runTestEpilogue Index: readme.htm ================================================================== --- readme.htm +++ readme.htm @@ -3,12 +3,12 @@ADO.NET SQLite Data Provider
-Version 1.0.95.0 November XX, 2014 (release scheduled)
-Using SQLite 3.8.7
+Version 1.0.94.0 September 9, 2014
+Using SQLite 3.8.6
Originally written by Robert Simpson
Released to the public domain, use at your own risk!
Official provider website: http://system.data.sqlite.org/
Legacy versions: http://sqlite.phxsoftware.com/
@@ -144,11 +144,11 @@ <configuration> <system.data> <DbProviderFactories> <remove invariant="System.Data.SQLite" /> <add name="SQLite Data Provider" invariant="System.Data.SQLite" description=".NET Framework Data Provider for SQLite" - type="System.Data.SQLite.SQLiteFactory, System.Data.SQLite, Version=1.0.95.0, Culture=neutral, PublicKeyToken=db937bc2d44ff139" /> + type="System.Data.SQLite.SQLiteFactory, System.Data.SQLite, Version=1.0.94.0, Culture=neutral, PublicKeyToken=db937bc2d44ff139" /> </DbProviderFactories> </system.data> </configuration>@@ -207,18 +207,10 @@
Version History
- 1.0.95.0 - November XX, 2014 (release scheduled) -
--
-
- Updated to SQLite 3.8.7. -
- Make sure SQL statements generated for DbUpdateCommandTree objects are properly delimited. -
- In the SQLiteDataReader.VerifyType method, remove duplicate "if" statement for the DbType.SByte value and move the remaining "if" to the Int64 affinity. Fix for [c5cc2fb334]. ** Potentially Incompatible Change ** -
1.0.94.0 - September 9, 2014
- Updated to SQLite 3.8.6.
- Updated to Entity Framework 6.1.1. Index: test/AssemblyInfo.cs ================================================================== --- test/AssemblyInfo.cs +++ test/AssemblyInfo.cs @@ -36,7 +36,7 @@ // Major Version // Minor Version // Build Number // Revision // -[assembly: AssemblyVersion("1.0.95.0")] -[assembly: AssemblyFileVersion("1.0.95.0")] +[assembly: AssemblyVersion("1.0.94.0")] +[assembly: AssemblyFileVersion("1.0.94.0")] Index: test/app.config ================================================================== --- test/app.config +++ test/app.config @@ -1,8 +1,8 @@
- Updated to [http://www.sqlite.org/src/info/trunk|SQLite 3.8.7]. -
- Make sure SQL statements generated for DbUpdateCommandTree objects are properly delimited. -
- In the SQLiteDataReader.VerifyType method, remove duplicate "if" statement for the DbType.SByte value and move the remaining "if" to the Int64 affinity. Fix for [c5cc2fb334]. ** Potentially Incompatible Change ** -
- Updated to [http://www.sqlite.org/releaselog/3_8_6.html|SQLite 3.8.6].
- Updated to [http://www.nuget.org/packages/EntityFramework/6.1.1|Entity Framework 6.1.1].
- + - + - + - + - + - + - + - + - + - + - + - - + Index: testlinq/NorthwindModel.EF6.2012.edmx ================================================================== --- testlinq/NorthwindModel.EF6.2012.edmx +++ testlinq/NorthwindModel.EF6.2012.edmx @@ -147,17 +147,11 @@ - - + Index: testlinq/NorthwindModel.EF6.2013.edmx ================================================================== --- testlinq/NorthwindModel.EF6.2013.edmx +++ testlinq/NorthwindModel.EF6.2013.edmx @@ -147,17 +147,11 @@ - - + Index: testlinq/NorthwindModel.Linq.2008.edmx ================================================================== --- testlinq/NorthwindModel.Linq.2008.edmx +++ testlinq/NorthwindModel.Linq.2008.edmx @@ -147,17 +147,11 @@ - - + Index: testlinq/NorthwindModel.Linq.2010.edmx ================================================================== --- testlinq/NorthwindModel.Linq.2010.edmx +++ testlinq/NorthwindModel.Linq.2010.edmx @@ -147,17 +147,11 @@ - - + Index: testlinq/NorthwindModel.Linq.2012.edmx ================================================================== --- testlinq/NorthwindModel.Linq.2012.edmx +++ testlinq/NorthwindModel.Linq.2012.edmx @@ -147,17 +147,11 @@ - - + Index: testlinq/NorthwindModel.Linq.2013.edmx ================================================================== --- testlinq/NorthwindModel.Linq.2013.edmx +++ testlinq/NorthwindModel.Linq.2013.edmx @@ -147,17 +147,11 @@ - - + Index: testlinq/Program.cs ================================================================== --- testlinq/Program.cs +++ testlinq/Program.cs @@ -120,14 +120,10 @@ } } return EFTransactionTest(value); } - case "update": - { - return UpdateTest(); - } default: { Console.WriteLine("unknown test \"{0}\"", arg); return 1; } @@ -329,81 +325,10 @@ } #endif } } - return 0; - } - - // - // NOTE: Used to test the UPDATE fix (i.e. the missing semi-colon - // in the SQL statement between the actual UPDATE statement - // and the follow-up SELECT statement). - // - private static int UpdateTest() - { - long[] orderIds = new long[] { - 0 - }; - - using (northwindEFEntities db = new northwindEFEntities()) - { - int[] counts = { 0, 0 }; - - // - // NOTE: *REQUIRED* This is required so that the - // Entity Framework is prevented from opening - // multiple connections to the underlying SQLite - // database (i.e. which would result in multiple - // IMMEDIATE transactions, thereby failing [later - // on] with locking errors). - // - db.Connection.Open(); - - for (int index = 0; index < orderIds.Length; index++) - { - Orders newOrders = new Orders(); - - newOrders.ShipAddress = String.Format( - "Test Order Ship Address, Index #{0}", - index); - - db.Orders.AddObject(newOrders); - - try - { - db.SaveChanges(); - counts[0]++; - - // StoreGeneratedPattern="Identity" - orderIds[index] = newOrders.OrderID; - - // StoreGeneratedPattern="None" - newOrders.ShipAddress = String.Format( - "New Order Ship Address #{0}", - orderIds[index]); - - // StoreGeneratedPattern="Computed" - newOrders.Freight = 1; - - db.SaveChanges(); - counts[1]++; - } - catch (Exception e) - { - Console.WriteLine(e); - } - finally - { - db.AcceptAllChanges(); - } - } - - Console.WriteLine( - "inserted {0} updated {1}", counts[0], counts[1]); - } - return 0; } private static int DateTimeTest() { Index: testlinq/Properties/AssemblyInfo.cs ================================================================== --- testlinq/Properties/AssemblyInfo.cs +++ testlinq/Properties/AssemblyInfo.cs @@ -44,7 +44,7 @@ // Revision // // You can specify all the values or you can default the Build and Revision Numbers // by using the '*' as shown below: // [assembly: AssemblyVersion("1.0.*")] -[assembly: AssemblyVersion("1.0.95.0")] -[assembly: AssemblyFileVersion("1.0.95.0")] +[assembly: AssemblyVersion("1.0.94.0")] +[assembly: AssemblyFileVersion("1.0.94.0")] Index: tools/install/Properties/AssemblyInfo.cs ================================================================== --- tools/install/Properties/AssemblyInfo.cs +++ tools/install/Properties/AssemblyInfo.cs @@ -26,7 +26,7 @@ // Major Version // Minor Version // Build Number // Revision // -[assembly: AssemblyVersion("1.0.95.0")] -[assembly: AssemblyFileVersion("1.0.95.0")] +[assembly: AssemblyVersion("1.0.94.0")] +[assembly: AssemblyFileVersion("1.0.94.0")] Index: www/downloads.wiki ================================================================== --- www/downloads.wiki +++ www/downloads.wiki @@ -431,20 +431,20 @@ - sqlite-netFx-source-1.0.94.1.zip + sqlite-netFx-source-1.0.94.0.zip
(4.93 MiB)This ZIP archive contains all current source code for System.Data.SQLite 1.0.94.0 (3.8.6) combined into a single archive file.
