Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
Overview
Comment: | Merge the latest trunk changes, especially the ORDER BY optimizer enhancements but also other fixes, onto the sessions branch. |
---|---|
Downloads: | Tarball | ZIP archive |
Timelines: | family | ancestors | descendants | both | sessions |
Files: | files | file ages | folders |
SHA1: |
f1fbb8c5bfa84e84e0b8e2872d83b06a |
User & Date: | drh 2012-10-09 14:58:15.175 |
Context
2012-10-15
| ||
14:25 | Merge all the latest core changes into the sessions branch. (check-in: 76767d651f user: drh tags: sessions) | |
2012-10-09
| ||
14:58 | Merge the latest trunk changes, especially the ORDER BY optimizer enhancements but also other fixes, onto the sessions branch. (check-in: f1fbb8c5bf user: drh tags: sessions) | |
14:36 | Avoid an assertion fault and/or freeing memory while it is still in use when an error occurs during virtual table construction. (check-in: a02599ad85 user: drh tags: trunk) | |
2012-09-28
| ||
13:05 | Merge the latest trunk changes (especially "PRAGMA busy_timeout" and the ORDER BY query planner optimizations) into the sessions branch. (check-in: 6ca8eae1f8 user: drh tags: sessions) | |
Changes
Changes to Makefile.msc.
︙ | ︙ | |||
25 26 27 28 29 30 31 32 33 34 35 36 37 38 | # XCOMPILE = 0 # Set this non-0 to use the native libraries paths for cross-compiling # the command line tools needed during the compilation process. # USE_NATIVE_LIBPATHS = 0 # Set this non-0 to compile binaries suitable for the WinRT environment. # This setting does not apply to any binaries that require Tcl to operate # properly (i.e. the text fixture, etc). # FOR_WINRT = 0 | > > > > | 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 | # XCOMPILE = 0 # Set this non-0 to use the native libraries paths for cross-compiling # the command line tools needed during the compilation process. # USE_NATIVE_LIBPATHS = 0 # Set this 0 to skip the compiling and embedding of version resources. # USE_RC = 1 # Set this non-0 to compile binaries suitable for the WinRT environment. # This setting does not apply to any binaries that require Tcl to operate # properly (i.e. the text fixture, etc). # FOR_WINRT = 0 |
︙ | ︙ | |||
164 165 166 167 168 169 170 | # the following compile-time options must be used as well to # disable use of Win32 APIs that are not available and to enable # use of Win32 APIs that are specific to Windows 8 and/or WinRT. # !IF $(FOR_WINRT)!=0 TCC = $(TCC) -DSQLITE_OS_WINRT=1 RCC = $(RCC) -DSQLITE_OS_WINRT=1 | | | > > > > | 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 | # the following compile-time options must be used as well to # disable use of Win32 APIs that are not available and to enable # use of Win32 APIs that are specific to Windows 8 and/or WinRT. # !IF $(FOR_WINRT)!=0 TCC = $(TCC) -DSQLITE_OS_WINRT=1 RCC = $(RCC) -DSQLITE_OS_WINRT=1 TCC = $(TCC) -DWINAPI_FAMILY=WINAPI_FAMILY_APP RCC = $(RCC) -DWINAPI_FAMILY=WINAPI_FAMILY_APP !ENDIF # Also, we need to dynamically link to the correct MSVC runtime # when compiling for WinRT (e.g. debug or release) OR if the # USE_CRT_DLL option is set to force dynamically linking to the # MSVC runtime library. # !IF $(FOR_WINRT)!=0 || $(USE_CRT_DLL)!=0 !IF $(DEBUG)>0 TCC = $(TCC) -MDd BCC = $(BCC) -MDd !ELSE TCC = $(TCC) -MD BCC = $(BCC) -MD !ENDIF !ELSE !IF $(DEBUG)>0 TCC = $(TCC) -MTd BCC = $(BCC) -MTd !ELSE TCC = $(TCC) -MT BCC = $(BCC) -MT !ENDIF !ENDIF # The mksqlite3c.tcl and mksqlite3h.tcl scripts will pull in # any extension header files by default. For non-amalgamation # builds, we need to make sure the compiler can find these. # |
︙ | ︙ | |||
469 470 471 472 473 474 475 476 477 478 479 480 481 482 | # Determine the real value of LIBOBJ based on the 'configure' script # !IF $(USE_AMALGAMATION)==0 LIBOBJ = $(LIBOBJS0) !ELSE LIBOBJ = $(LIBOBJS1) !ENDIF # All of the source code files. # SRC = \ $(TOP)\src\alter.c \ $(TOP)\src\analyze.c \ $(TOP)\src\attach.c \ | > > > > > > > > | 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 | # Determine the real value of LIBOBJ based on the 'configure' script # !IF $(USE_AMALGAMATION)==0 LIBOBJ = $(LIBOBJS0) !ELSE LIBOBJ = $(LIBOBJS1) !ENDIF # Determine if embedded resource compilation and usage are enabled. # !IF $(USE_RC)!=0 LIBRESOBJS = sqlite3res.lo !ELSE LIBRESOBJS = !ENDIF # All of the source code files. # SRC = \ $(TOP)\src\alter.c \ $(TOP)\src\analyze.c \ $(TOP)\src\attach.c \ |
︙ | ︙ | |||
778 779 780 781 782 783 784 | libsqlite3.lib: $(LIBOBJ) $(LTLIB) $(LTLIBOPTS) /OUT:$@ $(LIBOBJ) $(TLIBS) libtclsqlite3.lib: tclsqlite.lo libsqlite3.lib $(LTLIB) $(LTLIBOPTS) $(LTLIBPATHS) /OUT:$@ tclsqlite.lo libsqlite3.lib $(LIBTCL:tcl=tclstub) $(TLIBS) | | | | 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 | libsqlite3.lib: $(LIBOBJ) $(LTLIB) $(LTLIBOPTS) /OUT:$@ $(LIBOBJ) $(TLIBS) libtclsqlite3.lib: tclsqlite.lo libsqlite3.lib $(LTLIB) $(LTLIBOPTS) $(LTLIBPATHS) /OUT:$@ tclsqlite.lo libsqlite3.lib $(LIBTCL:tcl=tclstub) $(TLIBS) sqlite3.exe: $(TOP)\src\shell.c libsqlite3.lib $(LIBRESOBJS) sqlite3.h $(LTLINK) $(READLINE_FLAGS) \ $(TOP)\src\shell.c \ /link $(LTLINKOPTS) $(LTLIBPATHS) libsqlite3.lib $(LIBRESOBJS) $(LIBREADLINE) $(LTLIBS) $(TLIBS) # This target creates a directory named "tsrc" and fills it with # copies of all of the C source code and header files needed to # build on the target system. Some of the C source code and header # files are automatically generated. This target takes care of # all that automatic generation. # |
︙ | ︙ | |||
831 832 833 834 835 836 837 | $(LTCOMPILE) -c parse.c opcodes.lo: opcodes.c $(LTCOMPILE) -c opcodes.c # Rule to build the Win32 resources object file. # | > | | > | 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 | $(LTCOMPILE) -c parse.c opcodes.lo: opcodes.c $(LTCOMPILE) -c opcodes.c # Rule to build the Win32 resources object file. # !IF $(USE_RC)!=0 $(LIBRESOBJS): $(TOP)\src\sqlite3.rc $(HDR) echo #ifndef SQLITE_RESOURCE_VERSION > sqlite3rc.h for /F %%V in ('type "$(TOP)\VERSION"') do ( \ echo #define SQLITE_RESOURCE_VERSION %%V \ | $(NAWK) "/.*/ { gsub(/[.]/,\",\");print }" >> sqlite3rc.h \ ) echo #endif >> sqlite3rc.h $(LTRCOMPILE) -fo $(LIBRESOBJS) $(TOP)\src\sqlite3.rc !ENDIF # Rules to build individual *.lo files from files in the src directory. # alter.lo: $(TOP)\src\alter.c $(HDR) $(LTCOMPILE) -c $(TOP)\src\alter.c analyze.lo: $(TOP)\src\analyze.c $(HDR) |
︙ | ︙ | |||
1055 1056 1057 1058 1059 1060 1061 | tclsqlite.lo: $(TOP)\src\tclsqlite.c $(HDR) $(LTCOMPILE) -DUSE_TCL_STUBS=1 -DBUILD_sqlite -I$(TCLINCDIR) -c $(TOP)\src\tclsqlite.c tclsqlite-shell.lo: $(TOP)\src\tclsqlite.c $(HDR) $(LTCOMPILE) -DTCLSH=1 -DBUILD_sqlite -I$(TCLINCDIR) -c $(TOP)\src\tclsqlite.c | | | | 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 | tclsqlite.lo: $(TOP)\src\tclsqlite.c $(HDR) $(LTCOMPILE) -DUSE_TCL_STUBS=1 -DBUILD_sqlite -I$(TCLINCDIR) -c $(TOP)\src\tclsqlite.c tclsqlite-shell.lo: $(TOP)\src\tclsqlite.c $(HDR) $(LTCOMPILE) -DTCLSH=1 -DBUILD_sqlite -I$(TCLINCDIR) -c $(TOP)\src\tclsqlite.c tclsqlite3.exe: tclsqlite-shell.lo libsqlite3.lib $(LIBRESOBJS) $(LD) $(LDFLAGS) $(LTLINKOPTS) $(LTLIBPATHS) /OUT:$@ libsqlite3.lib tclsqlite-shell.lo $(LIBRESOBJS) $(LTLIBS) $(TLIBS) # Rules to build opcodes.c and opcodes.h # opcodes.c: opcodes.h $(TOP)\mkopcodec.awk $(NAWK) -f $(TOP)\mkopcodec.awk opcodes.h > opcodes.c opcodes.h: parse.h $(TOP)\src\vdbe.c $(TOP)\mkopcodeh.awk |
︙ | ︙ | |||
1169 1170 1171 1172 1173 1174 1175 | TESTFIXTURE_SRC1 = $(TESTSRC3) sqlite3.c !IF $(USE_AMALGAMATION)==0 TESTFIXTURE_SRC = $(TESTSRC) $(TOP)\src\tclsqlite.c $(TESTFIXTURE_SRC0) !ELSE TESTFIXTURE_SRC = $(TESTSRC) $(TOP)\src\tclsqlite.c $(TESTFIXTURE_SRC1) !ENDIF | | | | | | 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 | TESTFIXTURE_SRC1 = $(TESTSRC3) sqlite3.c !IF $(USE_AMALGAMATION)==0 TESTFIXTURE_SRC = $(TESTSRC) $(TOP)\src\tclsqlite.c $(TESTFIXTURE_SRC0) !ELSE TESTFIXTURE_SRC = $(TESTSRC) $(TOP)\src\tclsqlite.c $(TESTFIXTURE_SRC1) !ENDIF testfixture.exe: $(TESTFIXTURE_SRC) $(LIBRESOBJS) $(HDR) $(LTLINK) -DSQLITE_NO_SYNC=1 $(TESTFIXTURE_FLAGS) \ -DBUILD_sqlite -I$(TCLINCDIR) \ $(TESTFIXTURE_SRC) \ /link $(LTLINKOPTS) $(LTLIBPATHS) $(LIBRESOBJS) $(LTLIBS) $(TLIBS) fulltest: testfixture.exe sqlite3.exe .\testfixture.exe $(TOP)\test\all.test soaktest: testfixture.exe sqlite3.exe .\testfixture.exe $(TOP)\test\all.test -soak=1 fulltestonly: testfixture.exe sqlite3.exe .\testfixture.exe $(TOP)\test\full.test test: testfixture.exe sqlite3.exe .\testfixture.exe $(TOP)\test\veryquick.test sqlite3_analyzer.c: sqlite3.c $(TOP)\src\test_stat.c $(TOP)\src\tclsqlite.c $(TOP)\tool\spaceanal.tcl copy sqlite3.c + $(TOP)\src\test_stat.c + $(TOP)\src\tclsqlite.c $@ echo static const char *tclsh_main_loop(void){ >> $@ echo static const char *zMainloop = >> $@ $(NAWK) -f $(TOP)\tool\tostr.awk $(TOP)\tool\spaceanal.tcl >> $@ echo ; return zMainloop; } >> $@ sqlite3_analyzer.exe: sqlite3_analyzer.c $(LIBRESOBJS) $(LTLINK) -DBUILD_sqlite -DTCLSH=2 -I$(TCLINCDIR) sqlite3_analyzer.c \ /link $(LTLINKOPTS) $(LTLIBPATHS) $(LIBRESOBJS) $(LTLIBS) $(TLIBS) clean: del /Q *.lo *.ilk *.lib *.obj *.pdb sqlite3.exe libsqlite3.lib del /Q *.da *.bb *.bbg gmon.out del /Q sqlite3.h opcodes.c opcodes.h del /Q lemon.exe lempar.c parse.* del /Q mkkeywordhash.exe keywordhash.h |
︙ | ︙ | |||
1229 1230 1231 1232 1233 1234 1235 | sqlite3.def: libsqlite3.lib echo EXPORTS > sqlite3.def dumpbin /all libsqlite3.lib \ | $(NAWK) "/ 1 _?sqlite3_/ { sub(/^.* _?/,\"\");print }" \ | sort >> sqlite3.def | | | | 1247 1248 1249 1250 1251 1252 1253 1254 1255 | sqlite3.def: libsqlite3.lib echo EXPORTS > sqlite3.def dumpbin /all libsqlite3.lib \ | $(NAWK) "/ 1 _?sqlite3_/ { sub(/^.* _?/,\"\");print }" \ | sort >> sqlite3.def sqlite3.dll: $(LIBOBJ) $(LIBRESOBJS) sqlite3.def $(LD) $(LDFLAGS) $(LTLINKOPTS) $(LTLIBPATHS) /DLL /DEF:sqlite3.def /OUT:$@ $(LIBOBJ) $(LIBRESOBJS) $(LTLIBS) $(TLIBS) |
Changes to ext/rtree/rtree.c.
︙ | ︙ | |||
2656 2657 2658 2659 2660 2661 2662 | } /* ** Remove the entry with rowid=iDelete from the r-tree structure. */ static int rtreeDeleteRowid(Rtree *pRtree, sqlite3_int64 iDelete){ int rc; /* Return code */ | | | 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 | } /* ** Remove the entry with rowid=iDelete from the r-tree structure. */ static int rtreeDeleteRowid(Rtree *pRtree, sqlite3_int64 iDelete){ int rc; /* Return code */ RtreeNode *pLeaf = 0; /* Leaf node containing record iDelete */ int iCell; /* Index of iDelete cell in pLeaf */ RtreeNode *pRoot; /* Root node of rtree structure */ /* Obtain a reference to the root node to initialise Rtree.iDepth */ rc = nodeAcquire(pRtree, 1, 0, &pRoot); |
︙ | ︙ | |||
2859 2860 2861 2862 2863 2864 2865 | /* If the azData[] array contains more than one element, elements ** (azData[2]..azData[argc-1]) contain a new record to insert into ** the r-tree structure. */ if( rc==SQLITE_OK && nData>1 ){ /* Insert the new record into the r-tree */ | | | 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 | /* If the azData[] array contains more than one element, elements ** (azData[2]..azData[argc-1]) contain a new record to insert into ** the r-tree structure. */ if( rc==SQLITE_OK && nData>1 ){ /* Insert the new record into the r-tree */ RtreeNode *pLeaf = 0; /* Figure out the rowid of the new row. */ if( bHaveRowid==0 ){ rc = newRowid(pRtree, &cell.iRowid); } *pRowid = cell.iRowid; |
︙ | ︙ |
Changes to src/alter.c.
︙ | ︙ | |||
410 411 412 413 414 415 416 | int savedDbFlags; /* Saved value of db->flags */ savedDbFlags = db->flags; if( NEVER(db->mallocFailed) ) goto exit_rename_table; assert( pSrc->nSrc==1 ); assert( sqlite3BtreeHoldsAllMutexes(pParse->db) ); | | | 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 | int savedDbFlags; /* Saved value of db->flags */ savedDbFlags = db->flags; if( NEVER(db->mallocFailed) ) goto exit_rename_table; assert( pSrc->nSrc==1 ); assert( sqlite3BtreeHoldsAllMutexes(pParse->db) ); pTab = sqlite3LocateTableItem(pParse, 0, &pSrc->a[0]); if( !pTab ) goto exit_rename_table; iDb = sqlite3SchemaToIndex(pParse->db, pTab->pSchema); zDb = db->aDb[iDb].zName; db->flags |= SQLITE_PreferBuiltin; /* Get a NULL terminated version of the new table name. */ zName = sqlite3NameFromToken(db, pName); |
︙ | ︙ | |||
753 754 755 756 757 758 759 | int nAlloc; sqlite3 *db = pParse->db; /* Look up the table being altered. */ assert( pParse->pNewTable==0 ); assert( sqlite3BtreeHoldsAllMutexes(db) ); if( db->mallocFailed ) goto exit_begin_add_column; | | | 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 | int nAlloc; sqlite3 *db = pParse->db; /* Look up the table being altered. */ assert( pParse->pNewTable==0 ); assert( sqlite3BtreeHoldsAllMutexes(db) ); if( db->mallocFailed ) goto exit_begin_add_column; pTab = sqlite3LocateTableItem(pParse, 0, &pSrc->a[0]); if( !pTab ) goto exit_begin_add_column; #ifndef SQLITE_OMIT_VIRTUALTABLE if( IsVirtual(pTab) ){ sqlite3ErrorMsg(pParse, "virtual tables may not be altered"); goto exit_begin_add_column; } |
︙ | ︙ |
Changes to src/attach.c.
︙ | ︙ | |||
430 431 432 433 434 435 436 437 438 439 440 441 442 443 | sqlite3 *db; if( NEVER(iDb<0) || iDb==1 ) return 0; db = pParse->db; assert( db->nDb>iDb ); pFix->pParse = pParse; pFix->zDb = db->aDb[iDb].zName; pFix->zType = zType; pFix->pName = pName; return 1; } /* ** The following set of routines walk through the parse tree and assign | > | 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 | sqlite3 *db; if( NEVER(iDb<0) || iDb==1 ) return 0; db = pParse->db; assert( db->nDb>iDb ); pFix->pParse = pParse; pFix->zDb = db->aDb[iDb].zName; pFix->pSchema = db->aDb[iDb].pSchema; pFix->zType = zType; pFix->pName = pName; return 1; } /* ** The following set of routines walk through the parse tree and assign |
︙ | ︙ | |||
460 461 462 463 464 465 466 | int i; const char *zDb; struct SrcList_item *pItem; if( NEVER(pList==0) ) return 0; zDb = pFix->zDb; for(i=0, pItem=pList->a; i<pList->nSrc; i++, pItem++){ | | < < > > > | 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 | int i; const char *zDb; struct SrcList_item *pItem; if( NEVER(pList==0) ) return 0; zDb = pFix->zDb; for(i=0, pItem=pList->a; i<pList->nSrc; i++, pItem++){ if( pItem->zDatabase && sqlite3StrICmp(pItem->zDatabase, zDb) ){ sqlite3ErrorMsg(pFix->pParse, "%s %T cannot reference objects in database %s", pFix->zType, pFix->pName, pItem->zDatabase); return 1; } sqlite3_free(pItem->zDatabase); pItem->zDatabase = 0; pItem->pSchema = pFix->pSchema; #if !defined(SQLITE_OMIT_VIEW) || !defined(SQLITE_OMIT_TRIGGER) if( sqlite3FixSelect(pFix, pItem->pSelect) ) return 1; if( sqlite3FixExpr(pFix, pItem->pOn) ) return 1; #endif } return 0; } |
︙ | ︙ |
Changes to src/backup.c.
︙ | ︙ | |||
215 216 217 218 219 220 221 | static int backupOnePage(sqlite3_backup *p, Pgno iSrcPg, const u8 *zSrcData){ Pager * const pDestPager = sqlite3BtreePager(p->pDest); const int nSrcPgsz = sqlite3BtreeGetPageSize(p->pSrc); int nDestPgsz = sqlite3BtreeGetPageSize(p->pDest); const int nCopy = MIN(nSrcPgsz, nDestPgsz); const i64 iEnd = (i64)iSrcPg*(i64)nSrcPgsz; #ifdef SQLITE_HAS_CODEC | > > > | < > | 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 | static int backupOnePage(sqlite3_backup *p, Pgno iSrcPg, const u8 *zSrcData){ Pager * const pDestPager = sqlite3BtreePager(p->pDest); const int nSrcPgsz = sqlite3BtreeGetPageSize(p->pSrc); int nDestPgsz = sqlite3BtreeGetPageSize(p->pDest); const int nCopy = MIN(nSrcPgsz, nDestPgsz); const i64 iEnd = (i64)iSrcPg*(i64)nSrcPgsz; #ifdef SQLITE_HAS_CODEC /* Use BtreeGetReserveNoMutex() for the source b-tree, as although it is ** guaranteed that the shared-mutex is held by this thread, handle ** p->pSrc may not actually be the owner. */ int nSrcReserve = sqlite3BtreeGetReserveNoMutex(p->pSrc); int nDestReserve = sqlite3BtreeGetReserve(p->pDest); #endif int rc = SQLITE_OK; i64 iOff; assert( sqlite3BtreeGetReserveNoMutex(p->pSrc)>=0 ); assert( p->bDestLocked ); assert( !isFatalError(p->rc) ); assert( iSrcPg!=PENDING_BYTE_PAGE(p->pSrc->pBt) ); assert( zSrcData ); /* Catch the case where the destination is an in-memory database and the ** page sizes of the source and destination differ. |
︙ | ︙ |
Changes to src/btree.c.
