Many hyperlinks are disabled.
Use anonymous login
to enable hyperlinks.
Overview
Comment: | Add sub-transactions to btree module. |
---|---|
Downloads: | Tarball | ZIP archive |
Timelines: | family | ancestors | descendants | both | trunk |
Files: | files | file ages | folders |
SHA1: |
9e00823074163c12c9cef8ea61d29136 |
User & Date: | dan 2013-10-30 18:37:28.976 |
Context
2013-10-30
| ||
19:57 | Btree fixes related to multiple client tests. check-in: 58f7282211 user: dan tags: trunk | |
18:37 | Add sub-transactions to btree module. check-in: 9e00823074 user: dan tags: trunk | |
2013-10-29
| ||
17:46 | Add code to prevent database writers from overwriting portions of the log that might be required by present or future database readers or recoverers. check-in: 407a82adbf user: dan tags: trunk | |
Changes
Changes to lsm-test/README.
︙ | ︙ | |||
29 30 31 32 33 34 35 36 37 38 39 | clients are still running) whereas lsmtest2.c tests recovery from a system or power failure. lsmtest9.c: More data tests. These focus on testing that calling lsm_work(nMerge=1) to compact the database does not corrupt it. In other words, that databases containing block-redirects can be read and written. | > | 29 30 31 32 33 34 35 36 37 38 39 40 | clients are still running) whereas lsmtest2.c tests recovery from a system or power failure. lsmtest9.c: More data tests. These focus on testing that calling lsm_work(nMerge=1) to compact the database does not corrupt it. In other words, that databases containing block-redirects can be read and written. |
Changes to src/bt.h.
︙ | ︙ | |||
58 59 60 61 62 63 64 | ** Begin(): If (L>=iLevel), this is a no-op. Otherwise, open the read and/or ** write transactions required to set L to iLevel. ** ** Commit(): If (L<=iLevel), this is a no-op. Otherwise, close as many ** transactions as required to set L to iLevel. Commit any write ** transactions closed by this action. ** | | | | < < < | | | < | 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 | ** Begin(): If (L>=iLevel), this is a no-op. Otherwise, open the read and/or ** write transactions required to set L to iLevel. ** ** Commit(): If (L<=iLevel), this is a no-op. Otherwise, close as many ** transactions as required to set L to iLevel. Commit any write ** transactions closed by this action. ** ** Rollback(): If (L<=iLevel), this is a no-op. Otherwise, rollback and ** close as many transactions as required to set L to iLevel. ** Then, if iLevel is greater than or equal to 2, rollback ** (but do not close) the innermost remaining sub-transaction. ** ** SAVEPOINT one; -- Begin(2) ** SAVEPOINT two; -- Begin(3) ** ROLLBACK TO one; -- Rollback(2) ** SAVEPOINT two; -- Begin(3) ** SAVEPOINT three; -- Begin(4) ** ROLLBACK TO three; -- Rollback(5) ** RELEASE three; -- Commit(3) ** RELEASE one; -- Commit(1) ** ** The TransactionLevel() method returns the current transaction level. */ int sqlite4BtBegin(bt_db*, int iLevel); int sqlite4BtCommit(bt_db*, int iLevel); int sqlite4BtRollback(bt_db*, int iLevel); int sqlite4BtTransactionLevel(bt_db*); /* ** Open and close a database cursor. */ int sqlite4BtCsrOpen(bt_db*, int nExtra, bt_cursor **ppCsr); |
︙ | ︙ |
Changes to src/bt_pager.c.