- (sha1: c59b48c22a0471634adb204f920ec397f69cfe1d) + (sha1: d581ac783315e51bde70496f15ee69f181273b80)@@ -1849,11 +1849,11 @@ - System.Data.SQLite.1.0.94.1.nupkg + System.Data.SQLite.Test.1.0.94.2.nupkg
(0.00 MiB)@@ -1873,18 +1873,18 @@ depend on the "System.Data.SQLite.Core", "System.Data.SQLite.Linq", and "System.Data.SQLite.EF6" packages.
- (sha1: 45eaa9eab15571878adc89750a491e1990d63e12) + (sha1: c8903e4875b8e4a45790d696f77a150de3022a20)- System.Data.SQLite.Core.1.0.94.0.nupkg + System.Data.SQLite.Core.Test.1.0.94.2.nupkg
(4.92 MiB)@@ -1896,20 +1896,20 @@ the .NET Framework used by the associated managed assembly is statically linked.
This package does not depend on any other package.
- (sha1: 1ccad1b567f2d5d86d6f487e3613e605d24b6b07) + (sha1: 9b1ee72e9b14452e75cf4c85200618cfff6cac05)- System.Data.SQLite.Core.MSIL.1.0.94.0.nupkg + System.Data.SQLite.Core.MSIL.Test.1.0.94.2.nupkg
- (0.68 MiB) + (0.69 MiB)This NuGet package contains the managed binaries for System.Data.SQLite 1.0.94.0. The .NET Framework 3.5 SP1, 4.0, 4.5, or 4.5.1 is required. @@ -1922,18 +1922,18 @@ assembly "SQLite.Interop.dll" compiled for the processor architecture of the host process being present in a directory in the native library search path.
- (sha1: 5ab704db605288ded3ac37787303d7074e4988bf) + (sha1: e2c0d88d3adf0ba1c3b7aa0d8ce4dfe368679018)- System.Data.SQLite.EF6.1.0.94.0.nupkg + System.Data.SQLite.EF6.Test.1.0.94.2.nupkg
(0.13 MiB)@@ -1943,18 +1943,18 @@
This package depends on both the "System.Data.SQLite.Core" and "EntityFramework" packages.
- (sha1: fc4b0046b083a9fc7a294df3af3b9d3918ee88d8) + (sha1: 537a9c7a8054f32973bed591d70636a70206fe9a)- System.Data.SQLite.Linq.1.0.94.1.nupkg + System.Data.SQLite.Linq.Test.1.0.94.2.nupkg
(0.17 MiB)@@ -1963,18 +1963,18 @@ 4.5, or 4.5.1 is required.
This package depends on the "System.Data.SQLite.Core" package.
- (sha1: a08052b4d350e3ba830ebf0618bc8c221a2ef70e) + (sha1: 6c827573479e5141e7c501f5f83e7c1b97d399b9)- System.Data.SQLite.MSIL.1.0.94.1.nupkg + System.Data.SQLite.MSIL.Test.1.0.94.2.nupkg
(0.99 MiB)@@ -1992,59 +1992,11 @@ assembly "SQLite.Interop.dll" compiled for the processor architecture of the host process being present in a directory in the native library search path. -
- (sha1: 7689db66a5b1f2c9c751dbece99ec2faa840f48d) -- - -- System.Data.SQLite.x86.1.0.94.1.nupkg - -
- (2.84 MiB) -- - This NuGet package contains all the binaries for the x86 version of - System.Data.SQLite 1.0.94.0 (3.8.6). The .NET Framework 3.5 SP1, - 4.0, 4.5, or 4.5.1 is required. For the included native binaries, the - version of the Visual C++ runtime corresponding to the .NET Framework - used by the associated managed assembly is statically linked. - -
- This is a legacy package; if possible, please use either the - "System.Data.SQLite" or "System.Data.SQLite.Core" - package instead. This package depends on the "EntityFramework" - package. -
- (sha1: 501c8252fc23134ab99bc6c170a66732578e4a73) -- - System.Data.SQLite.x64.1.0.94.1.nupkg - -
- (3.11 MiB) -- - This NuGet package contains all the binaries for the x64 version of - System.Data.SQLite 1.0.94.0 (3.8.6). The .NET Framework 3.5 SP1, - 4.0, 4.5, or 4.5.1 is required. For the included native binaries, the - version of the Visual C++ runtime corresponding to the .NET Framework - used by the associated managed assembly is statically linked. -
- This is a legacy package; if possible, please use either the - "System.Data.SQLite" or "System.Data.SQLite.Core" - package instead. This package depends on the - "EntityFramework" package. -
- (sha1: 0c9fffde7bd3ed7181ad5aa4dba8e372abe54f0c) + (sha1: 4f9b8b813d76d61f1f0ead5b5be30b89444ecb76)Index: www/news.wiki ================================================================== --- www/news.wiki +++ www/news.wiki @@ -1,18 +1,10 @@ News Version History- 1.0.95.0 - November XX, 2014 (release scheduled) -
--
-
1.0.94.0 - September 9, 2014
Distributing the Binaries (Compact Framework)
Both the System.Data.SQLite.DLL and SQLite.Interop.XXX.DLL files must be deployed on the Compact Framework. The XXX is the build number of - the System.Data.SQLite library (e.g. "095"). The + the System.Data.SQLite library (e.g. "094"). The SQLite.Interop.XXX.DLL file is a fully native assembly compiled for the ARM processor, and System.Data.SQLite is the fully-managed Compact Framework assembly.