︙ | ︙ | |||
2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 | /* ** Return the currently defined page size */ int sqlite3BtreeGetPageSize(Btree *p){ return p->pBt->pageSize; } #if !defined(SQLITE_OMIT_PAGER_PRAGMAS) || !defined(SQLITE_OMIT_VACUUM) /* ** Return the number of bytes of space at the end of every page that ** are intentually left unused. This is the "reserved" space that is ** sometimes used by extensions. */ | > > > > > > > > > > > > > > > > > > | 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 | /* ** Return the currently defined page size */ int sqlite3BtreeGetPageSize(Btree *p){ return p->pBt->pageSize; } #if defined(SQLITE_HAS_CODEC) || defined(SQLITE_DEBUG) /* ** This function is similar to sqlite3BtreeGetReserve(), except that it ** may only be called if it is guaranteed that the b-tree mutex is already ** held. ** ** This is useful in one special case in the backup API code where it is ** known that the shared b-tree mutex is held, but the mutex on the ** database handle that owns *p is not. In this case if sqlite3BtreeEnter() ** were to be called, it might collide with some other operation on the ** database handle that owns *p, causing undefined behaviour. */ int sqlite3BtreeGetReserveNoMutex(Btree *p){ assert( sqlite3_mutex_held(p->pBt->mutex) ); return p->pBt->pageSize - p->pBt->usableSize; } #endif /* SQLITE_HAS_CODEC || SQLITE_DEBUG */ #if !defined(SQLITE_OMIT_PAGER_PRAGMAS) || !defined(SQLITE_OMIT_VACUUM) /* ** Return the number of bytes of space at the end of every page that ** are intentually left unused. This is the "reserved" space that is ** sometimes used by extensions. */ |
︙ | ︙ | |||
5252 5253 5254 5255 5256 5257 5258 | assert( sqlite3_mutex_held(pPage->pBt->mutex) ); btreeParseCellPtr(pPage, pCell, &info); if( info.iOverflow==0 ){ return SQLITE_OK; /* No overflow pages. Return without doing anything */ } if( pCell+info.iOverflow+3 > pPage->aData+pPage->maskPage ){ | | | 5270 5271 5272 5273 5274 5275 5276 5277 5278 5279 5280 5281 5282 5283 5284 | assert( sqlite3_mutex_held(pPage->pBt->mutex) ); btreeParseCellPtr(pPage, pCell, &info); 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 */ } ovflPgno = get4byte(&pCell[info.iOverflow]); assert( pBt->usableSize > 4 ); ovflPageSize = pBt->usableSize - 4; nOvfl = (info.nPayload - info.nLocal + ovflPageSize - 1)/ovflPageSize; assert( ovflPgno==0 || nOvfl>0 ); while( nOvfl-- ){ |
︙ | ︙ | |||
5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928 5929 5930 5931 | ** size of a cell stored within an internal node is always less than 1/4 ** of the page-size, the aOvflSpace[] buffer is guaranteed to be large ** enough for all overflow cells. ** ** If aOvflSpace is set to a null pointer, this function returns ** SQLITE_NOMEM. */ static int balance_nonroot( MemPage *pParent, /* Parent page of siblings being balanced */ int iParentIdx, /* Index of "the page" in pParent */ u8 *aOvflSpace, /* page-size bytes of space for parent ovfl */ int isRoot, /* True if pParent is a root-page */ int bBulk /* True if this call is part of a bulk load */ ){ | > > > | 5936 5937 5938 5939 5940 5941 5942 5943 5944 5945 5946 5947 5948 5949 5950 5951 5952 | ** size of a cell stored within an internal node is always less than 1/4 ** of the page-size, the aOvflSpace[] buffer is guaranteed to be large ** enough for all overflow cells. ** ** If aOvflSpace is set to a null pointer, this function returns ** SQLITE_NOMEM. */ #if defined(_MSC_VER) && _MSC_VER >= 1700 && defined(_M_ARM) #pragma optimize("", off) #endif static int balance_nonroot( MemPage *pParent, /* Parent page of siblings being balanced */ int iParentIdx, /* Index of "the page" in pParent */ u8 *aOvflSpace, /* page-size bytes of space for parent ovfl */ int isRoot, /* True if pParent is a root-page */ int bBulk /* True if this call is part of a bulk load */ ){ |
︙ | ︙ | |||
6548 6549 6550 6551 6552 6553 6554 6555 6556 6557 6558 6559 6560 6561 | } for(i=0; i<nNew; i++){ releasePage(apNew[i]); } return rc; } /* ** This function is called when the root page of a b-tree structure is ** overfull (has one or more overflow pages). ** ** A new child page is allocated and the contents of the current root | > > > | 6569 6570 6571 6572 6573 6574 6575 6576 6577 6578 6579 6580 6581 6582 6583 6584 6585 | } for(i=0; i<nNew; i++){ releasePage(apNew[i]); } return rc; } #if defined(_MSC_VER) && _MSC_VER >= 1700 && defined(_M_ARM) #pragma optimize("", on) #endif /* ** This function is called when the root page of a b-tree structure is ** overfull (has one or more overflow pages). ** ** A new child page is allocated and the contents of the current root |
︙ | ︙ |
Changes to src/btree.h.
︙ | ︙ | |||
67 68 69 70 71 72 73 74 75 76 77 78 79 80 | int sqlite3BtreeSyncDisabled(Btree*); int sqlite3BtreeSetPageSize(Btree *p, int nPagesize, int nReserve, int eFix); int sqlite3BtreeGetPageSize(Btree*); int sqlite3BtreeMaxPageCount(Btree*,int); u32 sqlite3BtreeLastPage(Btree*); int sqlite3BtreeSecureDelete(Btree*,int); int sqlite3BtreeGetReserve(Btree*); int sqlite3BtreeSetAutoVacuum(Btree *, int); int sqlite3BtreeGetAutoVacuum(Btree *); int sqlite3BtreeBeginTrans(Btree*,int); int sqlite3BtreeCommitPhaseOne(Btree*, const char *zMaster); int sqlite3BtreeCommitPhaseTwo(Btree*, int); int sqlite3BtreeCommit(Btree*); int sqlite3BtreeRollback(Btree*,int); | > > > | 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 | int sqlite3BtreeSyncDisabled(Btree*); int sqlite3BtreeSetPageSize(Btree *p, int nPagesize, int nReserve, int eFix); int sqlite3BtreeGetPageSize(Btree*); int sqlite3BtreeMaxPageCount(Btree*,int); u32 sqlite3BtreeLastPage(Btree*); int sqlite3BtreeSecureDelete(Btree*,int); int sqlite3BtreeGetReserve(Btree*); #if defined(SQLITE_HAS_CODEC) || defined(SQLITE_DEBUG) int sqlite3BtreeGetReserveNoMutex(Btree *p); #endif int sqlite3BtreeSetAutoVacuum(Btree *, int); int sqlite3BtreeGetAutoVacuum(Btree *); int sqlite3BtreeBeginTrans(Btree*,int); int sqlite3BtreeCommitPhaseOne(Btree*, const char *zMaster); int sqlite3BtreeCommitPhaseTwo(Btree*, int); int sqlite3BtreeCommit(Btree*); int sqlite3BtreeRollback(Btree*,int); |
︙ | ︙ |
Changes to src/build.c.
︙ | ︙ | |||
314 315 316 317 318 319 320 321 322 323 324 325 326 327 | }else{ sqlite3ErrorMsg(pParse, "%s: %s", zMsg, zName); } pParse->checkSchema = 1; } return p; } /* ** Locate the in-memory structure that describes ** a particular index given the name of that index ** and the name of the database that contains the index. ** Return NULL if not found. ** | > > > > > > > > > > > > > > > > > > > > > > > > > | 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 | }else{ sqlite3ErrorMsg(pParse, "%s: %s", zMsg, zName); } pParse->checkSchema = 1; } return p; } /* ** Locate the table identified by *p. ** ** This is a wrapper around sqlite3LocateTable(). The difference between ** sqlite3LocateTable() and this function is that this function restricts ** the search to schema (p->pSchema) if it is not NULL. p->pSchema may be ** non-NULL if it is part of a view or trigger program definition. See ** sqlite3FixSrcList() for details. */ Table *sqlite3LocateTableItem( Parse *pParse, int isView, struct SrcList_item *p ){ const char *zDb; assert( p->pSchema==0 || p->zDatabase==0 ); if( p->pSchema ){ int iDb = sqlite3SchemaToIndex(pParse->db, p->pSchema); zDb = pParse->db->aDb[iDb].zName; }else{ zDb = p->zDatabase; } return sqlite3LocateTable(pParse, isView, p->zName, zDb); } /* ** Locate the in-memory structure that describes ** a particular index given the name of that index ** and the name of the database that contains the index. ** Return NULL if not found. ** |
︙ | ︙ | |||
1291 1292 1293 1294 1295 1296 1297 | sqlite3 *db = pParse->db; u8 enc = ENC(db); u8 initbusy = db->init.busy; CollSeq *pColl; pColl = sqlite3FindCollSeq(db, enc, zName, initbusy); if( !initbusy && (!pColl || !pColl->xCmp) ){ | | < < < | 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 | sqlite3 *db = pParse->db; u8 enc = ENC(db); u8 initbusy = db->init.busy; CollSeq *pColl; pColl = sqlite3FindCollSeq(db, enc, zName, initbusy); if( !initbusy && (!pColl || !pColl->xCmp) ){ pColl = sqlite3GetCollSeq(pParse, enc, pColl, zName); } return pColl; } /* |
︙ | ︙ | |||
2110 2111 2112 2113 2114 2115 2116 | if( db->mallocFailed ){ goto exit_drop_table; } assert( pParse->nErr==0 ); assert( pName->nSrc==1 ); if( noErr ) db->suppressErr++; | | < | 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 | if( db->mallocFailed ){ goto exit_drop_table; } assert( pParse->nErr==0 ); assert( pName->nSrc==1 ); if( noErr ) db->suppressErr++; pTab = sqlite3LocateTableItem(pParse, isView, &pName->a[0]); if( noErr ) db->suppressErr--; if( pTab==0 ){ if( noErr ) sqlite3CodeVerifyNamedSchema(pParse, pName->a[0].zDatabase); goto exit_drop_table; } iDb = sqlite3SchemaToIndex(db, pTab->pSchema); |
︙ | ︙ | |||
2551 2552 2553 2554 2555 2556 2557 | if( sqlite3FixInit(&sFix, pParse, iDb, "index", pName) && sqlite3FixSrcList(&sFix, pTblName) ){ /* Because the parser constructs pTblName from a single identifier, ** sqlite3FixSrcList can never fail. */ assert(0); } | | | | | 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 | if( sqlite3FixInit(&sFix, pParse, iDb, "index", pName) && sqlite3FixSrcList(&sFix, pTblName) ){ /* Because the parser constructs pTblName from a single identifier, ** sqlite3FixSrcList can never fail. */ assert(0); } pTab = sqlite3LocateTableItem(pParse, 0, &pTblName->a[0]); assert( db->mallocFailed==0 || pTab==0 ); if( pTab==0 ) goto exit_create_index; assert( db->aDb[iDb].pSchema==pTab->pSchema ); }else{ assert( pName==0 ); assert( pStart==0 ); pTab = pParse->pNewTable; if( !pTab ) goto exit_create_index; iDb = sqlite3SchemaToIndex(db, pTab->pSchema); |
︙ | ︙ |
Changes to src/callback.c.
︙ | ︙ | |||
71 72 73 74 75 76 77 | ** requested collation sequence is not available in the desired encoding. ** ** If it is not NULL, then pColl must point to the database native encoding ** collation sequence with name zName, length nName. ** ** The return value is either the collation sequence to be used in database ** db for collation type name zName, length nName, or NULL, if no collation | | | > > > > | < < | 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 | ** requested collation sequence is not available in the desired encoding. ** ** If it is not NULL, then pColl must point to the database native encoding ** collation sequence with name zName, length nName. ** ** The return value is either the collation sequence to be used in database ** db for collation type name zName, length nName, or NULL, if no collation ** sequence can be found. If no collation is found, leave an error message. ** ** See also: sqlite3LocateCollSeq(), sqlite3FindCollSeq() */ CollSeq *sqlite3GetCollSeq( Parse *pParse, /* Parsing context */ u8 enc, /* The desired encoding for the collating sequence */ CollSeq *pColl, /* Collating sequence with native encoding, or NULL */ const char *zName /* Collating sequence name */ ){ CollSeq *p; sqlite3 *db = pParse->db; p = pColl; if( !p ){ p = sqlite3FindCollSeq(db, enc, zName, 0); } if( !p || !p->xCmp ){ /* No collation sequence of this type for this encoding is registered. ** Call the collation factory to see if it can supply us with one. */ callCollNeeded(db, enc, zName); p = sqlite3FindCollSeq(db, enc, zName, 0); } if( p && !p->xCmp && synthCollSeq(db, p) ){ p = 0; } assert( !p || p->xCmp ); if( p==0 ){ sqlite3ErrorMsg(pParse, "no such collation sequence: %s", zName); } return p; } /* ** This routine is called on a collation sequence before it is used to ** check that it is defined. An undefined collation sequence exists when ** a database is loaded that contains references to collation sequences ** that have not been defined by sqlite3_create_collation() etc. ** ** If required, this routine calls the 'collation needed' callback to ** request a definition of the collating sequence. If this doesn't work, ** an equivalent collating sequence that uses a text encoding different ** from the main database is substituted, if one is available. */ int sqlite3CheckCollSeq(Parse *pParse, CollSeq *pColl){ if( pColl ){ const char *zName = pColl->zName; sqlite3 *db = pParse->db; CollSeq *p = sqlite3GetCollSeq(pParse, ENC(db), pColl, zName); if( !p ){ return SQLITE_ERROR; } assert( p==pColl ); } return SQLITE_OK; } |
︙ | ︙ |
Changes to src/delete.c.
︙ | ︙ | |||
28 29 30 31 32 33 34 | ** pSrc->a[0].pIndex Pointer to the INDEXED BY index, if there is one ** */ Table *sqlite3SrcListLookup(Parse *pParse, SrcList *pSrc){ struct SrcList_item *pItem = pSrc->a; Table *pTab; assert( pItem && pSrc->nSrc==1 ); | | | 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 | ** pSrc->a[0].pIndex Pointer to the INDEXED BY index, if there is one ** */ Table *sqlite3SrcListLookup(Parse *pParse, SrcList *pSrc){ struct SrcList_item *pItem = pSrc->a; Table *pTab; assert( pItem && pSrc->nSrc==1 ); pTab = sqlite3LocateTableItem(pParse, 0, pItem); sqlite3DeleteTable(pParse->db, pItem->pTab); pItem->pTab = pTab; if( pTab ){ pTab->nRef++; } if( sqlite3IndexedByLookup(pParse, pItem) ){ pTab = 0; |
︙ | ︙ |
Changes to src/expr.c.
︙ | ︙ | |||
926 927 928 929 930 931 932 933 934 935 936 937 938 939 | pNew = sqlite3DbMallocRaw(db, nByte ); if( pNew==0 ) return 0; pNew->nSrc = pNew->nAlloc = p->nSrc; for(i=0; i<p->nSrc; i++){ struct SrcList_item *pNewItem = &pNew->a[i]; struct SrcList_item *pOldItem = &p->a[i]; Table *pTab; pNewItem->zDatabase = sqlite3DbStrDup(db, pOldItem->zDatabase); pNewItem->zName = sqlite3DbStrDup(db, pOldItem->zName); pNewItem->zAlias = sqlite3DbStrDup(db, pOldItem->zAlias); pNewItem->jointype = pOldItem->jointype; pNewItem->iCursor = pOldItem->iCursor; pNewItem->addrFillSub = pOldItem->addrFillSub; pNewItem->regReturn = pOldItem->regReturn; | > | 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 | pNew = sqlite3DbMallocRaw(db, nByte ); if( pNew==0 ) return 0; pNew->nSrc = pNew->nAlloc = p->nSrc; for(i=0; i<p->nSrc; i++){ struct SrcList_item *pNewItem = &pNew->a[i]; struct SrcList_item *pOldItem = &p->a[i]; Table *pTab; pNewItem->pSchema = pOldItem->pSchema; pNewItem->zDatabase = sqlite3DbStrDup(db, pOldItem->zDatabase); pNewItem->zName = sqlite3DbStrDup(db, pOldItem->zName); pNewItem->zAlias = sqlite3DbStrDup(db, pOldItem->zAlias); pNewItem->jointype = pOldItem->jointype; pNewItem->iCursor = pOldItem->iCursor; pNewItem->addrFillSub = pOldItem->addrFillSub; pNewItem->regReturn = pOldItem->regReturn; |
︙ | ︙ |
Changes to src/insert.c.
︙ | ︙ | |||
1270 1271 1272 1273 1274 1275 1276 | #ifndef SQLITE_OMIT_CHECK if( pTab->pCheck && (db->flags & SQLITE_IgnoreChecks)==0 ){ ExprList *pCheck = pTab->pCheck; pParse->ckBase = regData; onError = overrideError!=OE_Default ? overrideError : OE_Abort; for(i=0; i<pCheck->nExpr; i++){ int allOk = sqlite3VdbeMakeLabel(v); | > > > | | | | | | | | | | | | | | > > | 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 | #ifndef SQLITE_OMIT_CHECK if( pTab->pCheck && (db->flags & SQLITE_IgnoreChecks)==0 ){ ExprList *pCheck = pTab->pCheck; pParse->ckBase = regData; onError = overrideError!=OE_Default ? overrideError : OE_Abort; for(i=0; i<pCheck->nExpr; i++){ int allOk = sqlite3VdbeMakeLabel(v); Expr *pDup = sqlite3ExprDup(db, pCheck->a[i].pExpr, 0); if( !db->mallocFailed ){ assert( pDup!=0 ); sqlite3ExprIfTrue(pParse, pDup, allOk, SQLITE_JUMPIFNULL); if( onError==OE_Ignore ){ sqlite3VdbeAddOp2(v, OP_Goto, 0, ignoreDest); }else{ char *zConsName = pCheck->a[i].zName; if( onError==OE_Replace ) onError = OE_Abort; /* IMP: R-15569-63625 */ if( zConsName ){ zConsName = sqlite3MPrintf(db, "constraint %s failed", zConsName); }else{ zConsName = 0; } sqlite3HaltConstraint(pParse, onError, zConsName, P4_DYNAMIC); } sqlite3VdbeResolveLabel(v, allOk); } sqlite3ExprDelete(db, pDup); } } #endif /* !defined(SQLITE_OMIT_CHECK) */ /* If we have an INTEGER PRIMARY KEY, make sure the primary key ** of the new record does not previously exist. Except, if this ** is an UPDATE and the primary key is not changing, that is OK. |
︙ | ︙ | |||
1749 1750 1751 1752 1753 1754 1755 | } /* At this point we have established that the statement is of the ** correct syntactic form to participate in this optimization. Now ** we have to check the semantics. */ pItem = pSelect->pSrc->a; | | | 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 | } /* At this point we have established that the statement is of the ** correct syntactic form to participate in this optimization. Now ** we have to check the semantics. */ pItem = pSelect->pSrc->a; pSrc = sqlite3LocateTableItem(pParse, 0, pItem); if( pSrc==0 ){ return 0; /* FROM clause does not contain a real table */ } if( pSrc==pDest ){ return 0; /* tab1 and tab2 may not be the same table */ } #ifndef SQLITE_OMIT_VIRTUALTABLE |
︙ | ︙ |
Changes to src/main.c.
︙ | ︙ | |||
3036 3037 3038 3039 3040 3041 3042 | ** operation N should be 0. The idea is that a test program (like the ** SQL Logic Test or SLT test module) can run the same SQL multiple times ** with various optimizations disabled to verify that the same answer ** is obtained in every case. */ case SQLITE_TESTCTRL_OPTIMIZATIONS: { sqlite3 *db = va_arg(ap, sqlite3*); | | | 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 | ** operation N should be 0. The idea is that a test program (like the ** SQL Logic Test or SLT test module) can run the same SQL multiple times ** with various optimizations disabled to verify that the same answer ** is obtained in every case. */ case SQLITE_TESTCTRL_OPTIMIZATIONS: { sqlite3 *db = va_arg(ap, sqlite3*); db->dbOptFlags = (u8)(va_arg(ap, int) & 0xff); break; } #ifdef SQLITE_N_KEYWORD /* sqlite3_test_control(SQLITE_TESTCTRL_ISKEYWORD, const char *zWord) ** ** If zWord is a keyword recognized by the parser, then return the |
︙ | ︙ |
Changes to src/os_unix.c.
︙ | ︙ | |||
3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 | static int seekAndRead(unixFile *id, sqlite3_int64 offset, void *pBuf, int cnt){ int got; int prior = 0; #if (!defined(USE_PREAD) && !defined(USE_PREAD64)) i64 newOffset; #endif TIMER_START; do{ #if defined(USE_PREAD) got = osPread(id->h, pBuf, cnt, offset); SimulateIOError( got = -1 ); #elif defined(USE_PREAD64) got = osPread64(id->h, pBuf, cnt, offset); SimulateIOError( got = -1 ); | > > | 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 | static int seekAndRead(unixFile *id, sqlite3_int64 offset, void *pBuf, int cnt){ int got; int prior = 0; #if (!defined(USE_PREAD) && !defined(USE_PREAD64)) i64 newOffset; #endif TIMER_START; assert( cnt==(cnt&0x1ffff) ); cnt &= 0x1ffff; do{ #if defined(USE_PREAD) got = osPread(id->h, pBuf, cnt, offset); SimulateIOError( got = -1 ); #elif defined(USE_PREAD64) got = osPread64(id->h, pBuf, cnt, offset); SimulateIOError( got = -1 ); |
︙ | ︙ | |||
3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 | ** is set before returning. */ static int seekAndWrite(unixFile *id, i64 offset, const void *pBuf, int cnt){ int got; #if (!defined(USE_PREAD) && !defined(USE_PREAD64)) i64 newOffset; #endif TIMER_START; #if defined(USE_PREAD) do{ got = osPwrite(id->h, pBuf, cnt, offset); }while( got<0 && errno==EINTR ); #elif defined(USE_PREAD64) do{ got = osPwrite64(id->h, pBuf, cnt, offset);}while( got<0 && errno==EINTR); #else do{ | > > | 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 | ** is set before returning. */ static int seekAndWrite(unixFile *id, i64 offset, const void *pBuf, int cnt){ int got; #if (!defined(USE_PREAD) && !defined(USE_PREAD64)) i64 newOffset; #endif assert( cnt==(cnt&0x1ffff) ); cnt &= 0x1ffff; TIMER_START; #if defined(USE_PREAD) do{ got = osPwrite(id->h, pBuf, cnt, offset); }while( got<0 && errno==EINTR ); #elif defined(USE_PREAD64) do{ got = osPwrite64(id->h, pBuf, cnt, offset);}while( got<0 && errno==EINTR); #else do{ |
︙ | ︙ |
Changes to src/os_win.c.