︙ | ︙ | |||
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 | /* By default auto-checkpoint is 1000 */ #define BT_DEFAULT_AUTOCKPT 1000 typedef struct BtPageHash BtPageHash; typedef struct BtDbhdr BtDbhdr; /* ** Hash table for page references currently in memory. Manipulated using ** the following functions: ** ** btHashAdd() ** btHashRemove() ** btHashSearch() ** btHashClear() */ struct BtPageHash { int nEntry; /* Number of entries in hash table */ int nHash; /* Size of aHash[] array */ BtPage **aHash; /* Hash array */ }; struct BtPage { BtPager *pPager; /* Pager object that owns this page handle */ u32 pgno; /* Current page number */ int nRef; /* Number of references to this page */ int flags; /* Mask of BTPAGE_XXX flags */ u8 *aData; /* Pointer to current data */ BtPage *pNextHash; /* Next entry with same hash key */ BtPage *pNextDirty; /* Next page in BtPager.pDirty list */ }; /* ** Candidate values for BtPage.flags */ #define BT_PAGE_DIRTY 0x0001 /* Set for pages in BtPager.pDirty list */ | > > > > > > > > > > > > > > > > > > > > > | 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 | /* By default auto-checkpoint is 1000 */ #define BT_DEFAULT_AUTOCKPT 1000 typedef struct BtPageHash BtPageHash; typedef struct BtDbhdr BtDbhdr; typedef struct BtSavepoint BtSavepoint; typedef struct BtSavepage BtSavepage; /* ** Hash table for page references currently in memory. Manipulated using ** the following functions: ** ** btHashAdd() ** btHashRemove() ** btHashSearch() ** btHashClear() */ struct BtPageHash { int nEntry; /* Number of entries in hash table */ int nHash; /* Size of aHash[] array */ BtPage **aHash; /* Hash array */ }; /* ** There is one object of this type for each open sub-transaction. Stored ** in the BtPager.aSavepoint[] array. */ struct BtSavepoint { int iLevel; /* Transaction level value (always >2) */ BtSavepage *pSavepage; /* First in linked list of saved data */ }; struct BtSavepage { BtPage *pPg; /* Pointer to page this object belongs to */ u8 *aData; /* Saved data */ BtSavepage *pNext; /* Next saved page in the same savepoint */ int iSavepoint; /* Transaction number of savepoint */ BtSavepage *pNextSavepage; /* Next saved page on the same BtPage */ }; struct BtPage { BtPager *pPager; /* Pager object that owns this page handle */ u32 pgno; /* Current page number */ int nRef; /* Number of references to this page */ int flags; /* Mask of BTPAGE_XXX flags */ u8 *aData; /* Pointer to current data */ BtPage *pNextHash; /* Next entry with same hash key */ BtPage *pNextDirty; /* Next page in BtPager.pDirty list */ BtSavepage *pSavepage; /* List of saved page images */ }; /* ** Candidate values for BtPage.flags */ #define BT_PAGE_DIRTY 0x0001 /* Set for pages in BtPager.pDirty list */ |
︙ | ︙ | |||
80 81 82 83 84 85 86 87 88 89 90 91 92 93 | int nFile; /* Length of string zFile in bytes */ BtPageHash hash; /* Hash table */ BtPage *pDirty; /* List of all dirty pages */ int nTotalRef; /* Total number of outstanding page refs */ int nAutoCkpt; /* Auto-checkpoint when log is this large */ int bDoAutoCkpt; /* Do auto-checkpoint after next unlock */ BtDbhdr dbhdr; }; /************************************************************************** ** Interface to BtPageHash object. */ | > > | 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 | int nFile; /* Length of string zFile in bytes */ BtPageHash hash; /* Hash table */ BtPage *pDirty; /* List of all dirty pages */ int nTotalRef; /* Total number of outstanding page refs */ int nAutoCkpt; /* Auto-checkpoint when log is this large */ int bDoAutoCkpt; /* Do auto-checkpoint after next unlock */ BtDbhdr dbhdr; BtSavepoint *aSavepoint; /* Savepoint array */ int nSavepoint; /* Number of entries in aSavepoint array */ }; /************************************************************************** ** Interface to BtPageHash object. */ |
︙ | ︙ | |||