Index: Externals/Eagle/bin/Eagle.dll
==================================================================
--- Externals/Eagle/bin/Eagle.dll
+++ Externals/Eagle/bin/Eagle.dll
cannot compute difference between binary files
Index: Externals/Eagle/bin/EagleShell.exe
==================================================================
--- Externals/Eagle/bin/EagleShell.exe
+++ Externals/Eagle/bin/EagleShell.exe
cannot compute difference between binary files
Index: Externals/Eagle/bin/EagleShell32.exe
==================================================================
--- Externals/Eagle/bin/EagleShell32.exe
+++ Externals/Eagle/bin/EagleShell32.exe
cannot compute difference between binary files
Index: Externals/Eagle/bin/x64/Spilornis.dll
==================================================================
--- Externals/Eagle/bin/x64/Spilornis.dll
+++ Externals/Eagle/bin/x64/Spilornis.dll
cannot compute difference between binary files
Index: Externals/Eagle/bin/x86/Spilornis.dll
==================================================================
--- Externals/Eagle/bin/x86/Spilornis.dll
+++ Externals/Eagle/bin/x86/Spilornis.dll
cannot compute difference between binary files
Index: Externals/Eagle/lib/Eagle1.0/init.eagle
==================================================================
--- Externals/Eagle/lib/Eagle1.0/init.eagle
+++ Externals/Eagle/lib/Eagle1.0/init.eagle
@@ -46,22 +46,10 @@
proc isWindows {} {
return [expr {[info exists ::tcl_platform(platform)] && \
$::tcl_platform(platform) eq "windows"}]
}
- #
- # NOTE: This procedure should return non-zero if and only if only there
- # is currently an interactive user that can respond to prompts and
- # other requests for input.
- #
- proc isInteractive {} {
- #
- # TODO: Is more complex checking required here?
- #
- return [expr {[info exists ::tcl_interactive] && $::tcl_interactive}]
- }
-
proc haveGaruda { {varName ""} } {
#
# NOTE: Check for a variable name to place the Garuda package Id into.
#
if {[string length $varName] > 0} then {
@@ -1307,13 +1295,11 @@
# NOTE: This proc is used to check for new versions -OR- new update
# scripts for the runtime when a user executes the interactive
# "#check" command. To disable this functionality, simply
# redefine this procedure to do nothing.
#
- proc checkForUpdate {
- {wantScripts false} {quiet false} {prompt false}
- {automatic false} } {
+ proc checkForUpdate { {wantScripts false} {quiet false} } {
#
# NOTE: This should work properly in Eagle only.
#
set updateUri [appendArgs \
[info engine UpdateBaseUri] [info engine UpdatePathAndQuery]]
@@ -1527,35 +1513,15 @@
# NOTE: The engine patch level from the line is greater,
# we are out-of-date. Return the result of our
# checking now.
#
if {$checkBuild} then {
- #
- # NOTE: Are we supposed to prompt the interactive user,
- # if any, to upgrade now?
- #
- set text [appendArgs \
+ return [list [appendArgs \
"latest build " $patchLevel ", dated " $dateTime \
", is newer than the running build " $enginePatchLevel \
- ", dated " $engineDateTime]
-
- if {$prompt && [isInteractive]} then {
- set caption [appendArgs \
- [info engine Name] " " [lindex [info level 0] 0]]
-
- if {[object invoke -flags +NonPublic \
- Eagle._Components.Private.WindowOps YesOrNo \
- [appendArgs $text \n\n "Run the updater now?"] \
- $caption false]} then {
- #
- # NOTE: Ok, run the updater now and then exit.
- #
- runUpdateAndExit $automatic
- }
- }
-
- return [list $text [list $baseUri $patchLevel] [list $notes]]
+ ", dated " $engineDateTime] [list $baseUri \
+ $patchLevel] [list $notes]]
}
#
# NOTE: The script patch level from the line matches the
# current engine patch level exactly, this script
@@ -2341,19 +2307,19 @@
#
# NOTE: Exports the necessary commands from this package and import them
# into the global namespace.
#
exportAndImportPackageCommands [namespace current] [list \
- isEagle isWindows isInteractive haveGaruda isTclThread isMono \
- isSameFileName getEnvironmentVariable combineFlags getCompileInfo \
- getPlatformInfo getPluginPath appendArgs lappendArgs \
- getDictionaryValue getColumnValue getRowColumnValue tqputs tqlog \
- readFile readSharedFile writeFile appendFile appendLogFile \
- appendSharedFile appendSharedLogFile readAsciiFile writeAsciiFile \
- readUnicodeFile writeUnicodeFile getDirResultPath addToPath \
- removeFromPath execShell lshuffle ldifference filter map reduce \
- getLengthModifier debug findDirectories findFiles findFilesRecursive \
+ isEagle isWindows haveGaruda isTclThread isMono isSameFileName \
+ getEnvironmentVariable combineFlags getCompileInfo getPlatformInfo \
+ getPluginPath appendArgs lappendArgs getDictionaryValue \
+ getColumnValue getRowColumnValue tqputs tqlog readFile \
+ readSharedFile writeFile appendFile appendLogFile appendSharedFile \
+ appendSharedLogFile readAsciiFile writeAsciiFile readUnicodeFile \
+ writeUnicodeFile getDirResultPath addToPath removeFromPath execShell \
+ lshuffle ldifference filter map reduce getLengthModifier debug \
+ findDirectories findFiles findFilesRecursive \
exportAndImportPackageCommands] false false
###########################################################################
############################## END Tcl ONLY ###############################
###########################################################################
Index: Externals/Eagle/lib/Eagle1.0/test.eagle
==================================================================
--- Externals/Eagle/lib/Eagle1.0/test.eagle
+++ Externals/Eagle/lib/Eagle1.0/test.eagle
@@ -1467,11 +1467,10 @@
set array(delegates,$index) 0
set array(tcl,$index) 0
set array(tclInterps,$index) 0
set array(tclThreads,$index) 0
set array(tclCommands,$index) 0
- set array(scriptThreads,$index) 0
catch {set array(connections,$index) [llength [info connections]]}
catch {set array(transactions,$index) [llength [info transactions]]}
catch {set array(modules,$index) [llength [info modules]]}
catch {set array(delegates,$index) [llength [info delegates]]}
@@ -1481,20 +1480,10 @@
}
catch {set array(tclInterps,$index) [llength [tcl interps]]}
catch {set array(tclThreads,$index) [llength [tcl threads]]}
catch {set array(tclCommands,$index) [llength [tcl command list]]}
-
- #
- # NOTE: Grab the number of active threads that are active because
- # of ScriptThread object instances. This only works if Eagle
- # is Beta 31 or higher.