︙ | ︙ | |||
29 30 31 32 33 34 35 36 37 38 39 40 41 42 | ** available in Windows platforms based on the NT kernel. */ #if !SQLITE_OS_WINNT && !defined(SQLITE_OMIT_WAL) # error "WAL mode requires support from the Windows NT kernel, compile\ with SQLITE_OMIT_WAL." #endif /* ** Macro to find the minimum of two numeric values. */ #ifndef MIN # define MIN(x,y) ((x)<(y)?(x):(y)) #endif | > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > | 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 | ** available in Windows platforms based on the NT kernel. */ #if !SQLITE_OS_WINNT && !defined(SQLITE_OMIT_WAL) # error "WAL mode requires support from the Windows NT kernel, compile\ with SQLITE_OMIT_WAL." #endif /* ** Are most of the Win32 ANSI APIs available (i.e. with certain exceptions ** based on the sub-platform)? */ #if !SQLITE_OS_WINCE && !SQLITE_OS_WINRT # define SQLITE_WIN32_HAS_ANSI #endif /* ** Are most of the Win32 Unicode APIs available (i.e. with certain exceptions ** based on the sub-platform)? */ #if SQLITE_OS_WINCE || SQLITE_OS_WINNT || SQLITE_OS_WINRT # define SQLITE_WIN32_HAS_WIDE #endif /* ** Do we need to manually define the Win32 file mapping APIs for use with WAL ** mode (e.g. these APIs are available in the Windows CE SDK; however, they ** are not present in the header file)? */ #if SQLITE_WIN32_FILEMAPPING_API && !defined(SQLITE_OMIT_WAL) /* ** Two of the file mapping APIs are different under WinRT. Figure out which ** set we need. */ #if SQLITE_OS_WINRT WINBASEAPI HANDLE WINAPI CreateFileMappingFromApp(HANDLE, \ LPSECURITY_ATTRIBUTES, ULONG, ULONG64, LPCWSTR); WINBASEAPI LPVOID WINAPI MapViewOfFileFromApp(HANDLE, ULONG, ULONG64, SIZE_T); #else #if defined(SQLITE_WIN32_HAS_ANSI) WINBASEAPI HANDLE WINAPI CreateFileMappingA(HANDLE, LPSECURITY_ATTRIBUTES, \ DWORD, DWORD, DWORD, LPCSTR); #endif /* defined(SQLITE_WIN32_HAS_ANSI) */ #if defined(SQLITE_WIN32_HAS_WIDE) WINBASEAPI HANDLE WINAPI CreateFileMappingW(HANDLE, LPSECURITY_ATTRIBUTES, \ DWORD, DWORD, DWORD, LPCWSTR); #endif /* defined(SQLITE_WIN32_HAS_WIDE) */ WINBASEAPI LPVOID WINAPI MapViewOfFile(HANDLE, DWORD, DWORD, DWORD, SIZE_T); #endif /* SQLITE_OS_WINRT */ /* ** This file mapping API is common to both Win32 and WinRT. */ WINBASEAPI BOOL WINAPI UnmapViewOfFile(LPCVOID); #endif /* SQLITE_WIN32_FILEMAPPING_API && !defined(SQLITE_OMIT_WAL) */ /* ** Macro to find the minimum of two numeric values. */ #ifndef MIN # define MIN(x,y) ((x)<(y)?(x):(y)) #endif |
︙ | ︙ | |||
234 235 236 237 238 239 240 | */ #ifdef SQLITE_TEST int sqlite3_os_type = 0; #else static int sqlite3_os_type = 0; #endif | < < < < < < < < | 285 286 287 288 289 290 291 292 293 294 295 296 297 298 | */ #ifdef SQLITE_TEST int sqlite3_os_type = 0; #else static int sqlite3_os_type = 0; #endif #ifndef SYSCALL # define SYSCALL sqlite3_syscall_ptr #endif /* ** This function is not available on Windows CE or WinRT. */ |
︙ | ︙ | |||
398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 | #else { "FormatMessageW", (SYSCALL)0, 0 }, #endif #define osFormatMessageW ((DWORD(WINAPI*)(DWORD,LPCVOID,DWORD,DWORD,LPWSTR, \ DWORD,va_list*))aSyscall[15].pCurrent) { "FreeLibrary", (SYSCALL)FreeLibrary, 0 }, #define osFreeLibrary ((BOOL(WINAPI*)(HMODULE))aSyscall[16].pCurrent) { "GetCurrentProcessId", (SYSCALL)GetCurrentProcessId, 0 }, #define osGetCurrentProcessId ((DWORD(WINAPI*)(VOID))aSyscall[17].pCurrent) | > > > > | 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 | #else { "FormatMessageW", (SYSCALL)0, 0 }, #endif #define osFormatMessageW ((DWORD(WINAPI*)(DWORD,LPCVOID,DWORD,DWORD,LPWSTR, \ DWORD,va_list*))aSyscall[15].pCurrent) #if !defined(SQLITE_OMIT_LOAD_EXTENSION) { "FreeLibrary", (SYSCALL)FreeLibrary, 0 }, #else { "FreeLibrary", (SYSCALL)0, 0 }, #endif #define osFreeLibrary ((BOOL(WINAPI*)(HMODULE))aSyscall[16].pCurrent) { "GetCurrentProcessId", (SYSCALL)GetCurrentProcessId, 0 }, #define osGetCurrentProcessId ((DWORD(WINAPI*)(VOID))aSyscall[17].pCurrent) |
︙ | ︙ | |||
479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 | #define osGetFullPathNameW ((DWORD(WINAPI*)(LPCWSTR,DWORD,LPWSTR, \ LPWSTR*))aSyscall[25].pCurrent) { "GetLastError", (SYSCALL)GetLastError, 0 }, #define osGetLastError ((DWORD(WINAPI*)(VOID))aSyscall[26].pCurrent) #if SQLITE_OS_WINCE /* The GetProcAddressA() routine is only available on Windows CE. */ { "GetProcAddressA", (SYSCALL)GetProcAddressA, 0 }, #else /* All other Windows platforms expect GetProcAddress() to take ** an ANSI string regardless of the _UNICODE setting */ { "GetProcAddressA", (SYSCALL)GetProcAddress, 0 }, #endif #define osGetProcAddressA ((FARPROC(WINAPI*)(HMODULE, \ LPCSTR))aSyscall[27].pCurrent) #if !SQLITE_OS_WINRT { "GetSystemInfo", (SYSCALL)GetSystemInfo, 0 }, #else | > > > > | 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 | #define osGetFullPathNameW ((DWORD(WINAPI*)(LPCWSTR,DWORD,LPWSTR, \ LPWSTR*))aSyscall[25].pCurrent) { "GetLastError", (SYSCALL)GetLastError, 0 }, #define osGetLastError ((DWORD(WINAPI*)(VOID))aSyscall[26].pCurrent) #if !defined(SQLITE_OMIT_LOAD_EXTENSION) #if SQLITE_OS_WINCE /* The GetProcAddressA() routine is only available on Windows CE. */ { "GetProcAddressA", (SYSCALL)GetProcAddressA, 0 }, #else /* All other Windows platforms expect GetProcAddress() to take ** an ANSI string regardless of the _UNICODE setting */ { "GetProcAddressA", (SYSCALL)GetProcAddress, 0 }, #endif #else { "GetProcAddressA", (SYSCALL)0, 0 }, #endif #define osGetProcAddressA ((FARPROC(WINAPI*)(HMODULE, \ LPCSTR))aSyscall[27].pCurrent) #if !SQLITE_OS_WINRT { "GetSystemInfo", (SYSCALL)GetSystemInfo, 0 }, #else |
︙ | ︙ | |||
590 591 592 593 594 595 596 | #else { "HeapValidate", (SYSCALL)0, 0 }, #endif #define osHeapValidate ((BOOL(WINAPI*)(HANDLE,DWORD, \ LPCVOID))aSyscall[41].pCurrent) | | | > | 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 | #else { "HeapValidate", (SYSCALL)0, 0 }, #endif #define osHeapValidate ((BOOL(WINAPI*)(HANDLE,DWORD, \ LPCVOID))aSyscall[41].pCurrent) #if defined(SQLITE_WIN32_HAS_ANSI) && !defined(SQLITE_OMIT_LOAD_EXTENSION) { "LoadLibraryA", (SYSCALL)LoadLibraryA, 0 }, #else { "LoadLibraryA", (SYSCALL)0, 0 }, #endif #define osLoadLibraryA ((HMODULE(WINAPI*)(LPCSTR))aSyscall[42].pCurrent) #if !SQLITE_OS_WINRT && defined(SQLITE_WIN32_HAS_WIDE) && \ !defined(SQLITE_OMIT_LOAD_EXTENSION) { "LoadLibraryW", (SYSCALL)LoadLibraryW, 0 }, #else { "LoadLibraryW", (SYSCALL)0, 0 }, #endif #define osLoadLibraryW ((HMODULE(WINAPI*)(LPCWSTR))aSyscall[43].pCurrent) |
︙ | ︙ | |||
787 788 789 790 791 792 793 | #else { "CreateFile2", (SYSCALL)0, 0 }, #endif #define osCreateFile2 ((HANDLE(WINAPI*)(LPCWSTR,DWORD,DWORD,DWORD, \ LPCREATEFILE2_EXTENDED_PARAMETERS))aSyscall[66].pCurrent) | | | 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 | #else { "CreateFile2", (SYSCALL)0, 0 }, #endif #define osCreateFile2 ((HANDLE(WINAPI*)(LPCWSTR,DWORD,DWORD,DWORD, \ LPCREATEFILE2_EXTENDED_PARAMETERS))aSyscall[66].pCurrent) #if SQLITE_OS_WINRT && !defined(SQLITE_OMIT_LOAD_EXTENSION) { "LoadPackagedLibrary", (SYSCALL)LoadPackagedLibrary, 0 }, #else { "LoadPackagedLibrary", (SYSCALL)0, 0 }, #endif #define osLoadPackagedLibrary ((HMODULE(WINAPI*)(LPCWSTR, \ DWORD))aSyscall[67].pCurrent) |
︙ | ︙ |
Changes to src/pager.c.
︙ | ︙ | |||
2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 | if( rc==SQLITE_OK ){ pPager->dbFileSize = nPage; } } } return rc; } /* ** 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 used ** to determine the size and alignment of journal header and ** master journal pointers within created journal files. | > > > > > > > > > > > > > > > | 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 | if( rc==SQLITE_OK ){ pPager->dbFileSize = nPage; } } } return rc; } /* ** Return a sanitized version of the sector-size of OS file pFile. The ** return value is guaranteed to lie between 32 and MAX_SECTOR_SIZE. */ int sqlite3SectorSize(sqlite3_file *pFile){ int iRet = sqlite3OsSectorSize(pFile); if( iRet<32 ){ iRet = 512; }else if( iRet>MAX_SECTOR_SIZE ){ assert( MAX_SECTOR_SIZE>=512 ); iRet = MAX_SECTOR_SIZE; } return iRet; } /* ** 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 used ** to determine the size and alignment of journal header and ** master journal pointers within created journal files. |
︙ | ︙ | |||
2540 2541 2542 2543 2544 2545 2546 | SQLITE_IOCAP_POWERSAFE_OVERWRITE)!=0 ){ /* Sector size doesn't matter for temporary files. Also, the file ** may not have been opened yet, in which case the OsSectorSize() ** call will segfault. */ pPager->sectorSize = 512; }else{ | < < | < < < < < | 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 | SQLITE_IOCAP_POWERSAFE_OVERWRITE)!=0 ){ /* Sector size doesn't matter for temporary files. Also, the file ** may not have been opened yet, in which case the OsSectorSize() ** call will segfault. */ pPager->sectorSize = 512; }else{ pPager->sectorSize = sqlite3SectorSize(pPager->fd); } } /* ** Playback the journal and thus restore the database file to ** the state it was in before we started making changes. ** |
︙ | ︙ | |||
3464 3465 3466 3467 3468 3469 3470 | ** retried. If it returns zero, then the SQLITE_BUSY error is ** returned to the caller of the pager API function. */ void sqlite3PagerSetBusyhandler( Pager *pPager, /* Pager object */ int (*xBusyHandler)(void *), /* Pointer to busy-handler function */ void *pBusyHandlerArg /* Argument to pass to xBusyHandler */ | | > > > > > > > | 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 | ** retried. If it returns zero, then the SQLITE_BUSY error is ** returned to the caller of the pager API function. */ void sqlite3PagerSetBusyhandler( Pager *pPager, /* Pager object */ int (*xBusyHandler)(void *), /* Pointer to busy-handler function */ void *pBusyHandlerArg /* Argument to pass to xBusyHandler */ ){ pPager->xBusyHandler = xBusyHandler; pPager->pBusyHandlerArg = pBusyHandlerArg; if( isOpen(pPager->fd) ){ void **ap = (void **)&pPager->xBusyHandler; assert( ((int(*)(void *))(ap[0]))==xBusyHandler ); assert( ap[1]==pBusyHandlerArg ); sqlite3OsFileControl(pPager->fd, SQLITE_FCNTL_BUSYHANDLER, (void *)ap); } } /* ** Change the page size used by the Pager object. The new page size ** is passed in *pPageSize. ** ** If the pager is in the error state when this function is called, it |
︙ | ︙ |
Changes to src/pager.h.
︙ | ︙ | |||
156 157 158 159 160 161 162 163 164 165 166 167 168 169 | sqlite3_file *sqlite3PagerFile(Pager*); const char *sqlite3PagerJournalname(Pager*); int sqlite3PagerNosync(Pager*); void *sqlite3PagerTempSpace(Pager*); int sqlite3PagerIsMemdb(Pager*); void sqlite3PagerCacheStat(Pager *, int, int, int *); void sqlite3PagerClearCache(Pager *); /* Functions used to truncate the database file. */ void sqlite3PagerTruncateImage(Pager*,Pgno); #if defined(SQLITE_HAS_CODEC) && !defined(SQLITE_OMIT_WAL) void *sqlite3PagerCodec(DbPage *); #endif | > | 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 | sqlite3_file *sqlite3PagerFile(Pager*); const char *sqlite3PagerJournalname(Pager*); int sqlite3PagerNosync(Pager*); void *sqlite3PagerTempSpace(Pager*); int sqlite3PagerIsMemdb(Pager*); void sqlite3PagerCacheStat(Pager *, int, int, int *); void sqlite3PagerClearCache(Pager *); int sqlite3SectorSize(sqlite3_file *); /* Functions used to truncate the database file. */ void sqlite3PagerTruncateImage(Pager*,Pgno); #if defined(SQLITE_HAS_CODEC) && !defined(SQLITE_OMIT_WAL) void *sqlite3PagerCodec(DbPage *); #endif |
︙ | ︙ |
Changes to src/pragma.c.
︙ | ︙ | |||
353 354 355 356 357 358 359 360 361 362 363 364 365 366 | ** connection. If it returns SQLITE_OK, then assume that the VFS ** handled the pragma and generate a no-op prepared statement. */ aFcntl[0] = 0; aFcntl[1] = zLeft; aFcntl[2] = zRight; aFcntl[3] = 0; rc = sqlite3_file_control(db, zDb, SQLITE_FCNTL_PRAGMA, (void*)aFcntl); if( rc==SQLITE_OK ){ if( aFcntl[0] ){ int mem = ++pParse->nMem; sqlite3VdbeAddOp4(v, OP_String8, 0, mem, 0, aFcntl[0], 0); sqlite3VdbeSetNumCols(v, 1); sqlite3VdbeSetColName(v, 0, COLNAME_NAME, "result", SQLITE_STATIC); | > | 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 | ** connection. If it returns SQLITE_OK, then assume that the VFS ** handled the pragma and generate a no-op prepared statement. */ aFcntl[0] = 0; aFcntl[1] = zLeft; aFcntl[2] = zRight; aFcntl[3] = 0; db->busyHandler.nBusy = 0; rc = sqlite3_file_control(db, zDb, SQLITE_FCNTL_PRAGMA, (void*)aFcntl); if( rc==SQLITE_OK ){ if( aFcntl[0] ){ int mem = ++pParse->nMem; sqlite3VdbeAddOp4(v, OP_String8, 0, mem, 0, aFcntl[0], 0); sqlite3VdbeSetNumCols(v, 1); sqlite3VdbeSetColName(v, 0, COLNAME_NAME, "result", SQLITE_STATIC); |
︙ | ︙ |
Changes to src/select.c.
︙ | ︙ | |||
3328 3329 3330 3331 3332 3333 3334 | pTab->iPKey = -1; pTab->nRowEst = 1000000; pTab->tabFlags |= TF_Ephemeral; #endif }else{ /* An ordinary table or view name in the FROM clause */ assert( pFrom->pTab==0 ); | | < | 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 | pTab->iPKey = -1; pTab->nRowEst = 1000000; pTab->tabFlags |= TF_Ephemeral; #endif }else{ /* An ordinary table or view name in the FROM clause */ assert( pFrom->pTab==0 ); pFrom->pTab = pTab = sqlite3LocateTableItem(pParse, 0, pFrom); if( pTab==0 ) return WRC_Abort; pTab->nRef++; #if !defined(SQLITE_OMIT_VIEW) || !defined (SQLITE_OMIT_VIRTUALTABLE) 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 ); |
︙ | ︙ |
Changes to src/sqlite.h.in.
︙ | ︙ | |||
848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 | ** prepared statement. ^If the [SQLITE_FCNTL_PRAGMA] file control returns ** any result code other than [SQLITE_OK] or [SQLITE_NOTFOUND], that means ** that the VFS encountered an error while handling the [PRAGMA] and the ** compilation of the PRAGMA fails with an error. ^The [SQLITE_FCNTL_PRAGMA] ** file control occurs at the beginning of pragma statement analysis and so ** it is able to override built-in [PRAGMA] statements. ** </ul> */ #define SQLITE_FCNTL_LOCKSTATE 1 #define SQLITE_GET_LOCKPROXYFILE 2 #define SQLITE_SET_LOCKPROXYFILE 3 #define SQLITE_LAST_ERRNO 4 #define SQLITE_FCNTL_SIZE_HINT 5 #define SQLITE_FCNTL_CHUNK_SIZE 6 #define SQLITE_FCNTL_FILE_POINTER 7 #define SQLITE_FCNTL_SYNC_OMITTED 8 #define SQLITE_FCNTL_WIN32_AV_RETRY 9 #define SQLITE_FCNTL_PERSIST_WAL 10 #define SQLITE_FCNTL_OVERWRITE 11 #define SQLITE_FCNTL_VFSNAME 12 #define SQLITE_FCNTL_POWERSAFE_OVERWRITE 13 #define SQLITE_FCNTL_PRAGMA 14 /* ** CAPI3REF: Mutex Handle ** ** The mutex module within SQLite defines [sqlite3_mutex] to be an ** abstract type for a mutex object. The SQLite core never looks ** at the internal representation of an [sqlite3_mutex]. It only | > > > > > > > > > > > > | 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 | ** prepared statement. ^If the [SQLITE_FCNTL_PRAGMA] file control returns ** any result code other than [SQLITE_OK] or [SQLITE_NOTFOUND], that means ** that the VFS encountered an error while handling the [PRAGMA] and the ** compilation of the PRAGMA fails with an error. ^The [SQLITE_FCNTL_PRAGMA] ** file control occurs at the beginning of pragma statement analysis and so ** it is able to override built-in [PRAGMA] statements. ** </ul> ** ** <li>[[SQLITE_FCNTL_BUSYHANDLER]] ** ^This file-control may be invoked by SQLite on the database file handle ** shortly after it is opened in order to provide a custom VFS with access ** to the connections busy-handler callback. The argument is of type (void **) ** - an array of two (void *) values. The first (void *) actually points ** to a function of type (int (*)(void *)). In order to invoke the connections ** busy-handler, this function should be invoked with the second (void *) in ** the array as the only argument. If it returns non-zero, then the operation ** should be retried. If it returns zero, the custom VFS should abandon the ** current operation. */ #define SQLITE_FCNTL_LOCKSTATE 1 #define SQLITE_GET_LOCKPROXYFILE 2 #define SQLITE_SET_LOCKPROXYFILE 3 #define SQLITE_LAST_ERRNO 4 #define SQLITE_FCNTL_SIZE_HINT 5 #define SQLITE_FCNTL_CHUNK_SIZE 6 #define SQLITE_FCNTL_FILE_POINTER 7 #define SQLITE_FCNTL_SYNC_OMITTED 8 #define SQLITE_FCNTL_WIN32_AV_RETRY 9 #define SQLITE_FCNTL_PERSIST_WAL 10 #define SQLITE_FCNTL_OVERWRITE 11 #define SQLITE_FCNTL_VFSNAME 12 #define SQLITE_FCNTL_POWERSAFE_OVERWRITE 13 #define SQLITE_FCNTL_PRAGMA 14 #define SQLITE_FCNTL_BUSYHANDLER 15 /* ** CAPI3REF: Mutex Handle ** ** The mutex module within SQLite defines [sqlite3_mutex] to be an ** abstract type for a mutex object. The SQLite core never looks ** at the internal representation of an [sqlite3_mutex]. It only |
︙ | ︙ |
Changes to src/sqliteInt.h.