241 242 243 244 245 246 247 248 249 250 251 252 | static int btCleanup(BtLock *pLock){ BtPager *p = (BtPager*)pLock; int rc = sqlite4BtLogClose(p->pLog, 1); p->pLog = 0; return rc; } /* ** Close a pager database handle. */ int sqlite4BtPagerClose(BtPager *p){ | > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > < > > | 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 | static int btCleanup(BtLock *pLock){ BtPager *p = (BtPager*)pLock; int rc = sqlite4BtLogClose(p->pLog, 1); p->pLog = 0; return rc; } static int btOpenSavepoints(BtPager *p, int iLevel){ int rc = SQLITE4_OK; /* Return code */ int nReq = iLevel - 2; /* Required number of savepoints */ if( nReq>p->nSavepoint ){ BtSavepoint *aNew; int nByte = (nReq * sizeof(BtSavepoint)); aNew = sqlite4_realloc(p->btl.pEnv, p->aSavepoint, nByte); if( aNew ){ int i; for(i=p->nSavepoint; i<nReq; i++){ aNew[i].pSavepage = 0; aNew[i].iLevel = i+3; } p->aSavepoint = aNew; p->nSavepoint = nReq; }else{ rc = btErrorBkpt(SQLITE4_NOMEM); } } return rc; } static void btCloseSavepoints(BtPager *p, int iLevel){ int nReq = MAX(iLevel - 2, 0); if( nReq<p->nSavepoint ){ int i; for(i=p->nSavepoint-1; i>=nReq; i--){ BtSavepoint *pSavepoint = &p->aSavepoint[i]; BtSavepage *pSavepage; BtSavepage *pNext; /* Loop through each of the BtSavepage objects associated with this ** savepoint. Detach them from the BtPage objects and free all ** allocated memory. */ for(pSavepage=pSavepoint->pSavepage; pSavepage; pSavepage=pNext){ BtPage *pPg = pSavepage->pPg; pNext = pSavepage->pNext; /* Detach the BtSavepage from its BtPage object */ assert( pSavepage==pPg->pSavepage ); pPg->pSavepage = pSavepage->pNextSavepage; /* Free associated memory allocations */ sqlite4_free(p->btl.pEnv, pSavepage->aData); sqlite4_free(p->btl.pEnv, pSavepage); } } p->nSavepoint = nReq; } } /* ** If it has not already been added, add page pPg to the innermost ** savepoint. */ static int btAddToSavepoint(BtPager *p, BtPage *pPg){ int rc = SQLITE4_OK; BtSavepage *pSavepage; int iLevel = p->iTransactionLevel; /* Check if there is already a BtSavepage for this page associated with ** the innermost open savepoint. If so, this function becomes a no-op. */ for(pSavepage=pPg->pSavepage; pSavepage; pSavepage=pSavepage->pNextSavepage){ if( pSavepage->iSavepoint==iLevel ) break; } if( pSavepage==0 ){ pSavepage = sqlite4_malloc(p->btl.pEnv, sizeof(BtSavepage)); if( pSavepage==0 ){ rc = btErrorBkpt(SQLITE4_NOMEM); }else{ memset(pSavepage, 0, sizeof(BtSavepage)); } if( rc==SQLITE4_OK && (1 || (pPg->flags & BT_PAGE_DIRTY)) ){ pSavepage->aData = (u8*)sqlite4_malloc(p->btl.pEnv, p->dbhdr.pgsz); if( pSavepage->aData==0 ){ sqlite4_free(p->btl.pEnv, pSavepage); rc = btErrorBkpt(SQLITE4_NOMEM); }else{ memcpy(pSavepage->aData, pPg->aData, p->dbhdr.pgsz); } } if( rc==SQLITE4_OK ){ pSavepage->pPg = pPg; pSavepage->iSavepoint = iLevel; pSavepage->pNextSavepage = pPg->pSavepage; pPg->pSavepage = pSavepage; assert( p->aSavepoint[iLevel-3].iLevel==iLevel ); pSavepage->pNext = p->aSavepoint[iLevel-3].pSavepage; p->aSavepoint[iLevel-3].pSavepage = pSavepage; } } return rc; } /* ** Close a pager database handle. */ int sqlite4BtPagerClose(BtPager *p){ int rc; rc = sqlite4BtLockDisconnect((BtLock*)p, btCheckpoint, btCleanup); if( p->btl.pFd ){ p->btl.pVfs->xClose(p->btl.pFd); } btCloseSavepoints(p, 0); btPurgeCache(p); sqlite4BtLogClose(p->pLog, 0); sqlite4_free(p->btl.pEnv, p->zFile); sqlite4_free(p->btl.pEnv, p->aSavepoint); sqlite4_free(p->btl.pEnv, p); return rc; } /* ** Return a pointer to the nExtra bytes of space allocated by PagerNew(). */ |
︙ | ︙ | |||
382 383 384 385 386 387 388 389 390 391 392 393 394 395 | ** Commit the current write transaction to disk. */ static int btCommitTransaction(BtPager *p){ int rc = SQLITE4_OK; BtPage *pPg; BtPage *pNext; assert( p->iTransactionLevel>=2 ); for(pPg=p->pDirty; pPg; pPg=pNext){ pNext = pPg->pNextDirty; pPg->flags &= ~(BT_PAGE_DIRTY); pPg->pNextDirty = 0; if( rc==SQLITE4_OK ){ int bCommit = (pNext==0); | > > | 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 | ** Commit the current write transaction to disk. */ static int btCommitTransaction(BtPager *p){ int rc = SQLITE4_OK; BtPage *pPg; BtPage *pNext; assert( p->iTransactionLevel>=2 ); btCloseSavepoints(p, 2); for(pPg=p->pDirty; pPg; pPg=pNext){ pNext = pPg->pNextDirty; pPg->flags &= ~(BT_PAGE_DIRTY); pPg->pNextDirty = 0; if( rc==SQLITE4_OK ){ int bCommit = (pNext==0); |