- #
- catch {
- set array(scriptThreads,$index) [object invoke -flags +NonPublic \
- ScriptThread activeCount]
- }
}
}
proc reportTestStatistics { channel fileName statsVarName filesVarName } {
set statistics [list afters variables commands procedures namespaces \
@@ -1504,14 +1493,14 @@
#
# TODO: For now, tracking "leaked" assemblies is meaningless because
# the .NET Framework has no way to unload them without tearing
# down the entire application domain.
#
- lappend statistics \
- scopes assemblies processes objects objectCallbacks objectTypes \
- objectInterfaces objectNamespaces connections transactions modules \
- delegates tcl tclInterps tclThreads tclCommands scriptThreads
+ lappend statistics scopes processes objects objectCallbacks \
+ objectTypes objectInterfaces objectNamespaces connections \
+ transactions modules delegates tcl tclInterps tclThreads \
+ tclCommands; # assemblies
}
#
# NOTE: Show what leaked, if anything.
#
Index: Externals/Eagle/lib/Test1.0/constraints.eagle
==================================================================
--- Externals/Eagle/lib/Test1.0/constraints.eagle
+++ Externals/Eagle/lib/Test1.0/constraints.eagle
@@ -56,40 +56,10 @@
[list 2 0] [list 2 2] [list 2 4] [list 2 6] [list 2 8] [list 2 10] \
[list 2 11] [list 2 12] [list 3 0] [list 3 1] [list 3 2] [list 3 3] \
[list 3 4] [list 3 5] [list 3 6]]
}
- #
- # NOTE: This procedure was adapted from the one listed on the Tcl Wiki page
- # at "http://wiki.tcl.tk/43". It is only intended to be used on very
- # small lists because of its heavy use of recursion and complexity on
- # the order of O(N!).
- #
- proc lpermute { list } {
- set length [llength $list]
- if {$length < 2} {
- return [list $list]
- }
- set index 0
- foreach element1 $list {
- if {$index == 0} then {
- set rest [lrange $list 1 end]
- } elseif {$index == $length - 1} then {
- set rest [lrange $list 0 end-1]
- } else {
- set rest [concat \
- [lrange $list 0 [expr {$index - 1}]] \
- [lrange $list [expr {$index + 1}] end]]
- }
- incr index
- foreach list2 [lpermute $rest] {
- lappend result [concat [list $element1] $list2]
- }
- }
- return $result
- }
-
proc alwaysFullInterpReady {} {
#
# NOTE: The [interp readylimit] sub-command is only in Eagle.
#
if {![isEagle]} then {
@@ -1025,12 +995,12 @@
tputs $channel unknown\n
}
}
proc checkForFile { channel name {constraint ""} } {
- tputs $channel [appendArgs "---- checking for file \"" \
- [file normalize $name] "\"... "]
+ tputs $channel [appendArgs "---- checking for file \"" $name \
+ "\"... "]
if {[file exists $name]} then {
#
# NOTE: Yes, it appears that it is available.
#
@@ -1568,15 +1538,15 @@
tputs $channel [appendArgs \
"---- checking for network connectivity to host \"" $host "\"... "]
if {[isEagle]} then {
#
- # NOTE: Running this check on the Mono 3.3.0 (or 3.4.0?) release build
- # will lock up the process; therefore, skip it in that case.
+ # NOTE: Running this check on the Mono 3.3.0 release build will lock
+ # up the process; therefore, attempt to skip it in that case.
#
if {[info exists ::no(mono)] || ![isMono] || \
- (![haveConstraint mono33] && ![haveConstraint mono34])} then {
+ ![haveConstraint mono33]} then {
#
# BUGBUG: Tcl 8.4 does not like this expression (and Tcl tries to
# compile it even though it will only actually ever be
# evaluated in Eagle).
#
@@ -1598,11 +1568,11 @@
tputs $channel [appendArgs "yes (" $response ")\n"]
} else {
tputs $channel no\n
}
} else {
- tputs $channel "skipped, may hang on Mono 3.3.0 and 3.4.0\n"
+ tputs $channel "skipped, broken on Mono 3.3.0\n"
}
} else {
#
# HACK: Running in Tcl, just assume we have network access.
#
@@ -2174,23 +2144,10 @@
} else {
tputs $channel no\n
}
}
- proc checkForProcessBits { channel } {
- tputs $channel "---- checking for process bits... "
-
- if {[info exists ::tcl_platform(processBits)] && \
- [string is integer -strict $::tcl_platform(processBits)]} then {
- addConstraint [appendArgs $::tcl_platform(processBits) bit]
-
- tputs $channel [appendArgs $::tcl_platform(processBits) -bit\n]
- } else {
- tputs $channel unknown\n
- }
- }
-
proc checkForMachine { channel bits machine } {
tputs $channel [appendArgs "---- checking for machine \"" $bits \
"-bit " $machine "\"... "]
#
@@ -3229,26 +3186,25 @@
#
# NOTE: We need several of our test constraint related commands in the
# global namespace.
#
exportAndImportPackageCommands [namespace current] [list \
- getKnownCompileOptions getKnownMonoVersions lpermute \
- alwaysFullInterpReady canExecTclShell canExecFossil \
- checkForTestSuiteFiles checkForPlatform checkForWindowsVersion \
- checkForScriptLibrary checkForVariable checkForTclOptions \
- checkForWindowsCommandProcessor checkForFossil checkForEagle \
- checkForSymbols checkForLogFile checkForGaruda checkForShell \
- checkForDebug checkForTk checkForVersion checkForCommand \
- checkForNamespaces checkForTestExec checkForTestMachine \
- checkForTestPlatform checkForTestConfiguration checkForTestSuffix \
- checkForFile checkForPathFile checkForNativeCode checkForTip127 \
- checkForTip194 checkForTip207 checkForTip241 checkForTip285 \
- checkForTip405 checkForTip426 checkForTip429 checkForTiming \
- checkForPerformance checkForBigLists checkForMemoryIntensive \
- checkForStackIntensive checkForInteractive checkForInteractiveCommand \
- checkForUserInteraction checkForNetwork checkForCompileOption \
- checkForKnownCompileOptions] false false
+ getKnownCompileOptions getKnownMonoVersions alwaysFullInterpReady \
+ canExecTclShell canExecFossil checkForTestSuiteFiles checkForPlatform \
+ checkForWindowsVersion checkForScriptLibrary checkForVariable \
+ checkForTclOptions checkForWindowsCommandProcessor checkForFossil \
+ checkForEagle checkForSymbols checkForLogFile checkForGaruda \
+ checkForShell checkForDebug checkForTk checkForVersion \
+ checkForCommand checkForNamespaces checkForTestExec \
+ checkForTestMachine checkForTestPlatform checkForTestConfiguration \
+ checkForTestSuffix checkForFile checkForPathFile checkForNativeCode \
+ checkForTip127 checkForTip194 checkForTip207 checkForTip241 \
+ checkForTip285 checkForTip405 checkForTip426 checkForTip429 \
+ checkForTiming checkForPerformance checkForBigLists \
+ checkForMemoryIntensive checkForStackIntensive checkForInteractive \
+ checkForInteractiveCommand checkForUserInteraction checkForNetwork \
+ checkForCompileOption checkForKnownCompileOptions] false false
###########################################################################
############################## END Tcl ONLY ###############################
###########################################################################
}
Index: Externals/Eagle/lib/Test1.0/prologue.eagle
==================================================================
--- Externals/Eagle/lib/Test1.0/prologue.eagle
+++ Externals/Eagle/lib/Test1.0/prologue.eagle
@@ -977,17 +977,10 @@
#
if {![info exists no(imageRuntimeVersion)]} then {
checkForImageRuntimeVersion $test_channel
}
- #
- # NOTE: Has process bits detection support been disabled?