︙ | ︙ | |||
1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 | ** In the colUsed field, the high-order bit (bit 63) is set if the table ** contains more than 63 columns and the 64-th or later column is used. */ struct SrcList { i16 nSrc; /* Number of tables or subqueries in the FROM clause */ i16 nAlloc; /* Number of entries allocated in a[] below */ struct SrcList_item { char *zDatabase; /* Name of database holding this table */ char *zName; /* Name of the table */ char *zAlias; /* The "B" part of a "A AS B" phrase. zName is the "A" */ Table *pTab; /* An SQL table corresponding to zName */ Select *pSelect; /* A SELECT statement used in place of a table name */ int addrFillSub; /* Address of subroutine to manifest a subquery */ int regReturn; /* Register holding return address of addrFillSub */ | > | 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 | ** In the colUsed field, the high-order bit (bit 63) is set if the table ** contains more than 63 columns and the 64-th or later column is used. */ struct SrcList { i16 nSrc; /* Number of tables or subqueries in the FROM clause */ i16 nAlloc; /* Number of entries allocated in a[] below */ struct SrcList_item { Schema *pSchema; /* Schema to which this item is fixed */ char *zDatabase; /* Name of database holding this table */ char *zName; /* Name of the table */ char *zAlias; /* The "B" part of a "A AS B" phrase. zName is the "A" */ Table *pTab; /* An SQL table corresponding to zName */ Select *pSelect; /* A SELECT statement used in place of a table name */ int addrFillSub; /* Address of subroutine to manifest a subquery */ int regReturn; /* Register holding return address of addrFillSub */ |
︙ | ︙ | |||
2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 | ** The following structure contains information used by the sqliteFix... ** routines as they walk the parse tree to make database references ** explicit. */ typedef struct DbFixer DbFixer; struct DbFixer { Parse *pParse; /* The parsing context. Error messages written here */ const char *zDb; /* Make sure all objects are contained in this database */ const char *zType; /* Type of the container - used for error messages */ const Token *pName; /* Name of the container - used for error messages */ }; /* ** An objected used to accumulate the text of a string where we | > | 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 | ** The following structure contains information used by the sqliteFix... ** routines as they walk the parse tree to make database references ** explicit. */ typedef struct DbFixer DbFixer; struct DbFixer { Parse *pParse; /* The parsing context. Error messages written here */ Schema *pSchema; /* Fix items to this schema */ const char *zDb; /* Make sure all objects are contained in this database */ const char *zType; /* Type of the container - used for error messages */ const Token *pName; /* Name of the container - used for error messages */ }; /* ** An objected used to accumulate the text of a string where we |
︙ | ︙ | |||
2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 | int sqlite3ExprCodeAndCache(Parse*, Expr*, int); void sqlite3ExprCodeConstants(Parse*, Expr*); int sqlite3ExprCodeExprList(Parse*, ExprList*, int, int); void sqlite3ExprIfTrue(Parse*, Expr*, int, int); void sqlite3ExprIfFalse(Parse*, Expr*, int, int); Table *sqlite3FindTable(sqlite3*,const char*, const char*); Table *sqlite3LocateTable(Parse*,int isView,const char*, const char*); Index *sqlite3FindIndex(sqlite3*,const char*, const char*); void sqlite3UnlinkAndDeleteTable(sqlite3*,int,const char*); void sqlite3UnlinkAndDeleteIndex(sqlite3*,int,const char*); void sqlite3Vacuum(Parse*); int sqlite3RunVacuum(char**, sqlite3*); char *sqlite3NameFromToken(sqlite3*, Token*); int sqlite3ExprCompare(Expr*, Expr*); | > | 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 | int sqlite3ExprCodeAndCache(Parse*, Expr*, int); void sqlite3ExprCodeConstants(Parse*, Expr*); int sqlite3ExprCodeExprList(Parse*, ExprList*, int, int); void sqlite3ExprIfTrue(Parse*, Expr*, int, int); void sqlite3ExprIfFalse(Parse*, Expr*, int, int); Table *sqlite3FindTable(sqlite3*,const char*, const char*); Table *sqlite3LocateTable(Parse*,int isView,const char*, const char*); Table *sqlite3LocateTableItem(Parse*,int isView,struct SrcList_item *); Index *sqlite3FindIndex(sqlite3*,const char*, const char*); void sqlite3UnlinkAndDeleteTable(sqlite3*,int,const char*); void sqlite3UnlinkAndDeleteIndex(sqlite3*,int,const char*); void sqlite3Vacuum(Parse*); int sqlite3RunVacuum(char**, sqlite3*); char *sqlite3NameFromToken(sqlite3*, Token*); int sqlite3ExprCompare(Expr*, Expr*); |
︙ | ︙ | |||
3085 3086 3087 3088 3089 3090 3091 | void sqlite3SelectPrep(Parse*, Select*, NameContext*); int sqlite3ResolveExprNames(NameContext*, Expr*); void sqlite3ResolveSelectNames(Parse*, Select*, NameContext*); int sqlite3ResolveOrderGroupBy(Parse*, Select*, ExprList*, const char*); void sqlite3ColumnDefault(Vdbe *, Table *, int, int); void sqlite3AlterFinishAddColumn(Parse *, Token *); void sqlite3AlterBeginAddColumn(Parse *, SrcList *); | | | 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 | void sqlite3SelectPrep(Parse*, Select*, NameContext*); int sqlite3ResolveExprNames(NameContext*, Expr*); void sqlite3ResolveSelectNames(Parse*, Select*, NameContext*); int sqlite3ResolveOrderGroupBy(Parse*, Select*, ExprList*, const char*); void sqlite3ColumnDefault(Vdbe *, Table *, int, int); void sqlite3AlterFinishAddColumn(Parse *, Token *); void sqlite3AlterBeginAddColumn(Parse *, SrcList *); CollSeq *sqlite3GetCollSeq(Parse*, u8, CollSeq *, const char*); char sqlite3AffinityType(const char*); void sqlite3Analyze(Parse*, Token*, Token*); int sqlite3InvokeBusyHandler(BusyHandler*); int sqlite3FindDb(sqlite3*, Token*); int sqlite3FindDbName(sqlite3 *, const char *); int sqlite3AnalysisLoad(sqlite3*,int iDB); void sqlite3DeleteIndexSamples(sqlite3*,Index*); |
︙ | ︙ |
Changes to src/tclsqlite.c.
︙ | ︙ | |||
49 50 51 52 53 54 55 | #undef TCL_STORAGE_CLASS #define TCL_STORAGE_CLASS DLLEXPORT #endif /* BUILD_sqlite */ #define NUM_PREPARED_STMTS 10 #define MAX_PREPARED_STMTS 100 | | | < < < < < < < > | 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 | #undef TCL_STORAGE_CLASS #define TCL_STORAGE_CLASS DLLEXPORT #endif /* BUILD_sqlite */ #define NUM_PREPARED_STMTS 10 #define MAX_PREPARED_STMTS 100 /* Forward declaration */ typedef struct SqliteDb SqliteDb; /* ** New SQL functions can be created as TCL scripts. Each such function ** is described by an instance of the following structure. */ typedef struct SqlFunc SqlFunc; struct SqlFunc { Tcl_Interp *interp; /* The TCL interpret to execute the function */ Tcl_Obj *pScript; /* The Tcl_Obj representation of the script */ SqliteDb *pDb; /* Database connection that owns this function */ int useEvalObjv; /* True if it is safe to use Tcl_EvalObjv */ char *zName; /* Name of this function */ SqlFunc *pNext; /* Next function on the list of them all */ }; /* ** New collation sequences function can be created as TCL scripts. Each such |
︙ | ︙ | |||
109 110 111 112 113 114 115 | ** that has been opened by the SQLite TCL interface. ** ** If this module is built with SQLITE_TEST defined (to create the SQLite ** testfixture executable), then it may be configured to use either ** sqlite3_prepare_v2() or sqlite3_prepare() to prepare SQL statements. ** If SqliteDb.bLegacyPrepare is true, sqlite3_prepare() is used. */ | < | 103 104 105 106 107 108 109 110 111 112 113 114 115 116 | ** that has been opened by the SQLite TCL interface. ** ** If this module is built with SQLITE_TEST defined (to create the SQLite ** testfixture executable), then it may be configured to use either ** sqlite3_prepare_v2() or sqlite3_prepare() to prepare SQL statements. ** If SqliteDb.bLegacyPrepare is true, sqlite3_prepare() is used. */ struct SqliteDb { sqlite3 *db; /* The "real" database structure. MUST BE FIRST */ Tcl_Interp *interp; /* The interpreter used for this database */ char *zBusy; /* The busy callback routine */ char *zCommit; /* The commit hook callback routine */ char *zTrace; /* The trace callback routine */ char *zProfile; /* The profile callback routine */ |
︙ | ︙ | |||
428 429 430 431 432 433 434 435 436 437 438 439 440 441 | for(p=pDb->pFunc; p; p=p->pNext){ if( strcmp(p->zName, pNew->zName)==0 ){ Tcl_Free((char*)pNew); return p; } } pNew->interp = pDb->interp; pNew->pScript = 0; pNew->pNext = pDb->pFunc; pDb->pFunc = pNew; return pNew; } /* | > | 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 | for(p=pDb->pFunc; p; p=p->pNext){ if( strcmp(p->zName, pNew->zName)==0 ){ Tcl_Free((char*)pNew); return p; } } pNew->interp = pDb->interp; pNew->pDb = pDb; pNew->pScript = 0; pNew->pNext = pDb->pFunc; pDb->pFunc = pNew; return pNew; } /* |
︙ | ︙ | |||
475 476 477 478 479 480 481 482 483 484 485 486 487 488 | SqliteDb *pDb = (SqliteDb*)db; flushStmtCache(pDb); closeIncrblobChannels(pDb); sqlite3_close(pDb->db); while( pDb->pFunc ){ SqlFunc *pFunc = pDb->pFunc; pDb->pFunc = pFunc->pNext; Tcl_DecrRefCount(pFunc->pScript); Tcl_Free((char*)pFunc); } while( pDb->pCollate ){ SqlCollate *pCollate = pDb->pCollate; pDb->pCollate = pCollate->pNext; Tcl_Free((char*)pCollate); | > | 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 | SqliteDb *pDb = (SqliteDb*)db; flushStmtCache(pDb); closeIncrblobChannels(pDb); sqlite3_close(pDb->db); while( pDb->pFunc ){ SqlFunc *pFunc = pDb->pFunc; pDb->pFunc = pFunc->pNext; assert( pFunc->pDb==pDb ); Tcl_DecrRefCount(pFunc->pScript); Tcl_Free((char*)pFunc); } while( pDb->pCollate ){ SqlCollate *pCollate = pDb->pCollate; pDb->pCollate = pCollate->pNext; Tcl_Free((char*)pCollate); |
︙ | ︙ | |||
834 835 836 837 838 839 840 | } case SQLITE_FLOAT: { double r = sqlite3_value_double(pIn); pVal = Tcl_NewDoubleObj(r); break; } case SQLITE_NULL: { | | | 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 | } case SQLITE_FLOAT: { double r = sqlite3_value_double(pIn); pVal = Tcl_NewDoubleObj(r); break; } case SQLITE_NULL: { pVal = Tcl_NewStringObj(p->pDb->zNull, -1); break; } default: { int bytes = sqlite3_value_bytes(pIn); pVal = Tcl_NewStringObj((char *)sqlite3_value_text(pIn), bytes); break; } |
︙ | ︙ | |||
973 974 975 976 977 978 979 | }else{ rc = 999; } return rc; } #endif /* SQLITE_OMIT_AUTHORIZATION */ | < < < < < < < < < < < < < < < < < < < < | 968 969 970 971 972 973 974 975 976 977 978 979 980 981 | }else{ rc = 999; } return rc; } #endif /* SQLITE_OMIT_AUTHORIZATION */ /* ** This routine reads a line of text from FILE in, stores ** the text in memory obtained from malloc() and returns a pointer ** to the text. NULL is returned at end of file, or if malloc() ** fails. ** ** The interface is like "readline" but no command-line editing |
︙ | ︙ | |||
1180 1181 1182 1183 1184 1185 1186 | /* If no prepared statement was found. Compile the SQL text. Also allocate ** a new SqlPreparedStmt structure. */ if( pPreStmt==0 ){ int nByte; if( SQLITE_OK!=dbPrepare(pDb, zSql, &pStmt, pzOut) ){ | | | | 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 | /* If no prepared statement was found. Compile the SQL text. Also allocate ** a new SqlPreparedStmt structure. */ if( pPreStmt==0 ){ int nByte; if( SQLITE_OK!=dbPrepare(pDb, zSql, &pStmt, pzOut) ){ Tcl_SetObjResult(interp, Tcl_NewStringObj(sqlite3_errmsg(pDb->db), -1)); return TCL_ERROR; } if( pStmt==0 ){ if( SQLITE_OK!=sqlite3_errcode(pDb->db) ){ /* A compile-time error in the statement. */ Tcl_SetObjResult(interp, Tcl_NewStringObj(sqlite3_errmsg(pDb->db), -1)); return TCL_ERROR; }else{ /* The statement was a no-op. Continue to the next statement ** in the SQL string. */ return TCL_OK; } |
︙ | ︙ | |||
1405 1406 1407 1408 1409 1410 1411 | int nCol; /* Number of columns returned by pStmt */ Tcl_Obj **apColName = 0; /* Array of column names */ p->nCol = nCol = sqlite3_column_count(pStmt); if( nCol>0 && (papColName || p->pArray) ){ apColName = (Tcl_Obj**)Tcl_Alloc( sizeof(Tcl_Obj*)*nCol ); for(i=0; i<nCol; i++){ | | | 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 | int nCol; /* Number of columns returned by pStmt */ Tcl_Obj **apColName = 0; /* Array of column names */ p->nCol = nCol = sqlite3_column_count(pStmt); if( nCol>0 && (papColName || p->pArray) ){ apColName = (Tcl_Obj**)Tcl_Alloc( sizeof(Tcl_Obj*)*nCol ); for(i=0; i<nCol; i++){ apColName[i] = Tcl_NewStringObj(sqlite3_column_name(pStmt,i), -1); Tcl_IncrRefCount(apColName[i]); } p->apColName = apColName; } /* If results are being stored in an array variable, then create ** the array(*) entry for that array |
︙ | ︙ | |||
1492 1493 1494 1495 1496 1497 1498 | ** interface, retry prepare()/step() on the same SQL statement. ** This only happens once. If there is a second SQLITE_SCHEMA ** error, the error will be returned to the caller. */ p->zSql = zPrevSql; continue; } #endif | | > | 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 | ** interface, retry prepare()/step() on the same SQL statement. ** This only happens once. If there is a second SQLITE_SCHEMA ** error, the error will be returned to the caller. */ p->zSql = zPrevSql; continue; } #endif Tcl_SetObjResult(pDb->interp, Tcl_NewStringObj(sqlite3_errmsg(pDb->db), -1)); return TCL_ERROR; }else{ dbReleaseStmt(pDb, pPreStmt, 0); } } } |
︙ | ︙ | |||
1549 1550 1551 1552 1553 1554 1555 | return Tcl_NewWideIntObj(v); } } case SQLITE_FLOAT: { return Tcl_NewDoubleObj(sqlite3_column_double(pStmt, iCol)); } case SQLITE_NULL: { | | | | 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 | return Tcl_NewWideIntObj(v); } } case SQLITE_FLOAT: { return Tcl_NewDoubleObj(sqlite3_column_double(pStmt, iCol)); } case SQLITE_NULL: { return Tcl_NewStringObj(p->pDb->zNull, -1); } } return Tcl_NewStringObj((char*)sqlite3_column_text(pStmt, iCol), -1); } /* ** If using Tcl version 8.6 or greater, use the NR functions to avoid ** recursive evalution of scripts by the [db eval] and [db trans] ** commands. Even if the headers used while compiling the extension ** are 8.6 or newer, the code still tests the Tcl version at runtime. |
︙ | ︙ | |||
2448 2449 2450 2451 2452 2453 2454 | Tcl_AppendResult(interp, "incrblob not available in this build", 0); return TCL_ERROR; #else int isReadonly = 0; const char *zDb = "main"; const char *zTable; const char *zColumn; | | | 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 | Tcl_AppendResult(interp, "incrblob not available in this build", 0); return TCL_ERROR; #else int isReadonly = 0; const char *zDb = "main"; const char *zTable; const char *zColumn; Tcl_WideInt iRow; /* Check for the -readonly option */ if( objc>3 && strcmp(Tcl_GetString(objv[2]), "-readonly")==0 ){ isReadonly = 1; } if( objc!=(5+isReadonly) && objc!=(6+isReadonly) ){ |
︙ | ︙ | |||
2514 2515 2516 2517 2518 2519 2520 | pDb->zNull = Tcl_Alloc( len + 1 ); memcpy(pDb->zNull, zNull, len); pDb->zNull[len] = '\0'; }else{ pDb->zNull = 0; } } | | | 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 | pDb->zNull = Tcl_Alloc( len + 1 ); memcpy(pDb->zNull, zNull, len); pDb->zNull[len] = '\0'; }else{ pDb->zNull = 0; } } Tcl_SetObjResult(interp, Tcl_NewStringObj(pDb->zNull, -1)); break; } /* ** $db last_insert_rowid ** ** Return an integer which is the ROWID for the most recent insert. |
︙ | ︙ |
Changes to src/test1.c.
︙ | ︙ | |||
3063 3064 3065 3066 3067 3068 3069 | void * clientData, Tcl_Interp *interp, int objc, Tcl_Obj *CONST objv[] ){ sqlite3_stmt *pStmt; int idx; | | | 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 | void * clientData, Tcl_Interp *interp, int objc, Tcl_Obj *CONST objv[] ){ sqlite3_stmt *pStmt; int idx; Tcl_WideInt value; int rc; if( objc!=4 ){ Tcl_AppendResult(interp, "wrong # args: should be \"", Tcl_GetStringFromObj(objv[0], 0), " STMT N VALUE", 0); return TCL_ERROR; } |
︙ | ︙ | |||
4699 4700 4701 4702 4703 4704 4705 | static int test_soft_heap_limit( void * clientData, Tcl_Interp *interp, int objc, Tcl_Obj *CONST objv[] ){ sqlite3_int64 amt; | | | 4699 4700 4701 4702 4703 4704 4705 4706 4707 4708 4709 4710 4711 4712 4713 | static int test_soft_heap_limit( void * clientData, Tcl_Interp *interp, int objc, Tcl_Obj *CONST objv[] ){ sqlite3_int64 amt; Tcl_WideInt N = -1; if( objc!=1 && objc!=2 ){ Tcl_WrongNumArgs(interp, 1, objv, "?N?"); return TCL_ERROR; } if( objc==2 ){ if( Tcl_GetWideIntFromObj(interp, objv[1], &N) ) return TCL_ERROR; } |
︙ | ︙ | |||
5092 5093 5094 5095 5096 5097 5098 | */ static int file_control_sizehint_test( ClientData clientData, /* Pointer to sqlite3_enable_XXX function */ Tcl_Interp *interp, /* The TCL interpreter that invoked this command */ int objc, /* Number of arguments */ Tcl_Obj *CONST objv[] /* Command arguments */ ){ | | | 5092 5093 5094 5095 5096 5097 5098 5099 5100 5101 5102 5103 5104 5105 5106 | */ static int file_control_sizehint_test( ClientData clientData, /* Pointer to sqlite3_enable_XXX function */ Tcl_Interp *interp, /* The TCL interpreter that invoked this command */ int objc, /* Number of arguments */ Tcl_Obj *CONST objv[] /* Command arguments */ ){ Tcl_WideInt nSize; /* Hinted size */ char *zDb; /* Db name ("main", "temp" etc.) */ sqlite3 *db; /* Database handle */ int rc; /* file_control() return code */ if( objc!=4 ){ Tcl_WrongNumArgs(interp, 1, objv, "DB DBNAME SIZE"); return TCL_ERROR; |
︙ | ︙ |
Changes to src/test_intarray.c.
︙ | ︙ | |||
342 343 344 345 346 347 348 | #ifndef SQLITE_OMIT_VIRTUALTABLE a = sqlite3_malloc( sizeof(a[0])*n ); if( a==0 ){ Tcl_AppendResult(interp, "SQLITE_NOMEM", (char*)0); return TCL_ERROR; } for(i=0; i<n; i++){ | | | > | 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 | #ifndef SQLITE_OMIT_VIRTUALTABLE a = sqlite3_malloc( sizeof(a[0])*n ); if( a==0 ){ Tcl_AppendResult(interp, "SQLITE_NOMEM", (char*)0); return TCL_ERROR; } for(i=0; i<n; i++){ Tcl_WideInt x = 0; Tcl_GetWideIntFromObj(0, objv[i+2], &x); a[i] = x; } rc = sqlite3_intarray_bind(pArray, n, a, sqlite3_free); if( rc!=SQLITE_OK ){ Tcl_AppendResult(interp, sqlite3TestErrorName(rc), (char*)0); return TCL_ERROR; } #endif |
︙ | ︙ |
Changes to src/test_quota.c.
︙ | ︙ | |||
1069 1070 1071 1072 1073 1074 1075 | pFile = 0; } rc = fwrite(pBuf, size, nmemb, p->f); /* If the write was incomplete, adjust the file size and group size ** downward */ if( rc<nmemb && pFile ){ | | | 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 | pFile = 0; } rc = fwrite(pBuf, size, nmemb, p->f); /* If the write was incomplete, adjust the file size and group size ** downward */ if( rc<nmemb && pFile ){ size_t nWritten = rc; sqlite3_int64 iNewEnd = iOfst + size*nWritten; if( iNewEnd<iEnd ) iNewEnd = iEnd; quotaEnter(); pFile->pGroup->iSize += iNewEnd - pFile->iSize; pFile->iSize = iNewEnd; quotaLeave(); } |
︙ | ︙ | |||
1350 1351 1352 1353 1354 1355 1356 1357 | Tcl_IncrRefCount(pEval); Tcl_ListObjAppendElement(0, pEval, Tcl_NewStringObj(zFilename, -1)); Tcl_ListObjAppendElement(0, pEval, pVarname); Tcl_ListObjAppendElement(0, pEval, Tcl_NewWideIntObj(iSize)); rc = Tcl_EvalObjEx(p->interp, pEval, TCL_EVAL_GLOBAL); if( rc==TCL_OK ){ Tcl_Obj *pLimit = Tcl_ObjGetVar2(p->interp, pVarname, 0, 0); | > | > | 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 | Tcl_IncrRefCount(pEval); Tcl_ListObjAppendElement(0, pEval, Tcl_NewStringObj(zFilename, -1)); Tcl_ListObjAppendElement(0, pEval, pVarname); Tcl_ListObjAppendElement(0, pEval, Tcl_NewWideIntObj(iSize)); rc = Tcl_EvalObjEx(p->interp, pEval, TCL_EVAL_GLOBAL); if( rc==TCL_OK ){ Tcl_WideInt x; Tcl_Obj *pLimit = Tcl_ObjGetVar2(p->interp, pVarname, 0, 0); rc = Tcl_GetWideIntFromObj(p->interp, pLimit, &x); *piLimit = x; Tcl_UnsetVar(p->interp, Tcl_GetString(pVarname), 0); } Tcl_DecrRefCount(pEval); Tcl_DecrRefCount(pVarname); if( rc!=TCL_OK ) Tcl_BackgroundError(p->interp); } |
︙ | ︙ | |||
1433 1434 1435 1436 1437 1438 1439 | static int test_quota_set( void * clientData, Tcl_Interp *interp, int objc, Tcl_Obj *CONST objv[] ){ const char *zPattern; /* File pattern to configure */ | | | 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 | static int test_quota_set( void * clientData, Tcl_Interp *interp, int objc, Tcl_Obj *CONST objv[] ){ const char *zPattern; /* File pattern to configure */ Tcl_WideInt iLimit; /* Initial quota in bytes */ Tcl_Obj *pScript; /* Tcl script to invoke to increase quota */ int rc; /* Value returned by quota_set() */ TclQuotaCallback *p; /* Callback object */ int nScript; /* Length of callback script */ void (*xDestroy)(void*); /* Optional destructor for pArg */ void (*xCallback)(const char *, sqlite3_int64 *, sqlite3_int64, void *); |
︙ | ︙ | |||
1609 1610 1611 1612 1613 1614 1615 | if( Tcl_GetIntFromObj(interp, objv[3], &nElem) ) return TCL_ERROR; zBuf = (char*)sqlite3_malloc( sz*nElem + 1 ); if( zBuf==0 ){ Tcl_SetResult(interp, "out of memory", TCL_STATIC); return TCL_ERROR; } got = sqlite3_quota_fread(zBuf, sz, nElem, p); | < | 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 | if( Tcl_GetIntFromObj(interp, objv[3], &nElem) ) return TCL_ERROR; zBuf = (char*)sqlite3_malloc( sz*nElem + 1 ); if( zBuf==0 ){ Tcl_SetResult(interp, "out of memory", TCL_STATIC); return TCL_ERROR; } got = sqlite3_quota_fread(zBuf, sz, nElem, p); zBuf[got*sz] = 0; Tcl_SetResult(interp, zBuf, TCL_VOLATILE); sqlite3_free(zBuf); return TCL_OK; } /* |
︙ | ︙ |
Changes to src/vdbeaux.c.
︙ | ︙ | |||
1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 | ** file is required for an atomic commit. */ for(i=0; rc==SQLITE_OK && i<db->nDb; i++){ Btree *pBt = db->aDb[i].pBt; if( sqlite3BtreeIsInTrans(pBt) ){ needXcommit = 1; if( i!=1 ) nTrans++; rc = sqlite3PagerExclusiveLock(sqlite3BtreePager(pBt)); } } if( rc!=SQLITE_OK ){ return rc; } /* If there are any write-transactions at all, invoke the commit hook */ | > > | 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 | ** file is required for an atomic commit. */ for(i=0; rc==SQLITE_OK && i<db->nDb; i++){ Btree *pBt = db->aDb[i].pBt; if( sqlite3BtreeIsInTrans(pBt) ){ needXcommit = 1; if( i!=1 ) nTrans++; sqlite3BtreeEnter(pBt); rc = sqlite3PagerExclusiveLock(sqlite3BtreePager(pBt)); sqlite3BtreeLeave(pBt); } } if( rc!=SQLITE_OK ){ return rc; } /* If there are any write-transactions at all, invoke the commit hook */ |
︙ | ︙ |
Changes to src/vtab.c.
︙ | ︙ | |||
260 261 262 263 264 265 266 | ** database connection. */ void sqlite3VtabClear(sqlite3 *db, Table *p){ if( !db || db->pnBytesFreed==0 ) vtabDisconnectAll(0, p); if( p->azModuleArg ){ int i; for(i=0; i<p->nModuleArg; i++){ | | | 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 | ** database connection. */ void sqlite3VtabClear(sqlite3 *db, Table *p){ if( !db || db->pnBytesFreed==0 ) vtabDisconnectAll(0, p); if( p->azModuleArg ){ int i; for(i=0; i<p->nModuleArg; i++){ if( i!=1 ) sqlite3DbFree(db, p->azModuleArg[i]); } sqlite3DbFree(db, p->azModuleArg); } } /* ** Add a new module argument to pTable->azModuleArg[]. |
︙ | ︙ | |||
320 321 322 323 324 325 326 | db = pParse->db; iDb = sqlite3SchemaToIndex(db, pTable->pSchema); assert( iDb>=0 ); pTable->tabFlags |= TF_Virtual; pTable->nModuleArg = 0; addModuleArgument(db, pTable, sqlite3NameFromToken(db, pModuleName)); | | | 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 | db = pParse->db; iDb = sqlite3SchemaToIndex(db, pTable->pSchema); assert( iDb>=0 ); pTable->tabFlags |= TF_Virtual; pTable->nModuleArg = 0; addModuleArgument(db, pTable, sqlite3NameFromToken(db, pModuleName)); addModuleArgument(db, pTable, 0); addModuleArgument(db, pTable, sqlite3DbStrDup(db, pTable->zName)); pParse->sNameToken.n = (int)(&pModuleName->z[pModuleName->n] - pName1->z); #ifndef SQLITE_OMIT_AUTHORIZATION /* Creating a virtual table invokes the authorization callback twice. ** The first invocation, to obtain permission to INSERT a row into the ** sqlite_master table, has already been made by sqlite3StartTable(). |
︙ | ︙ | |||
477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 | VtabCtx sCtx, *pPriorCtx; VTable *pVTable; int rc; const char *const*azArg = (const char *const*)pTab->azModuleArg; int nArg = pTab->nModuleArg; char *zErr = 0; char *zModuleName = sqlite3MPrintf(db, "%s", pTab->zName); if( !zModuleName ){ return SQLITE_NOMEM; } pVTable = sqlite3DbMallocZero(db, sizeof(VTable)); if( !pVTable ){ sqlite3DbFree(db, zModuleName); return SQLITE_NOMEM; } pVTable->db = db; pVTable->pMod = pMod; /* Invoke the virtual table constructor */ assert( &db->pVtabCtx ); assert( xConstruct ); sCtx.pTab = pTab; sCtx.pVTable = pVTable; pPriorCtx = db->pVtabCtx; db->pVtabCtx = &sCtx; rc = xConstruct(db, pMod->pAux, nArg, azArg, &pVTable->pVtab, &zErr); db->pVtabCtx = pPriorCtx; if( rc==SQLITE_NOMEM ) db->mallocFailed = 1; if( SQLITE_OK!=rc ){ if( zErr==0 ){ *pzErr = sqlite3MPrintf(db, "vtable constructor failed: %s", zModuleName); }else { *pzErr = sqlite3MPrintf(db, "%s", zErr); sqlite3_free(zErr); | > > > > > > | 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 | VtabCtx sCtx, *pPriorCtx; VTable *pVTable; int rc; const char *const*azArg = (const char *const*)pTab->azModuleArg; int nArg = pTab->nModuleArg; char *zErr = 0; char *zModuleName = sqlite3MPrintf(db, "%s", pTab->zName); int iDb; if( !zModuleName ){ return SQLITE_NOMEM; } pVTable = sqlite3DbMallocZero(db, sizeof(VTable)); if( !pVTable ){ sqlite3DbFree(db, zModuleName); return SQLITE_NOMEM; } pVTable->db = db; pVTable->pMod = pMod; assert( pTab->azModuleArg[1]==0 ); iDb = sqlite3SchemaToIndex(db, pTab->pSchema); pTab->azModuleArg[1] = db->aDb[iDb].zName; /* Invoke the virtual table constructor */ assert( &db->pVtabCtx ); assert( xConstruct ); sCtx.pTab = pTab; sCtx.pVTable = pVTable; pPriorCtx = db->pVtabCtx; db->pVtabCtx = &sCtx; rc = xConstruct(db, pMod->pAux, nArg, azArg, &pVTable->pVtab, &zErr); db->pVtabCtx = pPriorCtx; if( rc==SQLITE_NOMEM ) db->mallocFailed = 1; pTab->azModuleArg[1] = 0; if( SQLITE_OK!=rc ){ if( zErr==0 ){ *pzErr = sqlite3MPrintf(db, "vtable constructor failed: %s", zModuleName); }else { *pzErr = sqlite3MPrintf(db, "%s", zErr); sqlite3_free(zErr); |
︙ | ︙ |
Changes to src/wal.c.
︙ | ︙ | |||
2824 2825 2826 2827 2828 2829 2830 | ** final frame is repeated (with its commit mark) until the next sector ** boundary is crossed. Only the part of the WAL prior to the last ** sector boundary is synced; the part of the last frame that extends ** past the sector boundary is written after the sync. */ if( isCommit && (sync_flags & WAL_SYNC_TRANSACTIONS)!=0 ){ if( pWal->padToSectorBoundary ){ | | | 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 | ** final frame is repeated (with its commit mark) until the next sector ** boundary is crossed. Only the part of the WAL prior to the last ** sector boundary is synced; the part of the last frame that extends ** past the sector boundary is written after the sync. */ if( isCommit && (sync_flags & WAL_SYNC_TRANSACTIONS)!=0 ){ if( pWal->padToSectorBoundary ){ int sectorSize = sqlite3SectorSize(pWal->pWalFd); w.iSyncPoint = ((iOffset+sectorSize-1)/sectorSize)*sectorSize; while( iOffset<w.iSyncPoint ){ rc = walWriteOneFrame(&w, pLast, nTruncate, iOffset); if( rc ) return rc; iOffset += szFrame; nExtra++; } |
︙ | ︙ |
Changes to src/where.c.
︙ | ︙ | |||
253 254 255 256 257 258 259 | #define WHERE_COLUMN_NULL 0x00080000 /* x IS NULL */ #define WHERE_INDEXED 0x000f0000 /* Anything that uses an index */ #define WHERE_NOT_FULLSCAN 0x100f3000 /* Does not do a full table scan */ #define WHERE_IN_ABLE 0x000f1000 /* Able to support an IN operator */ #define WHERE_TOP_LIMIT 0x00100000 /* x<EXPR or x<=EXPR constraint */ #define WHERE_BTM_LIMIT 0x00200000 /* x>EXPR or x>=EXPR constraint */ #define WHERE_BOTH_LIMIT 0x00300000 /* Both x>EXPR and x<EXPR */ | | | | | > | 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 | #define WHERE_COLUMN_NULL 0x00080000 /* x IS NULL */ #define WHERE_INDEXED 0x000f0000 /* Anything that uses an index */ #define WHERE_NOT_FULLSCAN 0x100f3000 /* Does not do a full table scan */ #define WHERE_IN_ABLE 0x000f1000 /* Able to support an IN operator */ #define WHERE_TOP_LIMIT 0x00100000 /* x<EXPR or x<=EXPR constraint */ #define WHERE_BTM_LIMIT 0x00200000 /* x>EXPR or x>=EXPR constraint */ #define WHERE_BOTH_LIMIT 0x00300000 /* Both x>EXPR and x<EXPR */ #define WHERE_IDX_ONLY 0x00400000 /* Use index only - omit table */ #define WHERE_ORDERED 0x00800000 /* Output will appear in correct order */ #define WHERE_REVERSE 0x01000000 /* Scan in reverse order */ #define WHERE_UNIQUE 0x02000000 /* Selects no more than one row */ #define WHERE_ALL_UNIQUE 0x04000000 /* This and all prior have one row */ #define WHERE_VIRTUALTABLE 0x08000000 /* Use virtual-table processing */ #define WHERE_MULTI_OR 0x10000000 /* OR using multiple indices */ #define WHERE_TEMP_INDEX 0x20000000 /* Uses an ephemeral index */ #define WHERE_DISTINCT 0x40000000 /* Correct order for DISTINCT */ #define WHERE_COVER_SCAN 0x80000000 /* Full scan of a covering index */ /* |
︙ | ︙ | |||
284 285 286 287 288 289 290 291 292 293 294 295 296 297 | ExprList *pOrderBy; /* The ORDER BY clause */ ExprList *pDistinct; /* The select-list if query is DISTINCT */ sqlite3_index_info **ppIdxInfo; /* Index information passed to xBestIndex */ int i, n; /* Which loop is being coded; # of loops */ WhereLevel *aLevel; /* Info about outer loops */ WhereCost cost; /* Lowest cost query plan */ }; /* ** Initialize a preallocated WhereClause structure. */ static void whereClauseInit( WhereClause *pWC, /* The WhereClause to be initialized */ Parse *pParse, /* The parsing context */ | > > > > > > > > > > > | 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 | ExprList *pOrderBy; /* The ORDER BY clause */ ExprList *pDistinct; /* The select-list if query is DISTINCT */ sqlite3_index_info **ppIdxInfo; /* Index information passed to xBestIndex */ int i, n; /* Which loop is being coded; # of loops */ WhereLevel *aLevel; /* Info about outer loops */ WhereCost cost; /* Lowest cost query plan */ }; /* ** Return TRUE if the probe cost is less than the baseline cost */ static int compareCost(const WhereCost *pProbe, const WhereCost *pBaseline){ if( pProbe->rCost<pBaseline->rCost ) return 1; if( pProbe->rCost>pBaseline->rCost ) return 0; if( pProbe->plan.nOBSat>pBaseline->plan.nOBSat ) return 1; if( pProbe->plan.nRow<pBaseline->plan.nRow ) return 1; return 0; } /* ** Initialize a preallocated WhereClause structure. */ static void whereClauseInit( WhereClause *pWC, /* The WhereClause to be initialized */ Parse *pParse, /* The parsing context */ |
︙ | ︙ | |||
1426 1427 1428 1429 1430 1431 1432 | /* Prevent ON clause terms of a LEFT JOIN from being used to drive ** an index for tables to the left of the join. */ pTerm->prereqRight |= extraRight; } | < < < < < < < < < < < < < < < | 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 | /* Prevent ON clause terms of a LEFT JOIN from being used to drive ** an index for tables to the left of the join. */ pTerm->prereqRight |= extraRight; } /* ** This function searches the expression list passed as the second argument ** for an expression of type TK_COLUMN that refers to the same column and ** uses the same collation sequence as the iCol'th column of index pIdx. ** Argument iBase is the cursor number used for the table that pIdx refers ** to. ** |
︙ | ︙ | |||
1500 1501 1502 1503 1504 1505 1506 | int base, /* Cursor number for the table pIdx is on */ ExprList *pDistinct, /* The DISTINCT expressions */ int nEqCol /* Number of index columns with == */ ){ Bitmask mask = 0; /* Mask of unaccounted for pDistinct exprs */ int i; /* Iterator variable */ | > | | 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 | int base, /* Cursor number for the table pIdx is on */ ExprList *pDistinct, /* The DISTINCT expressions */ int nEqCol /* Number of index columns with == */ ){ Bitmask mask = 0; /* Mask of unaccounted for pDistinct exprs */ int i; /* Iterator variable */ assert( pDistinct!=0 ); if( pIdx->zName==0 || pDistinct->nExpr>=BMS ) return 0; testcase( pDistinct->nExpr==BMS-1 ); /* Loop through all the expressions in the distinct list. If any of them ** are not simple column references, return early. Otherwise, test if the ** WHERE clause contains a "col=X" clause. If it does, the expression ** can be ignored. If it does not, and the column does not belong to the ** same table as index pIdx, return early. Finally, if there is no |
︙ | ︙ | |||
1602 1603 1604 1605 1606 1607 1608 | return 1; } } return 0; } | < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < < | 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 | return 1; } } return 0; } /* ** Prepare a crude estimate of the logarithm of the input value. ** The results need not be exact. This is only used for estimating ** the total cost of performing operations with O(logN) or O(NlogN) ** complexity. Because N is just a guess, it is no great tragedy if ** logN is a little off. */ |
︙ | ︙ | |||
1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 | ** less than the current cost stored in pCost, replace the contents ** of pCost. */ WHERETRACE(("... multi-index OR cost=%.9g nrow=%.9g\n", rTotal, nRow)); if( rTotal<p->cost.rCost ){ p->cost.rCost = rTotal; p->cost.used = used; p->cost.plan.nRow = nRow; p->cost.plan.wsFlags = flags; p->cost.plan.u.pTerm = pTerm; } } } #endif /* SQLITE_OMIT_OR_OPTIMIZATION */ } | > | 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 | ** less than the current cost stored in pCost, replace the contents ** of pCost. */ WHERETRACE(("... multi-index OR cost=%.9g nrow=%.9g\n", rTotal, nRow)); if( rTotal<p->cost.rCost ){ p->cost.rCost = rTotal; p->cost.used = used; p->cost.plan.nRow = nRow; p->cost.plan.nOBSat = p->i ? p->aLevel[p->i-1].plan.nOBSat : 0; p->cost.plan.wsFlags = flags; p->cost.plan.u.pTerm = pTerm; } } } #endif /* SQLITE_OMIT_OR_OPTIMIZATION */ } |
︙ | ︙ | |||
2457 2458 2459 2460 2461 2462 2463 | if( (SQLITE_BIG_DBL/((double)2))<rCost ){ p->cost.rCost = (SQLITE_BIG_DBL/((double)2)); }else{ p->cost.rCost = rCost; } p->cost.plan.u.pVtabIdx = pIdxInfo; if( pIdxInfo->orderByConsumed ){ | | > > > | 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 | if( (SQLITE_BIG_DBL/((double)2))<rCost ){ p->cost.rCost = (SQLITE_BIG_DBL/((double)2)); }else{ p->cost.rCost = rCost; } p->cost.plan.u.pVtabIdx = pIdxInfo; if( pIdxInfo->orderByConsumed ){ p->cost.plan.wsFlags |= WHERE_ORDERED; p->cost.plan.nOBSat = nOrderBy; }else{ p->cost.plan.nOBSat = p->i ? p->aLevel[p->i-1].plan.nOBSat : 0; } p->cost.plan.nEq = 0; pIdxInfo->nOrderBy = nOrderBy; /* Try to find a more efficient access pattern by using multiple indexes ** to optimize an OR expression within the WHERE clause. */ |
︙ | ︙ | |||
2553 2554 2555 2556 2557 2558 2559 | CollSeq *pColl; const u8 *z; if( eType==SQLITE_BLOB ){ z = (const u8 *)sqlite3_value_blob(pVal); pColl = db->pDfltColl; assert( pColl->enc==SQLITE_UTF8 ); }else{ | | < < | 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 | CollSeq *pColl; const u8 *z; if( eType==SQLITE_BLOB ){ z = (const u8 *)sqlite3_value_blob(pVal); pColl = db->pDfltColl; assert( pColl->enc==SQLITE_UTF8 ); }else{ pColl = sqlite3GetCollSeq(pParse, SQLITE_UTF8, 0, *pIdx->azColl); if( pColl==0 ){ return SQLITE_ERROR; } z = (const u8 *)sqlite3ValueText(pVal, pColl->enc); if( !z ){ return SQLITE_NOMEM; } assert( z && pColl && pColl->xCmp ); |
︙ | ︙ | |||