- #
- if {![info exists no(processBits)]} then {
- checkForProcessBits $test_channel
- }
-
#
# NOTE: Has machine detection support been disabled?
#
if {![info exists no(machine)]} then {
checkForMachine $test_channel 32 intel; # (i.e. x86)
@@ -2698,19 +2691,10 @@
}
}
tputs $test_channel [appendArgs $test_base_cops \n]
tputs $test_channel \
- "---- checking for baseline BogoCops multipliers... "
-
- if {![info exists test_base_cops_multipliers]} then {
- set test_base_cops_multipliers [list \
- 2 4 8 10 16 20 32 50 64 100 1000 10000 100000 1000000]
- }
-
- tputs $test_channel [appendArgs $test_base_cops_multipliers \n]
- tputs $test_channel \
"---- checking for current BogoCops (commands-per-second)... "
if {![info exists test_cops]} then {
set test_cops [calculateBogoCops]
}
@@ -2718,49 +2702,23 @@
tputs $test_channel [appendArgs [formatDecimal $test_cops] \n]
tputs $test_channel \
"---- checking for current BogoCops against baseline BogoCops... "
if {$test_cops > $test_base_cops} then {
- if {![info exists no(fasterBogoCops)]} then {
- addConstraint sameOrFasterBogoCops
- addConstraint fasterBogoCops
-
- if {![info exists no(bogoCopsMultipliers)]} then {
- foreach multiplier $test_base_cops_multipliers {
- if {$test_cops >= $test_base_cops * $multiplier} then {
- addConstraint [appendArgs fasterBogoCopsBy $multiplier X]
- }
- }
-
- unset multiplier
- }
- }
+ addConstraint sameOrFasterBogoCops
+ addConstraint fasterBogoCops
tputs $test_channel faster\n
} elseif {$test_cops < $test_base_cops} then {
- if {![info exists no(slowerBogoCops)]} then {
- addConstraint sameOrSlowerBogoCops
- addConstraint slowerBogoCops
-
- if {![info exists no(bogoCopsMultipliers)]} then {
- foreach multiplier $test_base_cops_multipliers {
- if {$test_cops <= $test_base_cops * (1.0 / $multiplier)} then {
- addConstraint [appendArgs slowerBogoCopsBy $multiplier X]
- }
- }
-
- unset multiplier
- }
- }
+ addConstraint sameOrSlowerBogoCops
+ addConstraint slowerBogoCops
tputs $test_channel slower\n
} else {
- if {![info exists no(sameBogoCops)]} then {
- addConstraint sameOrFasterBogoCops
- addConstraint sameOrSlowerBogoCops
- addConstraint sameBogoCops
- }
+ addConstraint sameOrFasterBogoCops
+ addConstraint sameOrSlowerBogoCops
+ addConstraint sameBogoCops
tputs $test_channel same\n
}
set percent [expr {[calculateRelativePerformance iterations 1] * 100}]
Index: NuGet/SQLite.Beta.nuspec
==================================================================
--- NuGet/SQLite.Beta.nuspec
+++ NuGet/SQLite.Beta.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.Beta
System.Data.SQLite (x86/x64) Beta
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
This is a "beta" package and is not intended for production use. The official SQLite database engine for both x86 and x64 along with the ADO.NET provider. This package includes support for LINQ and Entity Framework 6.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
@@ -20,41 +20,41 @@
http://www.sqlite.org/copyright.html
sqlite database ado.net provider interop
Public Domain
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
+
+
+
+
+
+
+
+
+
+
+
+
Index: NuGet/SQLite.Core.Beta.nuspec
==================================================================
--- NuGet/SQLite.Core.Beta.nuspec
+++ NuGet/SQLite.Core.Beta.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.Core.Beta
System.Data.SQLite Core (x86/x64) Beta
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
This is a "beta" package and is not intended for production use. The official SQLite database engine for both x86 and x64 along with the ADO.NET provider.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
Index: NuGet/SQLite.Core.MSIL.Beta.nuspec
==================================================================
--- NuGet/SQLite.Core.MSIL.Beta.nuspec
+++ NuGet/SQLite.Core.MSIL.Beta.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.Core.MSIL.Beta
System.Data.SQLite (MSIL) Beta
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
This is a "beta" package and is not intended for production use. An ADO.NET provider for SQLite (managed-only).
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
Index: NuGet/SQLite.Core.MSIL.Test.nuspec
==================================================================
--- NuGet/SQLite.Core.MSIL.Test.nuspec
+++ NuGet/SQLite.Core.MSIL.Test.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.Core.MSIL.Test
System.Data.SQLite (MSIL) Test
- 1.0.95.0
+ 1.0.94.2
SQLite Development Team
This is a pre-release package and is not intended for production use. An ADO.NET provider for SQLite (managed-only).
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
Index: NuGet/SQLite.Core.MSIL.nuspec
==================================================================
--- NuGet/SQLite.Core.MSIL.nuspec
+++ NuGet/SQLite.Core.MSIL.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.Core.MSIL
System.Data.SQLite (MSIL)
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
An ADO.NET provider for SQLite (managed-only).