2866 2867 2868 2869 2870 2871 2872 | } return rc; } #endif /* defined(SQLITE_ENABLE_STAT3) */ /* ** Check to see if column iCol of the table with cursor iTab will appear | | | > < | | > > | > > > > | > > > | > | | < < | < < | | > | > > > > > > > | > > > > > > > > > > | > > > > > | < < > > > > > > > > > > > > > > > > | > > > > > > | > > > > | > > > > > > > > > > > > > > > > > > > | > > > > > > > | > > > > > > > > > > > > > > > | > > > > | > > > > | | > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > | > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > | > > > | > > > > > > > > > > > > > > > > | 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 | } return rc; } #endif /* defined(SQLITE_ENABLE_STAT3) */ /* ** Check to see if column iCol of the table with cursor iTab will appear ** in sorted order according to the current query plan. ** ** Return values: ** ** 0 iCol is not ordered ** 1 iCol has only a single value ** 2 iCol is in ASC order ** 3 iCol is in DESC order */ static int isOrderedColumn( WhereBestIdx *p, int iTab, int iCol ){ int i, j; WhereLevel *pLevel = &p->aLevel[p->i-1]; Index *pIdx; u8 sortOrder; for(i=p->i-1; i>=0; i--, pLevel--){ if( pLevel->iTabCur!=iTab ) continue; if( (pLevel->plan.wsFlags & WHERE_ALL_UNIQUE)!=0 ){ return 1; } assert( (pLevel->plan.wsFlags & WHERE_ORDERED)!=0 ); if( (pIdx = pLevel->plan.u.pIdx)!=0 ){ if( iCol<0 ){ sortOrder = 0; testcase( (pLevel->plan.wsFlags & WHERE_REVERSE)!=0 ); }else{ int n = pIdx->nColumn; for(j=0; j<n; j++){ if( iCol==pIdx->aiColumn[j] ) break; } if( j>=n ) return 0; sortOrder = pIdx->aSortOrder[j]; testcase( (pLevel->plan.wsFlags & WHERE_REVERSE)!=0 ); } }else{ if( iCol!=(-1) ) return 0; sortOrder = 0; testcase( (pLevel->plan.wsFlags & WHERE_REVERSE)!=0 ); } if( (pLevel->plan.wsFlags & WHERE_REVERSE)!=0 ){ assert( sortOrder==0 || sortOrder==1 ); testcase( sortOrder==1 ); sortOrder = 1 - sortOrder; } return sortOrder+2; } return 0; } /* ** This routine decides if pIdx can be used to satisfy the ORDER BY ** clause, either in whole or in part. The return value is the ** cumulative number of terms in the ORDER BY clause that are satisfied ** by the index pIdx and other indices in outer loops. ** ** The table being queried has a cursor number of "base". pIdx is the ** index that is postulated for use to access the table. ** ** The *pbRev value is set to 0 order 1 depending on whether or not ** pIdx should be run in the forward order or in reverse order. */ static int isSortingIndex( WhereBestIdx *p, /* Best index search context */ Index *pIdx, /* The index we are testing */ int base, /* Cursor number for the table to be sorted */ int *pbRev /* Set to 1 for reverse-order scan of pIdx */ ){ int i; /* Number of pIdx terms used */ int j; /* Number of ORDER BY terms satisfied */ int sortOrder = 2; /* 0: forward. 1: backward. 2: unknown */ int nTerm; /* Number of ORDER BY terms */ struct ExprList_item *pOBItem;/* A term of the ORDER BY clause */ Table *pTab = pIdx->pTable; /* Table that owns index pIdx */ ExprList *pOrderBy; /* The ORDER BY clause */ Parse *pParse = p->pParse; /* Parser context */ sqlite3 *db = pParse->db; /* Database connection */ int nPriorSat; /* ORDER BY terms satisfied by outer loops */ int seenRowid = 0; /* True if an ORDER BY rowid term is seen */ int uniqueNotNull; /* pIdx is UNIQUE with all terms are NOT NULL */ if( p->i==0 ){ nPriorSat = 0; }else{ nPriorSat = p->aLevel[p->i-1].plan.nOBSat; if( (p->aLevel[p->i-1].plan.wsFlags & WHERE_ORDERED)==0 ){ /* This loop cannot be ordered unless the next outer loop is ** also ordered */ return nPriorSat; } if( OptimizationDisabled(db, SQLITE_OrderByIdxJoin) ){ /* Only look at the outer-most loop if the OrderByIdxJoin ** optimization is disabled */ return nPriorSat; } } pOrderBy = p->pOrderBy; assert( pOrderBy!=0 ); if( pIdx->bUnordered ){ /* Hash indices (indicated by the "unordered" tag on sqlite_stat1) cannot ** be used for sorting */ return nPriorSat; } nTerm = pOrderBy->nExpr; uniqueNotNull = pIdx->onError!=OE_None; assert( nTerm>0 ); /* Argument pIdx must either point to a 'real' named index structure, ** or an index structure allocated on the stack by bestBtreeIndex() to ** represent the rowid index that is part of every table. */ assert( pIdx->zName || (pIdx->nColumn==1 && pIdx->aiColumn[0]==-1) ); /* Match terms of the ORDER BY clause against columns of ** the index. ** ** Note that indices have pIdx->nColumn regular columns plus ** one additional column containing the rowid. The rowid column ** of the index is also allowed to match against the ORDER BY ** clause. */ j = nPriorSat; for(i=0,pOBItem=&pOrderBy->a[j]; j<nTerm && i<=pIdx->nColumn; i++){ Expr *pOBExpr; /* The expression of the ORDER BY pOBItem */ CollSeq *pColl; /* The collating sequence of pOBExpr */ int termSortOrder; /* Sort order for this term */ int iColumn; /* The i-th column of the index. -1 for rowid */ int iSortOrder; /* 1 for DESC, 0 for ASC on the i-th index term */ int isEq; /* Subject to an == or IS NULL constraint */ int isMatch; /* ORDER BY term matches the index term */ const char *zColl; /* Name of collating sequence for i-th index term */ WhereTerm *pConstraint; /* A constraint in the WHERE clause */ /* If the next term of the ORDER BY clause refers to anything other than ** a column in the "base" table, then this index will not be of any ** further use in handling the ORDER BY. */ pOBExpr = pOBItem->pExpr; if( pOBExpr->op!=TK_COLUMN || pOBExpr->iTable!=base ){ break; } /* Find column number and collating sequence for the next entry ** in the index */ if( pIdx->zName && i<pIdx->nColumn ){ iColumn = pIdx->aiColumn[i]; if( iColumn==pIdx->pTable->iPKey ){ iColumn = -1; } iSortOrder = pIdx->aSortOrder[i]; zColl = pIdx->azColl[i]; assert( zColl!=0 ); }else{ iColumn = -1; iSortOrder = 0; zColl = 0; } /* Check to see if the column number and collating sequence of the ** index match the column number and collating sequence of the ORDER BY ** clause entry. Set isMatch to 1 if they both match. */ if( pOBExpr->iColumn==iColumn ){ if( zColl ){ pColl = sqlite3ExprCollSeq(pParse, pOBExpr); if( !pColl ) pColl = db->pDfltColl; isMatch = sqlite3StrICmp(pColl->zName, zColl)==0; }else{ isMatch = 1; } }else{ isMatch = 0; } /* termSortOrder is 0 or 1 for whether or not the access loop should ** run forward or backwards (respectively) in order to satisfy this ** term of the ORDER BY clause. */ assert( pOBItem->sortOrder==0 || pOBItem->sortOrder==1 ); assert( iSortOrder==0 || iSortOrder==1 ); termSortOrder = iSortOrder ^ pOBItem->sortOrder; /* If X is the column in the index and ORDER BY clause, check to see ** if there are any X= or X IS NULL constraints in the WHERE clause. */ pConstraint = findTerm(p->pWC, base, iColumn, p->notReady, WO_EQ|WO_ISNULL|WO_IN, pIdx); if( pConstraint==0 ){ isEq = 0; }else if( pConstraint->eOperator==WO_IN ){ /* Constraints of the form: "X IN ..." cannot be used with an ORDER BY ** because we do not know in what order the values on the RHS of the IN ** operator will occur. */ break; }else if( pConstraint->eOperator==WO_ISNULL ){ uniqueNotNull = 0; isEq = 1; /* "X IS NULL" means X has only a single value */ }else if( pConstraint->prereqRight==0 ){ isEq = 1; /* Constraint "X=constant" means X has only a single value */ }else{ Expr *pRight = pConstraint->pExpr->pRight; if( pRight->op==TK_COLUMN ){ WHERETRACE((" .. isOrderedColumn(tab=%d,col=%d)", pRight->iTable, pRight->iColumn)); isEq = isOrderedColumn(p, pRight->iTable, pRight->iColumn); WHERETRACE((" -> isEq=%d\n", isEq)); /* If the constraint is of the form X=Y where Y is an ordered value ** in an outer loop, then make sure the sort order of Y matches the ** sort order required for X. */ if( isMatch && isEq>=2 && isEq!=pOBItem->sortOrder+2 ){ testcase( isEq==2 ); testcase( isEq==3 ); break; } }else{ isEq = 0; /* "X=expr" places no ordering constraints on X */ } } if( !isMatch ){ if( isEq==0 ){ break; }else{ continue; } }else if( isEq!=1 ){ if( sortOrder==2 ){ sortOrder = termSortOrder; }else if( termSortOrder!=sortOrder ){ break; } } j++; pOBItem++; if( iColumn<0 ){ seenRowid = 1; break; }else if( pTab->aCol[iColumn].notNull==0 && isEq!=1 ){ testcase( isEq==0 ); testcase( isEq==2 ); testcase( isEq==3 ); uniqueNotNull = 0; } } /* If we have not found at least one ORDER BY term that matches the ** index, then show no progress. */ if( pOBItem==&pOrderBy->a[nPriorSat] ) return nPriorSat; /* Return the necessary scan order back to the caller */ *pbRev = sortOrder & 1; /* If there was an "ORDER BY rowid" term that matched, or it is only ** possible for a single row from this table to match, then skip over ** any additional ORDER BY terms dealing with this table. */ if( seenRowid || (uniqueNotNull && i>=pIdx->nColumn) ){ /* Advance j over additional ORDER BY terms associated with base */ WhereMaskSet *pMS = p->pWC->pMaskSet; Bitmask m = ~getMask(pMS, base); while( j<nTerm && (exprTableUsage(pMS, pOrderBy->a[j].pExpr)&m)==0 ){ j++; } } return j; } /* ** Find the best query plan for accessing a particular table. Write the ** best query plan and its cost into the p->cost. ** ** The lowest cost plan wins. The cost is an estimate of the amount of ** CPU and disk I/O needed to process the requested result. |
︙ | ︙ | |||
3032 3033 3034 3035 3036 3037 3038 | pIdx = 0; } /* Loop over all indices looking for the best one to use */ for(; pProbe; pIdx=pProbe=pProbe->pNext){ const tRowcnt * const aiRowEst = pProbe->aiRowEst; | | < < < < | | 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 | pIdx = 0; } /* Loop over all indices looking for the best one to use */ for(; pProbe; pIdx=pProbe=pProbe->pNext){ const tRowcnt * const aiRowEst = pProbe->aiRowEst; WhereCost pc; /* Cost of using pProbe */ double log10N = (double)1; /* base-10 logarithm of nRow (inexact) */ /* The following variables are populated based on the properties of ** index being evaluated. They are then used to determine the expected ** cost and number of rows returned. ** ** pc.plan.nEq: ** Number of equality terms that can be implemented using the index. ** In other words, the number of initial fields in the index that ** are used in == or IN or NOT NULL constraints of the WHERE clause. ** ** nInMul: ** The "in-multiplier". This is an estimate of how many seek operations ** SQLite must perform on the index in question. For example, if the |
︙ | ︙ | |||
3068 3069 3070 3071 3072 3073 3074 | ** ** nInMul is set to 1. ** ** If there exists a WHERE term of the form "x IN (SELECT ...)", then ** the sub-select is assumed to return 25 rows for the purposes of ** determining nInMul. ** | < < < < | < < | > > > > > > > > > > > | | | > | | | | | | < < < | | | | | | | | > > | > > | | | | | | | | | | | | | < | > > > | > | | | | | | | | | | | | | > | | > | > | | | | | | | | > > > > | | | | | > > | | | | | | | | | < | | | < | | > | | < < < | < | < < | 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 | ** ** nInMul is set to 1. ** ** If there exists a WHERE term of the form "x IN (SELECT ...)", then ** the sub-select is assumed to return 25 rows for the purposes of ** determining nInMul. ** ** bInEst: ** Set to true if there was at least one "x IN (SELECT ...)" term used ** in determining the value of nInMul. Note that the RHS of the ** IN operator must be a SELECT, not a value list, for this variable ** to be true. ** ** rangeDiv: ** An estimate of a divisor by which to reduce the search space due ** to inequality constraints. In the absence of sqlite_stat3 ANALYZE ** data, a single inequality reduces the search space to 1/4rd its ** original size (rangeDiv==4). Two inequalities reduce the search ** space to 1/16th of its original size (rangeDiv==16). ** ** bSort: ** Boolean. True if there is an ORDER BY clause that will require an ** external sort (i.e. scanning the index being evaluated will not ** correctly order records). ** ** bDist: ** Boolean. True if there is a DISTINCT clause that will require an ** external btree. ** ** bLookup: ** Boolean. True if a table lookup is required for each index entry ** visited. In other words, true if this is not a covering index. ** This is always false for the rowid primary key index of a table. ** For other indexes, it is true unless all the columns of the table ** used by the SELECT statement are present in the index (such an ** index is sometimes described as a covering index). ** For example, given the index on (a, b), the second of the following ** two queries requires table b-tree lookups in order to find the value ** of column c, but the first does not because columns a and b are ** both available in the index. ** ** SELECT a, b FROM tbl WHERE a = 1; ** SELECT a, b, c FROM tbl WHERE a = 1; */ int bInEst = 0; /* True if "x IN (SELECT...)" seen */ int nInMul = 1; /* Number of distinct equalities to lookup */ double rangeDiv = (double)1; /* Estimated reduction in search space */ int nBound = 0; /* Number of range constraints seen */ int bSort; /* True if external sort required */ int bDist; /* True if index cannot help with DISTINCT */ int bLookup = 0; /* True if not a covering index */ int nPriorSat; /* ORDER BY terms satisfied by outer loops */ int nOrderBy; /* Number of ORDER BY terms */ WhereTerm *pTerm; /* A single term of the WHERE clause */ #ifdef SQLITE_ENABLE_STAT3 WhereTerm *pFirstTerm = 0; /* First term matching the index */ #endif WHERETRACE(( " %s(%s):\n", pSrc->pTab->zName, (pIdx ? pIdx->zName : "ipk") )); memset(&pc, 0, sizeof(pc)); nOrderBy = p->pOrderBy ? p->pOrderBy->nExpr : 0; if( p->i ){ nPriorSat = pc.plan.nOBSat = p->aLevel[p->i-1].plan.nOBSat; bSort = nPriorSat<nOrderBy; bDist = 0; }else{ nPriorSat = pc.plan.nOBSat = 0; bSort = nOrderBy>0; bDist = p->pDistinct!=0; } /* Determine the values of pc.plan.nEq and nInMul */ for(pc.plan.nEq=0; pc.plan.nEq<pProbe->nColumn; pc.plan.nEq++){ int j = pProbe->aiColumn[pc.plan.nEq]; pTerm = findTerm(pWC, iCur, j, p->notReady, eqTermMask, pIdx); if( pTerm==0 ) break; pc.plan.wsFlags |= (WHERE_COLUMN_EQ|WHERE_ROWID_EQ); testcase( pTerm->pWC!=pWC ); if( pTerm->eOperator & WO_IN ){ Expr *pExpr = pTerm->pExpr; pc.plan.wsFlags |= WHERE_COLUMN_IN; if( ExprHasProperty(pExpr, EP_xIsSelect) ){ /* "x IN (SELECT ...)": Assume the SELECT returns 25 rows */ nInMul *= 25; bInEst = 1; }else if( ALWAYS(pExpr->x.pList && pExpr->x.pList->nExpr) ){ /* "x IN (value, value, ...)" */ nInMul *= pExpr->x.pList->nExpr; } }else if( pTerm->eOperator & WO_ISNULL ){ pc.plan.wsFlags |= WHERE_COLUMN_NULL; } #ifdef SQLITE_ENABLE_STAT3 if( pc.plan.nEq==0 && pProbe->aSample ) pFirstTerm = pTerm; #endif pc.used |= pTerm->prereqRight; } /* If the index being considered is UNIQUE, and there is an equality ** constraint for all columns in the index, then this search will find ** at most a single row. In this case set the WHERE_UNIQUE flag to ** indicate this to the caller. ** ** Otherwise, if the search may find more than one row, test to see if ** there is a range constraint on indexed column (pc.plan.nEq+1) that can be ** optimized using the index. */ if( pc.plan.nEq==pProbe->nColumn && pProbe->onError!=OE_None ){ testcase( pc.plan.wsFlags & WHERE_COLUMN_IN ); testcase( pc.plan.wsFlags & WHERE_COLUMN_NULL ); if( (pc.plan.wsFlags & (WHERE_COLUMN_IN|WHERE_COLUMN_NULL))==0 ){ pc.plan.wsFlags |= WHERE_UNIQUE; if( p->i==0 || (p->aLevel[p->i-1].plan.wsFlags & WHERE_ALL_UNIQUE)!=0 ){ pc.plan.wsFlags |= WHERE_ALL_UNIQUE; } } }else if( pProbe->bUnordered==0 ){ int j; j = (pc.plan.nEq==pProbe->nColumn ? -1 : pProbe->aiColumn[pc.plan.nEq]); if( findTerm(pWC, iCur, j, p->notReady, WO_LT|WO_LE|WO_GT|WO_GE, pIdx) ){ WhereTerm *pTop, *pBtm; pTop = findTerm(pWC, iCur, j, p->notReady, WO_LT|WO_LE, pIdx); pBtm = findTerm(pWC, iCur, j, p->notReady, WO_GT|WO_GE, pIdx); whereRangeScanEst(pParse, pProbe, pc.plan.nEq, pBtm, pTop, &rangeDiv); if( pTop ){ nBound = 1; pc.plan.wsFlags |= WHERE_TOP_LIMIT; pc.used |= pTop->prereqRight; testcase( pTop->pWC!=pWC ); } if( pBtm ){ nBound++; pc.plan.wsFlags |= WHERE_BTM_LIMIT; pc.used |= pBtm->prereqRight; testcase( pBtm->pWC!=pWC ); } pc.plan.wsFlags |= (WHERE_COLUMN_RANGE|WHERE_ROWID_RANGE); } } /* If there is an ORDER BY clause and the index being considered will ** naturally scan rows in the required order, set the appropriate flags ** in pc.plan.wsFlags. Otherwise, if there is an ORDER BY clause but ** the index will scan rows in a different order, set the bSort ** variable. */ if( bSort && (pSrc->jointype & JT_LEFT)==0 ){ int bRev = 2; WHERETRACE((" --> before isSortingIndex: nPriorSat=%d\n",nPriorSat)); pc.plan.nOBSat = isSortingIndex(p, pProbe, iCur, &bRev); WHERETRACE((" --> after isSortingIndex: bRev=%d nOBSat=%d\n", bRev, pc.plan.nOBSat)); if( nPriorSat<pc.plan.nOBSat || (pc.plan.wsFlags & WHERE_UNIQUE)!=0 ){ pc.plan.wsFlags |= WHERE_ORDERED; } if( nOrderBy==pc.plan.nOBSat ){ bSort = 0; pc.plan.wsFlags |= WHERE_ROWID_RANGE|WHERE_COLUMN_RANGE; } if( bRev & 1 ) pc.plan.wsFlags |= WHERE_REVERSE; } /* If there is a DISTINCT qualifier and this index will scan rows in ** order of the DISTINCT expressions, clear bDist and set the appropriate ** flags in pc.plan.wsFlags. */ if( bDist && isDistinctIndex(pParse, pWC, pProbe, iCur, p->pDistinct, pc.plan.nEq) && (pc.plan.wsFlags & WHERE_COLUMN_IN)==0 ){ bDist = 0; pc.plan.wsFlags |= WHERE_ROWID_RANGE|WHERE_COLUMN_RANGE|WHERE_DISTINCT; } /* If currently calculating the cost of using an index (not the IPK ** index), determine if all required column data may be obtained without ** using the main table (i.e. if the index is a covering ** index for this query). If it is, set the WHERE_IDX_ONLY flag in ** pc.plan.wsFlags. Otherwise, set the bLookup variable to true. */ if( pIdx ){ Bitmask m = pSrc->colUsed; int j; for(j=0; j<pIdx->nColumn; j++){ int x = pIdx->aiColumn[j]; if( x<BMS-1 ){ m &= ~(((Bitmask)1)<<x); } } if( m==0 ){ pc.plan.wsFlags |= WHERE_IDX_ONLY; }else{ bLookup = 1; } } /* ** Estimate the number of rows of output. For an "x IN (SELECT...)" ** constraint, do not let the estimate exceed half the rows in the table. */ pc.plan.nRow = (double)(aiRowEst[pc.plan.nEq] * nInMul); if( bInEst && pc.plan.nRow*2>aiRowEst[0] ){ pc.plan.nRow = aiRowEst[0]/2; nInMul = (int)(pc.plan.nRow / aiRowEst[pc.plan.nEq]); } #ifdef SQLITE_ENABLE_STAT3 /* If the constraint is of the form x=VALUE or x IN (E1,E2,...) ** and we do not think that values of x are unique and if histogram ** data is available for column x, then it might be possible ** to get a better estimate on the number of rows based on ** VALUE and how common that value is according to the histogram. */ if( pc.plan.nRow>(double)1 && pc.plan.nEq==1 && pFirstTerm!=0 && aiRowEst[1]>1 ){ assert( (pFirstTerm->eOperator & (WO_EQ|WO_ISNULL|WO_IN))!=0 ); if( pFirstTerm->eOperator & (WO_EQ|WO_ISNULL) ){ testcase( pFirstTerm->eOperator==WO_EQ ); testcase( pFirstTerm->eOperator==WO_ISNULL ); whereEqualScanEst(pParse, pProbe, pFirstTerm->pExpr->pRight, &pc.plan.nRow); }else if( bInEst==0 ){ assert( pFirstTerm->eOperator==WO_IN ); whereInScanEst(pParse, pProbe, pFirstTerm->pExpr->x.pList, &pc.plan.nRow); } } #endif /* SQLITE_ENABLE_STAT3 */ /* Adjust the number of output rows and downward to reflect rows ** that are excluded by range constraints. */ pc.plan.nRow = pc.plan.nRow/rangeDiv; if( pc.plan.nRow<1 ) pc.plan.nRow = 1; /* Experiments run on real SQLite databases show that the time needed ** to do a binary search to locate a row in a table or index is roughly ** log10(N) times the time to move from one row to the next row within ** a table or index. The actual times can vary, with the size of ** records being an important factor. Both moves and searches are ** slower with larger records, presumably because fewer records fit ** on one page and hence more pages have to be fetched. ** ** The ANALYZE command and the sqlite_stat1 and sqlite_stat3 tables do ** not give us data on the relative sizes of table and index records. ** So this computation assumes table records are about twice as big ** as index records */ if( (pc.plan.wsFlags&~(WHERE_REVERSE|WHERE_ORDERED))==WHERE_IDX_ONLY && (pWC->wctrlFlags & WHERE_ONEPASS_DESIRED)==0 && sqlite3GlobalConfig.bUseCis && OptimizationEnabled(pParse->db, SQLITE_CoverIdxScan) ){ /* This index is not useful for indexing, but it is a covering index. ** A full-scan of the index might be a little faster than a full-scan ** of the table, so give this case a cost slightly less than a table ** scan. */ pc.rCost = aiRowEst[0]*3 + pProbe->nColumn; pc.plan.wsFlags |= WHERE_COVER_SCAN|WHERE_COLUMN_RANGE; }else if( (pc.plan.wsFlags & WHERE_NOT_FULLSCAN)==0 ){ /* The cost of a full table scan is a number of move operations equal ** to the number of rows in the table. ** ** We add an additional 4x penalty to full table scans. This causes ** the cost function to err on the side of choosing an index over ** choosing a full scan. This 4x full-scan penalty is an arguable ** decision and one which we expect to revisit in the future. But ** it seems to be working well enough at the moment. */ pc.rCost = aiRowEst[0]*4; pc.plan.wsFlags &= ~WHERE_IDX_ONLY; if( pIdx ){ pc.plan.wsFlags &= ~WHERE_ORDERED; pc.plan.nOBSat = nPriorSat; } }else{ log10N = estLog(aiRowEst[0]); pc.rCost = pc.plan.nRow; if( pIdx ){ if( bLookup ){ /* For an index lookup followed by a table lookup: ** nInMul index searches to find the start of each index range ** + nRow steps through the index ** + nRow table searches to lookup the table entry using the rowid */ pc.rCost += (nInMul + pc.plan.nRow)*log10N; }else{ /* For a covering index: ** nInMul index searches to find the initial entry ** + nRow steps through the index */ pc.rCost += nInMul*log10N; } }else{ /* For a rowid primary key lookup: ** nInMult table searches to find the initial entry for each range ** + nRow steps through the table */ pc.rCost += nInMul*log10N; } } /* Add in the estimated cost of sorting the result. Actual experimental ** measurements of sorting performance in SQLite show that sorting time ** adds C*N*log10(N) to the cost, where N is the number of rows to be ** sorted and C is a factor between 1.95 and 4.3. We will split the ** difference and select C of 3.0. */ if( bSort ){ double m = estLog(pc.plan.nRow*(nOrderBy - pc.plan.nOBSat)/nOrderBy); m *= (double)(pc.plan.nOBSat ? 2 : 3); pc.rCost += pc.plan.nRow*m; } if( bDist ){ pc.rCost += pc.plan.nRow*estLog(pc.plan.nRow)*3; } /**** Cost of using this index has now been computed ****/ /* If there are additional constraints on this table that cannot ** be used with the current index, but which might lower the number ** of output rows, adjust the nRow value accordingly. This only ** matters if the current index is the least costly, so do not bother ** with this step if we already know this index will not be chosen. ** Also, never reduce the output row count below 2 using this step. ** ** It is critical that the notValid mask be used here instead of ** the notReady mask. When computing an "optimal" index, the notReady ** mask will only have one bit set - the bit for the current table. ** The notValid mask, on the other hand, always has all bits set for ** tables that are not in outer loops. If notReady is used here instead ** of notValid, then a optimal index that depends on inner joins loops ** might be selected even when there exists an optimal index that has ** no such dependency. */ if( pc.plan.nRow>2 && pc.rCost<=p->cost.rCost ){ int k; /* Loop counter */ int nSkipEq = pc.plan.nEq; /* Number of == constraints to skip */ int nSkipRange = nBound; /* Number of < constraints to skip */ Bitmask thisTab; /* Bitmap for pSrc */ thisTab = getMask(pWC->pMaskSet, iCur); for(pTerm=pWC->a, k=pWC->nTerm; pc.plan.nRow>2 && k; k--, pTerm++){ if( pTerm->wtFlags & TERM_VIRTUAL ) continue; if( (pTerm->prereqAll & p->notValid)!=thisTab ) continue; if( pTerm->eOperator & (WO_EQ|WO_IN|WO_ISNULL) ){ if( nSkipEq ){ /* Ignore the first pc.plan.nEq equality matches since the index ** has already accounted for these */ nSkipEq--; }else{ /* Assume each additional equality match reduces the result ** set size by a factor of 10 */ pc.plan.nRow /= 10; } }else if( pTerm->eOperator & (WO_LT|WO_LE|WO_GT|WO_GE) ){ if( nSkipRange ){ /* Ignore the first nSkipRange range constraints since the index ** has already accounted for these */ nSkipRange--; }else{ /* Assume each additional range constraint reduces the result ** set size by a factor of 3. Indexed range constraints reduce ** the search space by a larger factor: 4. We make indexed range ** more selective intentionally because of the subjective ** observation that indexed range constraints really are more ** selective in practice, on average. */ pc.plan.nRow /= 3; } }else if( pTerm->eOperator!=WO_NOOP ){ /* Any other expression lowers the output row count by half */ pc.plan.nRow /= 2; } } if( pc.plan.nRow<2 ) pc.plan.nRow = 2; } WHERETRACE(( " nEq=%d nInMul=%d rangeDiv=%d bSort=%d bLookup=%d wsFlags=0x%08x\n" " notReady=0x%llx log10N=%.1f nRow=%.1f cost=%.1f\n" " used=0x%llx nOBSat=%d\n", pc.plan.nEq, nInMul, (int)rangeDiv, bSort, bLookup, pc.plan.wsFlags, p->notReady, log10N, pc.plan.nRow, pc.rCost, pc.used, pc.plan.nOBSat )); /* If this index is the best we have seen so far, then record this ** index and its cost in the p->cost structure. */ if( (!pIdx || pc.plan.wsFlags) && compareCost(&pc, &p->cost) ){ p->cost = pc; p->cost.plan.wsFlags &= wsFlagMask; p->cost.plan.u.pIdx = pIdx; } /* If there was an INDEXED BY clause, then only that one index is ** considered. */ if( pSrc->pIndex ) break; |
︙ | ︙ | |||
3463 3464 3465 3466 3467 3468 3469 | ** in. This is used for application testing, to help find cases ** where application behaviour depends on the (undefined) order that ** SQLite outputs rows in in the absence of an ORDER BY clause. */ if( !p->pOrderBy && pParse->db->flags & SQLITE_ReverseOrder ){ p->cost.plan.wsFlags |= WHERE_REVERSE; } | | | < | < | 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522 3523 3524 | ** in. This is used for application testing, to help find cases ** where application behaviour depends on the (undefined) order that ** SQLite outputs rows in in the absence of an ORDER BY clause. */ if( !p->pOrderBy && pParse->db->flags & SQLITE_ReverseOrder ){ p->cost.plan.wsFlags |= WHERE_REVERSE; } assert( p->pOrderBy || (p->cost.plan.wsFlags&WHERE_ORDERED)==0 ); assert( p->cost.plan.u.pIdx==0 || (p->cost.plan.wsFlags&WHERE_ROWID_EQ)==0 ); assert( pSrc->pIndex==0 || p->cost.plan.u.pIdx==0 || p->cost.plan.u.pIdx==pSrc->pIndex ); WHERETRACE((" best index is: %s\n", p->cost.plan.u.pIdx ? p->cost.plan.u.pIdx->zName : "ipk")); bestOrClauseIndex(p); bestAutomaticIndex(p); p->cost.plan.wsFlags |= eqTermMask; } /* |
︙ | ︙ | |||
4201 4202 4203 4204 4205 4206 4207 | ** query, then the caller will only allow the loop to run for ** a single iteration. This means that the first row returned ** should not have a NULL value stored in 'x'. If column 'x' is ** the first one after the nEq equality constraints in the index, ** this requires some special handling. */ if( (wctrlFlags&WHERE_ORDERBY_MIN)!=0 | | | 4238 4239 4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252 | ** query, then the caller will only allow the loop to run for ** a single iteration. This means that the first row returned ** should not have a NULL value stored in 'x'. If column 'x' is ** the first one after the nEq equality constraints in the index, ** this requires some special handling. */ if( (wctrlFlags&WHERE_ORDERBY_MIN)!=0 && (pLevel->plan.wsFlags&WHERE_ORDERED) && (pIdx->nColumn>nEq) ){ /* assert( pOrderBy->nExpr==1 ); */ /* assert( pOrderBy->a[0].pExpr->iColumn==pIdx->aiColumn[nEq] ); */ isMinQuery = 1; nExtraReg = 1; } |
︙ | ︙ | |||
5022 5023 5024 5025 5026 5027 5028 | if( (m & sWBI.notValid)==0 ){ if( j==iFrom ) iFrom++; continue; } sWBI.notReady = (isOptimal ? m : sWBI.notValid); if( sWBI.pSrc->pIndex==0 ) nUnconstrained++; | | | | 5059 5060 5061 5062 5063 5064 5065 5066 5067 5068 5069 5070 5071 5072 5073 5074 | if( (m & sWBI.notValid)==0 ){ if( j==iFrom ) iFrom++; continue; } sWBI.notReady = (isOptimal ? m : sWBI.notValid); if( sWBI.pSrc->pIndex==0 ) nUnconstrained++; WHERETRACE((" === trying table %d (%s) with isOptimal=%d ===\n", j, sWBI.pSrc->pTab->zName, isOptimal)); assert( sWBI.pSrc->pTab ); #ifndef SQLITE_OMIT_VIRTUALTABLE if( IsVirtual(sWBI.pSrc->pTab) ){ sWBI.ppIdxInfo = &pWInfo->a[j].pIdxInfo; bestVirtualIndex(&sWBI); }else #endif |
︙ | ︙ | |||
5064 5065 5066 5067 5068 5069 5070 | ** index specified by its INDEXED BY clause. This rule ensures ** that a best-so-far is always selected even if an impossible ** combination of INDEXED BY clauses are given. The error ** will be detected and relayed back to the application later. ** The NEVER() comes about because rule (2) above prevents ** An indexable full-table-scan from reaching rule (3). ** | | | | < < | | > | | | | > | < < < > | 5101 5102 5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142 5143 5144 5145 5146 5147 5148 5149 5150 | ** index specified by its INDEXED BY clause. This rule ensures ** that a best-so-far is always selected even if an impossible ** combination of INDEXED BY clauses are given. The error ** will be detected and relayed back to the application later. ** The NEVER() comes about because rule (2) above prevents ** An indexable full-table-scan from reaching rule (3). ** ** (4) The plan cost must be lower than prior plans, where "cost" ** is defined by the compareCost() function above. */ if( (sWBI.cost.used&sWBI.notValid)==0 /* (1) */ && (bestJ<0 || (notIndexed&m)!=0 /* (2) */ || (bestPlan.plan.wsFlags & WHERE_NOT_FULLSCAN)==0 || (sWBI.cost.plan.wsFlags & WHERE_NOT_FULLSCAN)!=0) && (nUnconstrained==0 || sWBI.pSrc->pIndex==0 /* (3) */ || NEVER((sWBI.cost.plan.wsFlags & WHERE_NOT_FULLSCAN)!=0)) && (bestJ<0 || compareCost(&sWBI.cost, &bestPlan)) /* (4) */ ){ WHERETRACE((" === table %d (%s) is best so far\n" " cost=%.1f, nRow=%.1f, nOBSat=%d, wsFlags=%08x\n", j, sWBI.pSrc->pTab->zName, sWBI.cost.rCost, sWBI.cost.plan.nRow, sWBI.cost.plan.nOBSat, sWBI.cost.plan.wsFlags)); bestPlan = sWBI.cost; bestJ = j; } if( doNotReorder ) break; } } assert( bestJ>=0 ); assert( sWBI.notValid & getMask(pMaskSet, pTabList->a[bestJ].iCursor) ); WHERETRACE(("*** Optimizer selects table %d (%s) for loop %d with:\n" " cost=%.1f, nRow=%.1f, nOBSat=%d, wsFlags=0x%08x\n", bestJ, pTabList->a[bestJ].pTab->zName, pLevel-pWInfo->a, bestPlan.rCost, bestPlan.plan.nRow, bestPlan.plan.nOBSat, bestPlan.plan.wsFlags)); if( (bestPlan.plan.wsFlags & WHERE_DISTINCT)!=0 ){ assert( pWInfo->eDistinct==0 ); pWInfo->eDistinct = WHERE_DISTINCT_ORDERED; } andFlags &= bestPlan.plan.wsFlags; pLevel->plan = bestPlan.plan; pLevel->iTabCur = pTabList->a[bestJ].iCursor; testcase( bestPlan.plan.wsFlags & WHERE_INDEXED ); testcase( bestPlan.plan.wsFlags & WHERE_TEMP_INDEX ); if( bestPlan.plan.wsFlags & (WHERE_INDEXED|WHERE_TEMP_INDEX) ){ if( (wctrlFlags & WHERE_ONETABLE_ONLY) && (bestPlan.plan.wsFlags & WHERE_TEMP_INDEX)==0 ){ pLevel->iIdxCur = iIdxCur; |
︙ | ︙ | |||
5143 5144 5145 5146 5147 5148 5149 5150 5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 | } } } WHERETRACE(("*** Optimizer Finished ***\n")); if( pParse->nErr || db->mallocFailed ){ goto whereBeginError; } /* If the total query only selects a single row, then the ORDER BY ** clause is irrelevant. */ if( (andFlags & WHERE_UNIQUE)!=0 && pOrderBy ){ pWInfo->nOBSat = pOrderBy->nExpr; } /* If the caller is an UPDATE or DELETE statement that is requesting ** to use a one-pass algorithm, determine if this is appropriate. ** The one-pass algorithm only works if the WHERE clause constraints ** the statement to update a single row. | > > > > > > > | 5178 5179 5180 5181 5182 5183 5184 5185 5186 5187 5188 5189 5190 5191 5192 5193 5194 5195 5196 5197 5198 5199 5200 5201 5202 5203 | } } } WHERETRACE(("*** Optimizer Finished ***\n")); if( pParse->nErr || db->mallocFailed ){ goto whereBeginError; } if( nTabList ){ pLevel--; pWInfo->nOBSat = pLevel->plan.nOBSat; }else{ pWInfo->nOBSat = 0; } /* If the total query only selects a single row, then the ORDER BY ** clause is irrelevant. */ if( (andFlags & WHERE_UNIQUE)!=0 && pOrderBy ){ assert( nTabList==0 || (pLevel->plan.wsFlags & WHERE_ALL_UNIQUE)!=0 ); pWInfo->nOBSat = pOrderBy->nExpr; } /* If the caller is an UPDATE or DELETE statement that is requesting ** to use a one-pass algorithm, determine if this is appropriate. ** The one-pass algorithm only works if the WHERE clause constraints ** the statement to update a single row. |
︙ | ︙ | |||
5175 5176 5177 5178 5179 5180 5181 | for(ii=0, pLevel=pWInfo->a; ii<nTabList; ii++, pLevel++){ Table *pTab; /* Table to open */ int iDb; /* Index of database containing table/index */ struct SrcList_item *pTabItem; pTabItem = &pTabList->a[pLevel->iFrom]; pTab = pTabItem->pTab; | < | 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226 5227 5228 5229 5230 | for(ii=0, pLevel=pWInfo->a; ii<nTabList; ii++, pLevel++){ Table *pTab; /* Table to open */ int iDb; /* Index of database containing table/index */ struct SrcList_item *pTabItem; pTabItem = &pTabList->a[pLevel->iFrom]; pTab = pTabItem->pTab; pWInfo->nRowOut *= pLevel->plan.nRow; iDb = sqlite3SchemaToIndex(db, pTab->pSchema); if( (pTab->tabFlags & TF_Ephemeral)!=0 || pTab->pSelect ){ /* Do nothing */ }else #ifndef SQLITE_OMIT_VIRTUALTABLE if( (pLevel->plan.wsFlags & WHERE_VIRTUALTABLE)!=0 ){ |
︙ | ︙ |
Changes to test/bigfile.test.
︙ | ︙ | |||
12 13 14 15 16 17 18 19 20 21 22 23 24 25 | # focus of this script testing the ability of SQLite to handle database # files larger than 4GB. # # $Id: bigfile.test,v 1.12 2009/03/05 04:27:08 shane Exp $ # if {[file exists skip-big-file]} return set testdir [file dirname $argv0] source $testdir/tester.tcl # Do not use a codec for this file, as the database is manipulated using # external methods (the [fake_big_file] and [hexio_write] commands). # | > | 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 | # focus of this script testing the ability of SQLite to handle database # files larger than 4GB. # # $Id: bigfile.test,v 1.12 2009/03/05 04:27:08 shane Exp $ # if {[file exists skip-big-file]} return if {$tcl_platform(os)=="Darwin"} return set testdir [file dirname $argv0] source $testdir/tester.tcl # Do not use a codec for this file, as the database is manipulated using # external methods (the [fake_big_file] and [hexio_write] commands). # |
︙ | ︙ |
Changes to test/bigfile2.test.
︙ | ︙ | |||
10 11 12 13 14 15 16 17 18 19 20 21 22 23 | #*********************************************************************** # This file implements regression tests for SQLite library. The # focus of this script testing the ability of SQLite to handle database # files larger than 4GB. # if {[file exists skip-big-file]} return set testdir [file dirname $argv0] source $testdir/tester.tcl set testprefix bigfile2 # Create a small database. # | > | 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 | #*********************************************************************** # This file implements regression tests for SQLite library. The # focus of this script testing the ability of SQLite to handle database # files larger than 4GB. # if {[file exists skip-big-file]} return if {$tcl_platform(os)=="Darwin"} return set testdir [file dirname $argv0] source $testdir/tester.tcl set testprefix bigfile2 # Create a small database. # |
︙ | ︙ |
Changes to test/fuzzer1.test.
︙ | ︙ | |||
1860 1861 1862 1863 1864 1865 1866 | INSERT INTO x5_rules VALUES(0, 'a', '0.1.2.3.4.5.6.7.8.9.a', 1); DROP TABLE x5; CREATE VIRTUAL TABLE x5 USING fuzzer(x5_rules); SELECT length(word) FROM x5 WHERE word MATCH 'a' LIMIT 50; } {1 21 41 61 81} finish_test | < < | 1860 1861 1862 1863 1864 1865 1866 | INSERT INTO x5_rules VALUES(0, 'a', '0.1.2.3.4.5.6.7.8.9.a', 1); DROP TABLE x5; CREATE VIRTUAL TABLE x5 USING fuzzer(x5_rules); SELECT length(word) FROM x5 WHERE word MATCH 'a' LIMIT 50; } {1 21 41 61 81} finish_test |
Changes to test/orderby1.test.
︙ | ︙ | |||
110 111 112 113 114 115 116 | } } {three-a three-c two-a two-b one-a one-c} ;# verify same order after sorting do_test 1.4c { db eval { EXPLAIN QUERY PLAN SELECT name FROM album JOIN track USING (aid) ORDER BY title DESC, tn } | | | | 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 | } } {three-a three-c two-a two-b one-a one-c} ;# verify same order after sorting do_test 1.4c { db eval { EXPLAIN QUERY PLAN SELECT name FROM album JOIN track USING (aid) ORDER BY title DESC, tn } } {~/ORDER BY/} ;# optimized out do_test 1.5a { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY title, tn DESC } } {one-c one-a two-b two-a three-c three-a} do_test 1.5b { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY +title, +tn DESC } } {one-c one-a two-b two-a three-c three-a} ;# verify same order after sorting do_test 1.5c { db eval { EXPLAIN QUERY PLAN SELECT name FROM album JOIN track USING (aid) ORDER BY title, tn DESC } } {~/ORDER BY/} ;# optimized out do_test 1.6a { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY title DESC, tn DESC } } {three-c three-a two-b two-a one-c one-a} do_test 1.6b { |
︙ | ︙ | |||
166 167 168 169 170 171 172 | CREATE INDEX album_i1 ON album(title, aid); CREATE TABLE track( aid INTEGER NOT NULL REFERENCES album, tn INTEGER NOT NULL, name TEXT, UNIQUE(aid, tn) ); | | | | | | | | > > > > > > > > > > > > | 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 | CREATE INDEX album_i1 ON album(title, aid); CREATE TABLE track( aid INTEGER NOT NULL REFERENCES album, tn INTEGER NOT NULL, name TEXT, UNIQUE(aid, tn) ); INSERT INTO album VALUES(1, '1-one'), (20, '2-two'), (3, '3-three'); INSERT INTO track VALUES (1, 1, 'one-a'), (20, 2, 'two-b'), (3, 3, 'three-c'), (1, 3, 'one-c'), (20, 1, 'two-a'), (3, 1, 'three-a'); COMMIT; } } {} do_test 2.1a { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY title, tn } } {one-a one-c two-a two-b three-a three-c} # Verify that the ORDER BY clause is optimized out # do_test 2.1b { db eval { EXPLAIN QUERY PLAN SELECT name FROM album JOIN track USING (aid) ORDER BY title, tn } } {~/ORDER BY/} ;# ORDER BY optimized out do_test 2.1c { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY title, aid, tn } } {one-a one-c two-a two-b three-a three-c} do_test 2.1d { db eval { EXPLAIN QUERY PLAN SELECT name FROM album JOIN track USING (aid) ORDER BY title, aid, tn } } {~/ORDER BY/} ;# ORDER BY optimized out # The same query with ORDER BY clause optimization disabled via + operators # should give exactly the same answer. # do_test 2.2a { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY +title, +tn } |
︙ | ︙ | |||
245 246 247 248 249 250 251 | } } {three-a three-c two-a two-b one-a one-c} ;# verify same order after sorting do_test 2.4c { db eval { EXPLAIN QUERY PLAN SELECT name FROM album JOIN track USING (aid) ORDER BY title DESC, tn } | | | | | 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 | } } {three-a three-c two-a two-b one-a one-c} ;# verify same order after sorting do_test 2.4c { db eval { EXPLAIN QUERY PLAN SELECT name FROM album JOIN track USING (aid) ORDER BY title DESC, tn } } {~/ORDER BY/} ;# optimized out do_test 2.5a { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY title, tn DESC } } {one-c one-a two-b two-a three-c three-a} do_test 2.5b { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY +title, +tn DESC } } {one-c one-a two-b two-a three-c three-a} ;# verify same order after sorting do_test 2.5c { db eval { EXPLAIN QUERY PLAN SELECT name FROM album JOIN track USING (aid) ORDER BY title, tn DESC } } {~/ORDER BY/} ;# optimized out do_test 2.6a { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY title DESC, tn DESC } } {three-c three-a two-b two-a one-c one-a} do_test 2.6b { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY +title DESC, +tn DESC } } {three-c three-a two-b two-a one-c one-a} ;# verify same order after sorting do_test 2.6c { db eval { EXPLAIN QUERY PLAN SELECT name FROM album JOIN track USING (aid) ORDER BY title DESC, tn DESC } } {~/ORDER BY/} ;# ORDER BY optimized out # Generate another test dataset, but this time using mixed ASC/DESC indices. # do_test 3.0 { db eval { BEGIN; |
︙ | ︙ | |||
380 381 382 383 384 385 386 | } } {one-a one-c two-a two-b three-a three-c} ;# verify same order after sorting do_test 3.4c { db eval { EXPLAIN QUERY PLAN SELECT name FROM album JOIN track USING (aid) ORDER BY title, tn } | | | | 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 | } } {one-a one-c two-a two-b three-a three-c} ;# verify same order after sorting do_test 3.4c { db eval { EXPLAIN QUERY PLAN SELECT name FROM album JOIN track USING (aid) ORDER BY title, tn } } {~/ORDER BY/} ;# optimized out do_test 3.5a { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY title DESC, tn DESC } } {three-c three-a two-b two-a one-c one-a} do_test 3.5b { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY +title DESC, +tn DESC } } {three-c three-a two-b two-a one-c one-a} ;# verify same order after sorting do_test 3.5c { db eval { EXPLAIN QUERY PLAN SELECT name FROM album JOIN track USING (aid) ORDER BY title DESC, tn DESC } } {~/ORDER BY/} ;# optimzed out do_test 3.6a { db eval { SELECT name FROM album JOIN track USING (aid) ORDER BY title DESC, tn } } {three-a three-c two-a two-b one-a one-c} |
︙ | ︙ |
Added test/orderby2.test.
> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 | # 2012 Sept 27 # # The author disclaims copyright to this source code. In place of # a legal notice, here is a blessing: # # 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 implements regression tests for SQLite library. The # focus of this file is testing that the optimizations that disable # ORDER BY clauses when the natural order of a query is correct. # set testdir [file dirname $argv0] source $testdir/tester.tcl set ::testprefix orderby2 # Generate test data for a join. Verify that the join gets the # correct answer. # do_test 1.0 { db eval { CREATE TABLE t1(a INTEGER PRIMARY KEY, b); INSERT INTO t1 VALUES(1,11), (2,22); CREATE TABLE t2(d, e, UNIQUE(d,e)); INSERT INTO t2 VALUES(10, 'ten'), (11,'eleven'), (12,'twelve'), (11, 'oneteen'); } } {} do_test 1.1a { db eval { SELECT e FROM t1, t2 WHERE a=1 AND d=b ORDER BY d, e; } } {eleven oneteen} do_test 1.1b { db eval { EXPLAIN QUERY PLAN SELECT e FROM t1, t2 WHERE a=1 AND d=b ORDER BY d, e; } } {~/ORDER BY/} do_test 1.2a { db eval { SELECT e FROM t1, t2 WHERE a=1 AND d=b ORDER BY e; } } {eleven oneteen} do_test 1.2b { db eval { EXPLAIN QUERY PLAN SELECT e FROM t1, t2 WHERE a=1 AND d=b ORDER BY e; } } {~/ORDER BY/} do_test 1.3a { db eval { SELECT e, b FROM t1, t2 WHERE a=1 ORDER BY d, e; } } {ten 11 eleven 11 oneteen 11 twelve 11} do_test 1.3b { db eval { EXPLAIN QUERY PLAN SELECT e, b FROM t1, t2 WHERE a=1 ORDER BY d, e; } } {~/ORDER BY/} # The following tests derived from TH3 test module cov1/where34.test # do_test 2.0 { db eval { CREATE TABLE t31(a,b); CREATE INDEX t31ab ON t31(a,b); CREATE TABLE t32(c,d); CREATE INDEX t32cd ON t32(c,d); CREATE TABLE t33(e,f); CREATE INDEX t33ef ON t33(e,f); CREATE TABLE t34(g,h); CREATE INDEX t34gh ON t34(g,h); INSERT INTO t31 VALUES(1,4), (2,3), (1,3); INSERT INTO t32 VALUES(4,5), (3,6), (3,7), (4,8); INSERT INTO t33 VALUES(5,9), (7,10), (6,11), (8,12), (8,13), (7,14); INSERT INTO t34 VALUES(11,20), (10,21), (12,22), (9,23), (13,24), (14,25), (12,26); SELECT a||','||c||','||e||','||g FROM t31, t32, t33, t34 WHERE c=b AND e=d AND g=f ORDER BY a ASC, c ASC, e DESC, g ASC; } } {1,3,7,10 1,3,7,14 1,3,6,11 1,4,8,12 1,4,8,12 1,4,8,13 1,4,5,9 2,3,7,10 2,3,7,14 2,3,6,11} do_test 2.1 { db eval { SELECT a||','||c||','||e||','||g FROM t31, t32, t33, t34 WHERE c=b AND e=d AND g=f ORDER BY +a ASC, +c ASC, +e DESC, +g ASC; } } {1,3,7,10 1,3,7,14 1,3,6,11 1,4,8,12 1,4,8,12 1,4,8,13 1,4,5,9 2,3,7,10 2,3,7,14 2,3,6,11} do_test 2.2 { db eval { SELECT a||','||c||','||e||','||g FROM t31, t32, t33, t34 WHERE c=b AND e=d AND g=f ORDER BY a ASC, c ASC, e ASC, g ASC; } } {1,3,6,11 1,3,7,10 1,3,7,14 1,4,5,9 1,4,8,12 1,4,8,12 1,4,8,13 2,3,6,11 2,3,7,10 2,3,7,14} do_test 2.3 { optimization_control db cover-idx-scan off db cache flush db eval { SELECT a||','||c||','||e||','||g FROM t31, t32, t33, t34 WHERE c=b AND e=d AND g=f ORDER BY a ASC, c ASC, e ASC, g ASC; } } {1,3,6,11 1,3,7,10 1,3,7,14 1,4,5,9 1,4,8,12 1,4,8,12 1,4,8,13 2,3,6,11 2,3,7,10 2,3,7,14} optimization_control db all on db cache flush finish_test |
Added test/shared9.test.
> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 | # 2012 October 5 # # The author disclaims copyright to this source code. In place of # a legal notice, here is a blessing: # # 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. # #*********************************************************************** # # The tests in this file are intended to show if two connections attach # to the same shared cache using different database names, views and # virtual tables may still be accessed. # set testdir [file dirname $argv0] source $testdir/tester.tcl source $testdir/lock_common.tcl set testprefix shared9 ifcapable !view||!trigger { finish_test return } db close set enable_shared_cache [sqlite3_enable_shared_cache 1] sqlite3 db1 test.db sqlite3 db2 test.db forcedelete test.db2 do_test 1.1 { db1 eval { ATTACH 'test.db2' AS 'fred'; CREATE TABLE fred.t1(a, b, c); CREATE VIEW fred.v1 AS SELECT * FROM t1; CREATE TABLE fred.t2(a, b); CREATE TABLE fred.t3(a, b); CREATE TRIGGER fred.trig AFTER INSERT ON t2 BEGIN DELETE FROM t3; INSERT INTO t3 SELECT * FROM t2; END; INSERT INTO t2 VALUES(1, 2); SELECT * FROM t3; } } {1 2} do_test 1.2 { db2 eval "ATTACH 'test.db2' AS 'jones'" } {} do_test 1.3 { db2 eval "SELECT * FROM v1" } {} do_test 1.4 { db2 eval "INSERT INTO t2 VALUES(3, 4)" } {} ifcapable fts3 { do_test 1.5 { db1 eval { CREATE VIRTUAL TABLE fred.t4 USING fts4; INSERT INTO t4 VALUES('hello world'); } } {} do_test 1.6 { db2 eval { INSERT INTO t4 VALUES('shared cache'); SELECT * FROM t4 WHERE t4 MATCH 'hello'; } } {{hello world}} do_test 1.7 { db1 eval { SELECT * FROM t4 WHERE t4 MATCH 'c*'; } } {{shared cache}} } db1 close db2 close #------------------------------------------------------------------------- # The following tests attempt to find a similar problem with collation # sequence names - pointers to database handle specific allocations leaking # into schema objects and being used after the original handle has been # closed. # forcedelete test.db test.db2 sqlite3 db1 test.db sqlite3 db2 test.db foreach x {collate1 collate2 collate3} { proc $x {a b} { string compare $a $b } db1 collate $x $x db2 collate $x $x } do_test 2.1 { db1 eval { CREATE TABLE t1(a, b, c COLLATE collate1); CREATE INDEX i1 ON t1(a COLLATE collate2, c, b); } } {} do_test 2.2 { db1 close db2 eval "INSERT INTO t1 VALUES('abc', 'def', 'ghi')" } {} db2 close #------------------------------------------------------------------------- # At one point, the following would cause a collation sequence belonging # to connection [db1] to be invoked by a call to [db2 eval]. Which is a # problem if [db1] has already been closed. # forcedelete test.db test.db2 sqlite3 db1 test.db sqlite3 db2 test.db proc mycollate_db1 {a b} {set ::invoked_mycollate_db1 1 ; string compare $a $b} proc mycollate_db2 {a b} {string compare $a $b} db1 collate mycollate mycollate_db1 db2 collate mycollate mycollate_db2 do_test 2.3 { set ::invoked_mycollate_db1 0 db1 eval { CREATE TABLE t1(a COLLATE mycollate, CHECK (a IN ('one', 'two', 'three'))); INSERT INTO t1 VALUES('one'); } db1 close set ::invoked_mycollate_db1 } {1} do_test 2.4 { set ::invoked_mycollate_db1 0 db2 eval { INSERT INTO t1 VALUES('two'); } db2 close set ::invoked_mycollate_db1 } {0} #------------------------------------------------------------------------- # This test verifies that a bug causing a busy-handler belonging to one # shared-cache connection to be executed as a result of an sqlite3_step() # on another has been fixed. # forcedelete test.db test.db2 sqlite3 db1 test.db sqlite3 db2 test.db proc busyhandler {handle args} { set ::busyhandler_invoked_for $handle return 1 } db1 busy [list busyhandler db1] db2 busy [list busyhandler db2] do_test 3.1 { db1 eval { BEGIN; CREATE TABLE t1(a, b); CREATE TABLE t2(a, b); INSERT INTO t1 VALUES(1, 2); INSERT INTO t2 VALUES(1, 2); } # Keep this next COMMIT as a separate statement. This ensures that COMMIT # has already been compiled and loaded into the tcl interface statement # cache when it is attempted below. db1 eval COMMIT db1 eval { BEGIN; INSERT INTO t1 VALUES(3, 4); } } {} do_test 3.2 { set ::tf [launch_testfixture] testfixture $::tf { sqlite3 db test.db db eval { BEGIN; SELECT * FROM t1; } } } {1 2} do_test 3.3 { db2 eval { SELECT * FROM t2 } } {1 2} do_test 3.4 { list [catch { db1 eval COMMIT } msg] $msg } {1 {database is locked}} # At one point the following would fail, showing that the busy-handler # belonging to [db2] was invoked instead. do_test 3.5 { set ::busyhandler_invoked_for } {db1} do_test 3.6 { close $::tf db1 eval COMMIT } {} db1 close db2 close sqlite3_enable_shared_cache $::enable_shared_cache finish_test |
Changes to test/tclsqlite.test.
︙ | ︙ | |||
315 316 317 318 319 320 321 322 323 324 325 | # modify and reset the NULL representation # do_test tcl-8.1 { db nullvalue NaN execsql {INSERT INTO t1 VALUES(30,NULL)} db eval {SELECT * FROM t1 WHERE b IS NULL} } {30 NaN} do_test tcl-8.2 { db nullvalue NULL db nullvalue } {NULL} | > > > > > | > > > > | 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 | # modify and reset the NULL representation # do_test tcl-8.1 { db nullvalue NaN execsql {INSERT INTO t1 VALUES(30,NULL)} db eval {SELECT * FROM t1 WHERE b IS NULL} } {30 NaN} proc concatFunc args {return [join $args {}]} do_test tcl-8.2 { db function concat concatFunc db eval {SELECT concat('a', b, 'z') FROM t1 WHERE b is NULL} } {aNaNz} do_test tcl-8.3 { db nullvalue NULL db nullvalue } {NULL} do_test tcl-8.4 { db nullvalue {} db eval {SELECT * FROM t1 WHERE b IS NULL} } {30 {}} do_test tcl-8.5 { db function concat concatFunc db eval {SELECT concat('a', b, 'z') FROM t1 WHERE b is NULL} } {az} # Test the return type of user-defined functions # do_test tcl-9.1 { db function ret_str {return "hi"} execsql {SELECT typeof(ret_str())} } {text} |
︙ | ︙ |
Changes to test/trigger1.test.
1 2 3 4 5 6 7 8 9 10 11 12 13 | # The author disclaims copyright to this source code. In place of # a legal notice, here is a blessing: # # 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 tests creating and dropping triggers, and interaction thereof # with the database COMMIT/ROLLBACK logic. # # 1. CREATE and DROP TRIGGER tests | | | | | | | | | | | | | | | | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 | # The author disclaims copyright to this source code. In place of # a legal notice, here is a blessing: # # 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 tests creating and dropping triggers, and interaction thereof # with the database COMMIT/ROLLBACK logic. # # 1. CREATE and DROP TRIGGER tests # trigger1-1.1: Error if table does not exist # trigger1-1.2: Error if trigger already exists # trigger1-1.3: Created triggers are deleted if the transaction is rolled back # trigger1-1.4: DROP TRIGGER removes trigger # trigger1-1.5: Dropped triggers are restored if the transaction is rolled back # trigger1-1.6: Error if dropped trigger doesn't exist # trigger1-1.7: Dropping the table automatically drops all triggers # trigger1-1.8: A trigger created on a TEMP table is not inserted into sqlite_master # trigger1-1.9: Ensure that we cannot create a trigger on sqlite_master # trigger1-1.10: # trigger1-1.11: # trigger1-1.12: Ensure that INSTEAD OF triggers cannot be created on tables # trigger1-1.13: Ensure that AFTER triggers cannot be created on views # trigger1-1.14: Ensure that BEFORE triggers cannot be created on views # set testdir [file dirname $argv0] source $testdir/tester.tcl ifcapable !trigger||!compound { finish_test return |
︙ | ︙ | |||
206 207 208 209 210 211 212 | do_test trigger1-1.12 { catchsql { create table t1(a,b); create trigger t1t instead of update on t1 for each row begin delete from t1 WHERE a=old.a+2; end; } | | | | | 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 | do_test trigger1-1.12 { catchsql { create table t1(a,b); create trigger t1t instead of update on t1 for each row begin delete from t1 WHERE a=old.a+2; end; } } {1 {cannot create INSTEAD OF trigger on table: t1}} ifcapable view { # Ensure that we cannot create BEFORE triggers on views do_test trigger1-1.13 { catchsql { create view v1 as select * from t1; create trigger v1t before update on v1 for each row begin delete from t1 WHERE a=old.a+2; end; } } {1 {cannot create BEFORE trigger on view: v1}} # Ensure that we cannot create AFTER triggers on views do_test trigger1-1.14 { catchsql { drop view v1; create view v1 as select * from t1; create trigger v1t AFTER update on v1 for each row begin delete from t1 WHERE a=old.a+2; end; } } {1 {cannot create AFTER trigger on view: v1}} } ;# ifcapable view # Check for memory leaks in the trigger parser # do_test trigger1-2.1 { catchsql { CREATE TRIGGER r1 AFTER INSERT ON t1 BEGIN |
︙ | ︙ | |||
261 262 263 264 265 266 267 | } catchsql { CREATE TRIGGER r1 AFTER INSERT ON t1 BEGIN INSERT INTO t2 VALUES(NEW.a,NEW.b); END; } } {0 {}} | | | | | | | | | | 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 | } catchsql { CREATE TRIGGER r1 AFTER INSERT ON t1 BEGIN INSERT INTO t2 VALUES(NEW.a,NEW.b); END; } } {0 {}} do_test trigger1-3.2 { catchsql { INSERT INTO t1 VALUES(1,2); SELECT * FROM t2; } } {1 {no such table: main.t2}} do_test trigger1-3.3 { db close set rc [catch {sqlite3 db test.db} err] if {$rc} {lappend rc $err} set rc } {0} do_test trigger1-3.4 { catchsql { INSERT INTO t1 VALUES(1,2); SELECT * FROM t2; } } {1 {no such table: main.t2}} do_test trigger1-3.5 { catchsql { CREATE TEMP TABLE t2(x,y); INSERT INTO t1 VALUES(1,2); SELECT * FROM t2; } } {1 {no such table: main.t2}} do_test trigger1-3.6.1 { catchsql { DROP TRIGGER r1; CREATE TEMP TRIGGER r1 AFTER INSERT ON t1 BEGIN INSERT INTO t2 VALUES(NEW.a,NEW.b), (NEW.b*100, NEW.a*100); END; INSERT INTO t1 VALUES(1,2); SELECT * FROM t2; } } {0 {1 2 200 100}} do_test trigger1-3.6.2 { catchsql { DROP TRIGGER r1; DELETE FROM t1; DELETE FROM t2; CREATE TEMP TRIGGER r1 AFTER INSERT ON t1 BEGIN INSERT INTO t2 VALUES(NEW.a,NEW.b); END; INSERT INTO t1 VALUES(1,2); SELECT * FROM t2; } } {0 {1 2}} do_test trigger1-3.7 { execsql { DROP TABLE t2; CREATE TABLE t2(x,y); SELECT * FROM t2; } } {} # There are two versions of trigger1-3.8 and trigger1-3.9. One that uses # compound SELECT statements, and another that does not. ifcapable compound { do_test trigger1-3.8 { execsql { INSERT INTO t1 VALUES(3,4); SELECT * FROM t1 UNION ALL SELECT * FROM t2; } |
︙ | ︙ | |||
442 443 444 445 446 447 448 | } {3 4 7 8} do_test trigger1-6.8 { db close sqlite3 db test.db execsql {SELECT * FROM t2} } {3 4 7 8} | | | 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 | } {3 4 7 8} do_test trigger1-6.8 { db close sqlite3 db test.db execsql {SELECT * FROM t2} } {3 4 7 8} integrity_check trigger1-7.1 # Check to make sure the name of a trigger can be quoted so that keywords # can be used as trigger names. Ticket #468 # do_test trigger1-8.1 { execsql { CREATE TRIGGER 'trigger' AFTER INSERT ON t2 BEGIN SELECT 1; END; |
︙ | ︙ | |||
487 488 489 490 491 492 493 | SELECT name FROM sqlite_master WHERE type='trigger'; } } {} ifcapable conflict { # Make sure REPLACE works inside of triggers. # | | | 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 | SELECT name FROM sqlite_master WHERE type='trigger'; } } {} ifcapable conflict { # Make sure REPLACE works inside of triggers. # # There are two versions of trigger1-9.1 and trigger1-9.2. One that uses # compound SELECT statements, and another that does not. ifcapable compound { do_test trigger1-9.1 { execsql { CREATE TABLE t3(a,b); CREATE TABLE t4(x UNIQUE, b); CREATE TRIGGER r34 AFTER INSERT ON t3 BEGIN |
︙ | ︙ | |||
608 609 610 611 612 613 614 | } } {} do_test trigger1-10.8 { execsql { SELECT * FROM insert_log; } } {main 11 12 13 temp 14 15 16 aux 17 18 19} | | | 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 | } } {} do_test trigger1-10.8 { execsql { SELECT * FROM insert_log; } } {main 11 12 13 temp 14 15 16 aux 17 18 19} do_test trigger1-10.9 { # Drop and re-create the insert_log table in a different database. Note # that we can change the column names because the trigger programs don't # use them explicitly. execsql { DROP TABLE insert_log; CREATE TABLE aux.insert_log(db, d, e, f); } |
︙ | ︙ |
Changes to test/where.test.
︙ | ︙ | |||
1084 1085 1086 1087 1088 1089 1090 | CREATE TABLE t8(a INTEGER PRIMARY KEY, b TEXT UNIQUE); INSERT INTO t8 VALUES(1,'one'); INSERT INTO t8 VALUES(4,'four'); } cksort { SELECT x.a || '/' || y.a FROM t8 x, t8 y ORDER BY x.a, y.b } | | | | 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 | CREATE TABLE t8(a INTEGER PRIMARY KEY, b TEXT UNIQUE); INSERT INTO t8 VALUES(1,'one'); INSERT INTO t8 VALUES(4,'four'); } cksort { SELECT x.a || '/' || y.a FROM t8 x, t8 y ORDER BY x.a, y.b } } {1/4 1/1 4/4 4/1 nosort} do_test where-14.2 { cksort { SELECT x.a || '/' || y.a FROM t8 x, t8 y ORDER BY x.a, y.b DESC } } {1/1 1/4 4/1 4/4 nosort} do_test where-14.3 { cksort { SELECT x.a || '/' || y.a FROM t8 x, t8 y ORDER BY x.a, x.b } } {1/4 1/1 4/4 4/1 nosort} do_test where-14.4 { cksort { |
︙ | ︙ |
Changes to tool/build-all-msvc.bat.
︙ | ︙ | |||
138 139 140 141 142 143 144 | REM FOR %%P IN (%PLATFORMS%) DO ( REM REM NOTE: Using the MSVC platform name, lookup the simpler platform name to REM be used for the name of the platform-specific binary directory via REM the environment variables setup earlier. REM | | | 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 | REM FOR %%P IN (%PLATFORMS%) DO ( REM REM NOTE: Using the MSVC platform name, lookup the simpler platform name to REM be used for the name of the platform-specific binary directory via REM the environment variables setup earlier. REM CALL :fn_CopyVariable %%P_NAME PLATFORMNAME REM REM NOTE: This is the inner loop. There should be exactly one iteration. REM This loop is necessary because the PlatformName environment REM variable was set above and that value is needed by some of the REM commands contained in the inner loop. If these commands were REM directly contained in the outer loop, the PlatformName environment |
︙ | ︙ | |||
226 227 228 229 230 231 232 | REM WindowsSdkDir environment variable because that location does REM not actually contain the necessary library files for x86. REM This must be done for each iteration because it relies upon REM the WindowsSdkDir environment variable being set by the batch REM file used to setup the MSVC environment. REM IF DEFINED SET_NSDKLIBPATH ( | | | 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 | REM WindowsSdkDir environment variable because that location does REM not actually contain the necessary library files for x86. REM This must be done for each iteration because it relies upon REM the WindowsSdkDir environment variable being set by the batch REM file used to setup the MSVC environment. REM IF DEFINED SET_NSDKLIBPATH ( CALL :fn_CopyVariable WindowsSdkDir NSDKLIBPATH CALL :fn_AppendVariable NSDKLIBPATH \lib\win8\um\x86 ) REM REM NOTE: Unless prevented from doing so, invoke NMAKE with the MSVC REM makefile to clean any stale build output from previous REM iterations of this loop and/or previous runs of this batch |
︙ | ︙ | |||
335 336 337 338 339 340 341 | VERIFY > NUL GOTO :EOF :fn_SetErrorLevel VERIFY MAYBE 2> NUL GOTO :EOF | | | 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 | VERIFY > NUL GOTO :EOF :fn_SetErrorLevel VERIFY MAYBE 2> NUL GOTO :EOF :fn_CopyVariable SETLOCAL IF NOT DEFINED %1 GOTO :EOF IF "%2" == "" GOTO :EOF SET __ECHO_CMD=ECHO %%%1%% FOR /F "delims=" %%V IN ('%__ECHO_CMD%') DO ( SET VALUE=%%V ) |
︙ | ︙ |