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
Index: NuGet/SQLite.Core.Test.nuspec
==================================================================
--- NuGet/SQLite.Core.Test.nuspec
+++ NuGet/SQLite.Core.Test.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.Core.Test
System.Data.SQLite Core (x86/x64) Test
- 1.0.95.0
+ 1.0.94.2
SQLite Development Team
This is a pre-release package and is not intended for production use. The official SQLite database engine for both x86 and x64 along with the ADO.NET provider.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
Index: NuGet/SQLite.Core.nuspec
==================================================================
--- NuGet/SQLite.Core.nuspec
+++ NuGet/SQLite.Core.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.Core
System.Data.SQLite Core (x86/x64)
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
The official SQLite database engine for both x86 and x64 along with the ADO.NET provider.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
Index: NuGet/SQLite.EF6.Beta.nuspec
==================================================================
--- NuGet/SQLite.EF6.Beta.nuspec
+++ NuGet/SQLite.EF6.Beta.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.EF6.Beta
System.Data.SQLite EF6 (x86/x64) Beta
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
This is a "beta" package and is not intended for production use. Support for Entity Framework 6 using System.Data.SQLite.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
@@ -20,17 +20,17 @@
http://www.sqlite.org/copyright.html
sqlite database ado.net provider interop
Public Domain
-
-
-
Index: NuGet/SQLite.EF6.Test.nuspec
==================================================================
--- NuGet/SQLite.EF6.Test.nuspec
+++ NuGet/SQLite.EF6.Test.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.EF6.Test
System.Data.SQLite EF6 (x86/x64) Test
- 1.0.95.0
+ 1.0.94.2
SQLite Development Team
This is a pre-release package and is not intended for production use. Support for Entity Framework 6 using System.Data.SQLite.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
@@ -20,17 +20,17 @@
http://www.sqlite.org/copyright.html
sqlite database ado.net provider interop
Public Domain
-
-
-
Index: NuGet/SQLite.EF6.nuspec
==================================================================
--- NuGet/SQLite.EF6.nuspec
+++ NuGet/SQLite.EF6.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.EF6
System.Data.SQLite EF6 (x86/x64)
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
Support for Entity Framework 6 using System.Data.SQLite.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
@@ -20,17 +20,17 @@
http://www.sqlite.org/copyright.html
sqlite database ado.net provider interop
Public Domain
-
-
-
Index: NuGet/SQLite.Linq.Beta.nuspec
==================================================================
--- NuGet/SQLite.Linq.Beta.nuspec
+++ NuGet/SQLite.Linq.Beta.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.Linq.Beta
System.Data.SQLite LINQ (x86/x64) Beta
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
This is a "beta" package and is not intended for production use. Support for LINQ using System.Data.SQLite.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
@@ -20,20 +20,20 @@
http://www.sqlite.org/copyright.html
sqlite database ado.net provider interop
Public Domain
-
-
-
-
System.Data.SQLite.Linq.Test
System.Data.SQLite LINQ (x86/x64) Test
- 1.0.95.0
+ 1.0.94.2
SQLite Development Team
This is a pre-release package and is not intended for production use. Support for LINQ using System.Data.SQLite.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
@@ -20,20 +20,20 @@
http://www.sqlite.org/copyright.html
sqlite database ado.net provider interop
Public Domain
-
-
-
-
System.Data.SQLite.Linq
System.Data.SQLite LINQ (x86/x64)
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
Support for LINQ using System.Data.SQLite.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
@@ -20,20 +20,20 @@
http://www.sqlite.org/copyright.html
sqlite database ado.net provider interop
Public Domain
-
-
-
-
System.Data.SQLite.MSIL.Beta
System.Data.SQLite (MSIL) Beta
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
This is a "beta" package and is not intended for production use. This is a legacy package; if possible, please use either the "System.Data.SQLite.Beta" or "System.Data.SQLite.Core.Beta" package instead. An ADO.NET provider for SQLite (managed-only).
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
Index: NuGet/SQLite.MSIL.Test.nuspec
==================================================================
--- NuGet/SQLite.MSIL.Test.nuspec
+++ NuGet/SQLite.MSIL.Test.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.MSIL.Test
System.Data.SQLite (MSIL) Test
- 1.0.95.0
+ 1.0.94.2
SQLite Development Team
This is a pre-release package and is not intended for production use. This is a legacy package; if possible, please use either the "System.Data.SQLite.Test" or "System.Data.SQLite.Core.Test" package instead. An ADO.NET provider for SQLite (managed-only).
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
Index: NuGet/SQLite.MSIL.nuspec
==================================================================
--- NuGet/SQLite.MSIL.nuspec
+++ NuGet/SQLite.MSIL.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.MSIL
System.Data.SQLite (MSIL)
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
This is a legacy package; if possible, please use either the "System.Data.SQLite" or "System.Data.SQLite.Core" package instead. An ADO.NET provider for SQLite (managed-only).
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
Index: NuGet/SQLite.Test.nuspec
==================================================================
--- NuGet/SQLite.Test.nuspec
+++ NuGet/SQLite.Test.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite.Test
System.Data.SQLite (x86/x64) Test
- 1.0.95.0
+ 1.0.94.2
SQLite Development Team
This is a pre-release package and is not intended for production use. The official SQLite database engine for both x86 and x64 along with the ADO.NET provider. This package includes support for LINQ and Entity Framework 6.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
@@ -20,41 +20,41 @@
http://www.sqlite.org/copyright.html
sqlite database ado.net provider interop
Public Domain
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
+
+
+
+
+
+
+
+
+
+
+
+
Index: NuGet/SQLite.nuspec
==================================================================
--- NuGet/SQLite.nuspec
+++ NuGet/SQLite.nuspec
@@ -9,11 +9,11 @@
-->
System.Data.SQLite
System.Data.SQLite (x86/x64)
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
The official SQLite database engine for both x86 and x64 along with the ADO.NET provider. This package includes support for LINQ and Entity Framework 6.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
@@ -20,41 +20,41 @@
http://www.sqlite.org/copyright.html
sqlite database ado.net provider interop
Public Domain
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
+
+
+
+
+
+
+
+
+
+
+
+
Index: NuGet/SQLite.x64.nuspec
==================================================================
--- NuGet/SQLite.x64.nuspec
+++ NuGet/SQLite.x64.nuspec
@@ -8,11 +8,11 @@
*
-->
System.Data.SQLite.x64
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
This is a legacy package; if possible, please use either the "System.Data.SQLite" or "System.Data.SQLite.Core" package instead. The official SQLite database engine combined with a complete ADO.NET provider all rolled into a single mixed-mode assembly for x64.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
Index: NuGet/SQLite.x86.nuspec
==================================================================
--- NuGet/SQLite.x86.nuspec
+++ NuGet/SQLite.x86.nuspec
@@ -8,11 +8,11 @@
*
-->
System.Data.SQLite.x86
- 1.0.95.0
+ 1.0.94.0
SQLite Development Team
This is a legacy package; if possible, please use either the "System.Data.SQLite" or "System.Data.SQLite.Core" package instead. The official SQLite database engine combined with a complete ADO.NET provider all rolled into a single mixed-mode assembly for x86.
en-US
http://system.data.sqlite.org/
http://system.data.sqlite.org/images/sqlite128.png
Index: NuGet/shared/Core/build/System.Data.SQLite.Core.targets
==================================================================
--- NuGet/shared/Core/build/System.Data.SQLite.Core.targets
+++ NuGet/shared/Core/build/System.Data.SQLite.Core.targets
@@ -27,37 +27,37 @@
******************************************************************************
-->
+ Outputs="@(SQLiteInteropFiles -> '$(OutputPath)%(RecursiveDir)%(Filename)%(Extension)')">
+ '$(OutputPath)' != '' And
+ HasTrailingSlash('$(OutputPath)') And
+ Exists('$(OutputPath)')">
-
2010
- 095
+ 094
1.0
- 1.0.95.0
- 1,0,95,0
+ 1.0.94.0
+ 1,0,94,0
src\core
INTEROP_DEBUG=0x31F;INTEROP_LOG=1;INTEROP_TEST_EXTENSION=1
INTEROP_EXTENSION_FUNCTIONS=1;INTEROP_CODEC=1;INTEROP_VIRTUAL_TABLE=1;INTEROP_PERCENTILE_EXTENSION=1;INTEROP_TOTYPE_EXTENSION=1;INTEROP_REGEXP_EXTENSION=1
Index: SQLite.Interop/props/SQLite.Interop.2012.props
==================================================================
--- SQLite.Interop/props/SQLite.Interop.2012.props
+++ SQLite.Interop/props/SQLite.Interop.2012.props
@@ -8,14 +8,14 @@
*
-->
2012
- 095
+ 094
1.0
- 1.0.95.0
- 1,0,95,0
+ 1.0.94.0
+ 1,0,94,0
src\core
INTEROP_DEBUG=0x31F;INTEROP_LOG=1;INTEROP_TEST_EXTENSION=1
INTEROP_EXTENSION_FUNCTIONS=1;INTEROP_CODEC=1;INTEROP_VIRTUAL_TABLE=1;INTEROP_PERCENTILE_EXTENSION=1;INTEROP_TOTYPE_EXTENSION=1;INTEROP_REGEXP_EXTENSION=1
Index: SQLite.Interop/props/SQLite.Interop.2013.props
==================================================================
--- SQLite.Interop/props/SQLite.Interop.2013.props
+++ SQLite.Interop/props/SQLite.Interop.2013.props
@@ -8,14 +8,14 @@
*
-->
2013
- 095
+ 094
1.0
- 1.0.95.0
- 1,0,95,0
+ 1.0.94.0
+ 1,0,94,0
src\core
INTEROP_DEBUG=0x31F;INTEROP_LOG=1;INTEROP_TEST_EXTENSION=1
INTEROP_EXTENSION_FUNCTIONS=1;INTEROP_CODEC=1;INTEROP_VIRTUAL_TABLE=1;INTEROP_PERCENTILE_EXTENSION=1;INTEROP_TOTYPE_EXTENSION=1;INTEROP_REGEXP_EXTENSION=1
Index: SQLite.Interop/props/sqlite3.props
==================================================================
--- SQLite.Interop/props/sqlite3.props
+++ SQLite.Interop/props/sqlite3.props
@@ -7,12 +7,12 @@
* Released to the public domain, use at your own risk!
*
-->
- 3.8.7
- 3,8,7
+ 3.8.6
+ 3,8,6
_CRT_SECURE_NO_DEPRECATE;_CRT_SECURE_NO_WARNINGS;_CRT_NONSTDC_NO_DEPRECATE;_CRT_NONSTDC_NO_WARNINGS;SQLITE_THREADSAFE=1;SQLITE_USE_URI=1;SQLITE_ENABLE_COLUMN_METADATA=1;SQLITE_ENABLE_STAT4=1;SQLITE_ENABLE_FTS3=1;SQLITE_ENABLE_LOAD_EXTENSION=1;SQLITE_ENABLE_RTREE=1;SQLITE_SOUNDEX=1;SQLITE_ENABLE_MEMORY_MANAGEMENT=1
SQLITE_HAS_CODEC=1
SQLITE_OMIT_WAL=1
HAVE_ERRNO_H=1;SQLITE_MSVC_LOCALTIME_API=1
SQLITE_DEBUG=1;SQLITE_MEMDEBUG=1;SQLITE_ENABLE_EXPENSIVE_ASSERT=1
Index: SQLite.Interop/props/sqlite3.vsprops
==================================================================
--- SQLite.Interop/props/sqlite3.vsprops
+++ SQLite.Interop/props/sqlite3.vsprops
@@ -12,16 +12,16 @@
Version="8.00"
Name="sqlite3"
>
= 3008007
- CollSeq *pColl = sqlite3GetFuncCollSeq(context);
-#else
CollSeq *pColl = context->pColl;
-#endif
assert( argc==3 ||argc==2);
if( SQLITE_NULL==sqlite3_value_type(argv[0]) || SQLITE_NULL==sqlite3_value_type(argv[1])){
sqlite3_result_null(context);
Index: SQLite.Interop/src/core/sqlite3.c
==================================================================
--- SQLite.Interop/src/core/sqlite3.c
+++ SQLite.Interop/src/core/sqlite3.c
@@ -1,8 +1,8 @@
/******************************************************************************
** This file is an amalgamation of many separate C source files from SQLite
-** version 3.8.7. By combining all the individual C code files into this
+** version 3.8.6. By combining all the individual C code files into this
** single large file, the entire code can be compiled as a single translation
** unit. This allows many compilers to do optimizations that would not be
** possible if the files were compiled separately. Performance improvements
** of 5% or more are commonly seen when SQLite is compiled as a single
** translation unit.
@@ -73,19 +73,10 @@
# define _FILE_OFFSET_BITS 64
# endif
# define _LARGEFILE_SOURCE 1
#endif
-/* Needed for various definitions... */
-#if defined(__GNUC__) && !defined(_GNU_SOURCE)
-# define _GNU_SOURCE
-#endif
-
-#if defined(__OpenBSD__) && !defined(_BSD_SOURCE)
-# define _BSD_SOURCE
-#endif
-
/*
** For MinGW, check to see if we can include the header file containing its
** version information, among other things. Normally, this internal MinGW
** header file would [only] be included automatically by other MinGW header
** files; however, the contained version information is now required by this
@@ -229,13 +220,13 @@
**
** See also: [sqlite3_libversion()],
** [sqlite3_libversion_number()], [sqlite3_sourceid()],
** [sqlite_version()] and [sqlite_source_id()].
*/
-#define SQLITE_VERSION "3.8.7"
-#define SQLITE_VERSION_NUMBER 3008007
-#define SQLITE_SOURCE_ID "2014-10-07 23:02:24 f672a380e2e52bede95ff11a533fd9f7d412d494"
+#define SQLITE_VERSION "3.8.6"
+#define SQLITE_VERSION_NUMBER 3008006
+#define SQLITE_SOURCE_ID "2014-08-15 11:46:33 9491ba7d738528f168657adb43a198238abde19e"
/*
** CAPI3REF: Run-Time Library Version Numbers
** KEYWORDS: sqlite3_version, sqlite3_sourceid
**
@@ -619,11 +610,10 @@
#define SQLITE_CONSTRAINT_VTAB (SQLITE_CONSTRAINT | (9<<8))
#define SQLITE_CONSTRAINT_ROWID (SQLITE_CONSTRAINT |(10<<8))
#define SQLITE_NOTICE_RECOVER_WAL (SQLITE_NOTICE | (1<<8))
#define SQLITE_NOTICE_RECOVER_ROLLBACK (SQLITE_NOTICE | (2<<8))
#define SQLITE_WARNING_AUTOINDEX (SQLITE_WARNING | (1<<8))
-#define SQLITE_AUTH_USER (SQLITE_AUTH | (1<<8))
/*
** CAPI3REF: Flags For File Open Operations
**
** These bit values are intended for use in the
@@ -2222,11 +2212,11 @@
**
** ^Calling this routine with an argument less than or equal to zero
** turns off all busy handlers.
**
** ^(There can only be a single busy handler for a particular
-** [database connection] at any given moment. If another busy handler
+** [database connection] any any given moment. If another busy handler
** was defined (using [sqlite3_busy_handler()]) prior to calling
** this routine, that other busy handler is cleared.)^
**
** See also: [PRAGMA busy_timeout]
*/
@@ -2426,14 +2416,10 @@
** ^If sqlite3_malloc() is unable to obtain sufficient free
** memory, it returns a NULL pointer. ^If the parameter N to
** sqlite3_malloc() is zero or negative then sqlite3_malloc() returns
** a NULL pointer.
**
-** ^The sqlite3_malloc64(N) routine works just like
-** sqlite3_malloc(N) except that N is an unsigned 64-bit integer instead
-** of a signed 32-bit integer.
-**
** ^Calling sqlite3_free() with a pointer previously returned
** by sqlite3_malloc() or sqlite3_realloc() releases that memory so
** that it might be reused. ^The sqlite3_free() routine is
** a no-op if is called with a NULL pointer. Passing a NULL pointer
** to sqlite3_free() is harmless. After being freed, memory
@@ -2441,42 +2427,28 @@
** memory might result in a segmentation fault or other severe error.
** Memory corruption, a segmentation fault, or other severe error
** might result if sqlite3_free() is called with a non-NULL pointer that
** was not obtained from sqlite3_malloc() or sqlite3_realloc().
**
-** ^The sqlite3_realloc(X,N) interface attempts to resize a
-** prior memory allocation X to be at least N bytes.
-** ^If the X parameter to sqlite3_realloc(X,N)
+** ^(The sqlite3_realloc() interface attempts to resize a
+** prior memory allocation to be at least N bytes, where N is the
+** second parameter. The memory allocation to be resized is the first
+** parameter.)^ ^ If the first parameter to sqlite3_realloc()
** is a NULL pointer then its behavior is identical to calling
-** sqlite3_malloc(N).
-** ^If the N parameter to sqlite3_realloc(X,N) is zero or
+** sqlite3_malloc(N) where N is the second parameter to sqlite3_realloc().
+** ^If the second parameter to sqlite3_realloc() is zero or
** negative then the behavior is exactly the same as calling
-** sqlite3_free(X).
-** ^sqlite3_realloc(X,N) returns a pointer to a memory allocation
-** of at least N bytes in size or NULL if insufficient memory is available.
+** sqlite3_free(P) where P is the first parameter to sqlite3_realloc().
+** ^sqlite3_realloc() returns a pointer to a memory allocation
+** of at least N bytes in size or NULL if sufficient memory is unavailable.
** ^If M is the size of the prior allocation, then min(N,M) bytes
** of the prior allocation are copied into the beginning of buffer returned
-** by sqlite3_realloc(X,N) and the prior allocation is freed.
-** ^If sqlite3_realloc(X,N) returns NULL and N is positive, then the
-** prior allocation is not freed.
-**
-** ^The sqlite3_realloc64(X,N) interfaces works the same as
-** sqlite3_realloc(X,N) except that N is a 64-bit unsigned integer instead
-** of a 32-bit signed integer.
-**
-** ^If X is a memory allocation previously obtained from sqlite3_malloc(),
-** sqlite3_malloc64(), sqlite3_realloc(), or sqlite3_realloc64(), then
-** sqlite3_msize(X) returns the size of that memory allocation in bytes.
-** ^The value returned by sqlite3_msize(X) might be larger than the number
-** of bytes requested when X was allocated. ^If X is a NULL pointer then
-** sqlite3_msize(X) returns zero. If X points to something that is not
-** the beginning of memory allocation, or if it points to a formerly
-** valid memory allocation that has now been freed, then the behavior
-** of sqlite3_msize(X) is undefined and possibly harmful.
-**
-** ^The memory returned by sqlite3_malloc(), sqlite3_realloc(),
-** sqlite3_malloc64(), and sqlite3_realloc64()
+** by sqlite3_realloc() and the prior allocation is freed.
+** ^If sqlite3_realloc() returns NULL, then the prior allocation
+** is not freed.
+**
+** ^The memory returned by sqlite3_malloc() and sqlite3_realloc()
** is always aligned to at least an 8 byte boundary, or to a
** 4 byte boundary if the [SQLITE_4_BYTE_ALIGNED_MALLOC] compile-time
** option is used.
**
** In SQLite version 3.5.0 and 3.5.1, it was possible to define
@@ -2500,15 +2472,12 @@
** The application must not read or write any part of
** a block of memory after it has been released using
** [sqlite3_free()] or [sqlite3_realloc()].
*/
SQLITE_API void *sqlite3_malloc(int);
-SQLITE_API void *sqlite3_malloc64(sqlite3_uint64);
SQLITE_API void *sqlite3_realloc(void*, int);
-SQLITE_API void *sqlite3_realloc64(void*, sqlite3_uint64);
SQLITE_API void sqlite3_free(void*);
-SQLITE_API sqlite3_uint64 sqlite3_msize(void*);
/*
** CAPI3REF: Memory Allocator Statistics
**
** SQLite provides these two interfaces for reporting on the status
@@ -2791,13 +2760,13 @@
** [SQLITE_OK] is returned. Otherwise an [error code] is returned.)^ ^The
** [sqlite3_errmsg()] or [sqlite3_errmsg16()] routines can be used to obtain
** an English language description of the error following a failure of any
** of the sqlite3_open() routines.
**
-** ^The default encoding will be UTF-8 for databases created using
-** sqlite3_open() or sqlite3_open_v2(). ^The default encoding for databases
-** created using sqlite3_open16() will be UTF-16 in the native byte order.
+** ^The default encoding for the database will be UTF-8 if
+** sqlite3_open() or sqlite3_open_v2() is called and
+** UTF-16 in the native byte order if sqlite3_open16() is used.
**
** Whether or not an error occurs when it is opened, resources
** associated with the [database connection] handle should be released by
** passing it to [sqlite3_close()] when it is no longer required.
**
@@ -2881,18 +2850,17 @@
** ^SQLite uses the path component of the URI as the name of the disk file
** which contains the database. ^If the path begins with a '/' character,
** then it is interpreted as an absolute path. ^If the path does not begin
** with a '/' (meaning that the authority section is omitted from the URI)
** then the path is interpreted as a relative path.
-** ^(On windows, the first component of an absolute path
-** is a drive specification (e.g. "C:").)^
+** ^On windows, the first component of an absolute path
+** is a drive specification (e.g. "C:").
**
** [[core URI query parameters]]
** The query component of a URI may contain parameters that are interpreted
** either by SQLite itself, or by a [VFS | custom VFS implementation].
-** SQLite and its built-in [VFSes] interpret the
-** following query parameters:
+** SQLite interprets the following three query parameters:
**
**
-
**