︙ | ︙ | |||
450 451 452 453 454 455 456 | } *ppPg = pRet; return rc; } /* | | > < > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > | 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 | } *ppPg = pRet; return rc; } /* ** Roll back, but do not close, the current write transaction. */ static int btRollbackTransaction(BtPager *p){ int rc = SQLITE4_OK; BtPage *pPg; BtPage *pNext; assert( p->iTransactionLevel>=2 ); btCloseSavepoints(p, 2); /* Load the old db header from disk */ rc = p->btl.pVfs->xRead(p->btl.pFd, 0, &p->dbhdr, sizeof(p->dbhdr)); /* Loop through all dirty pages in memory. Discard those with nRef==0. ** Reload data from disk for any others. */ for(pPg=p->pDirty; pPg; pPg=pNext){ pNext = pPg->pNextDirty; pPg->flags &= ~(BT_PAGE_DIRTY); pPg->pNextDirty = 0; if( pPg->nRef==0 ){ btHashRemove(p, pPg); btFreePage(p, pPg); }else if( rc==SQLITE4_OK && (pPg->pgno<=p->dbhdr.nPg) ){ rc = btLoadPageData(p, pPg); } } p->pDirty = 0; return rc; } static int btRollbackSavepoints(BtPager *p, int iLevel){ int nReq = iLevel - 2; assert( nReq>=1 ); if( nReq<=p->nSavepoint ){ int i; for(i=p->nSavepoint-1; i>=(nReq-1); i--){ BtSavepoint *pSavepoint = &p->aSavepoint[i]; BtSavepage *pSavepage; BtSavepage *pNext; /* Loop through each of the BtSavepage objects associated with this ** savepoint. Detach them from the BtPage objects and free all ** allocated memory. */ for(pSavepage=pSavepoint->pSavepage; pSavepage; pSavepage=pNext){ BtPage *pPg = pSavepage->pPg; pNext = pSavepage->pNext; /* Detach the BtSavepage from its BtPage object */ assert( pSavepage==pPg->pSavepage ); pPg->pSavepage = pSavepage->pNextSavepage; /* Restore the page data */ memcpy(pPg->aData, pSavepage->aData, p->dbhdr.pgsz); /* Free associated memory allocations */ assert( pSavepage->aData ); /* temp */ sqlite4_free(p->btl.pEnv, pSavepage->aData); sqlite4_free(p->btl.pEnv, pSavepage); } pSavepoint->pSavepage = 0; } p->nSavepoint = nReq; } return SQLITE4_OK; } /* ** Transactions. These methods are more or less the same as their ** counterparts in bt.h. */ int sqlite4BtPagerBegin(BtPager *p, int iLevel){ int rc = SQLITE4_OK; assert( p->btl.pFd ); if( p->iTransactionLevel<iLevel ){ /* Open a read transaction if one is not already open */ if( p->iTransactionLevel==0 ){ rc = btOpenReadTransaction(p); } /* Open a write transaction if one is required */ if( rc==SQLITE4_OK && p->iTransactionLevel<2 && iLevel>=2 ){ rc = btOpenWriteTransaction(p); } /* Open any required savepoints */ if( rc==SQLITE4_OK ){ rc = btOpenSavepoints(p, iLevel); } /* If nothing has gone wrong, update BtPager.iTransactionLevel */ if( rc==SQLITE4_OK ){ assert( p->iTransactionLevel>=1 && iLevel>=p->iTransactionLevel ); p->iTransactionLevel = iLevel; } } return rc; |
︙ | ︙ | |||
531 532 533 534 535 536 537 | } int sqlite4BtPagerRollback(BtPager *p, int iLevel){ int rc = SQLITE4_OK; assert( p->btl.pFd ); if( p->iTransactionLevel>=iLevel ){ | | > > > > | > > > | > > | > | 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 | } int sqlite4BtPagerRollback(BtPager *p, int iLevel){ int rc = SQLITE4_OK; assert( p->btl.pFd ); if( p->iTransactionLevel>=iLevel ){ /* If a write transaction is open and the requested level is 2 or ** lower, rollback the outermost write transaction. If the requested ** level is less than 2, also drop the WRITER lock. */ if( p->iTransactionLevel>=2 ){ if( iLevel<=2 ){ rc = btRollbackTransaction(p); if( iLevel<2 ) sqlite4BtLogSnapshotEndWrite(p->pLog); }else{ rc = btRollbackSavepoints(p, iLevel); } } p->iTransactionLevel = iLevel; if( iLevel==0 ){ int rc2 = btCloseReadTransaction(p); if( rc==SQLITE4_OK ) rc2 = rc; } } return rc; } int sqlite4BtPagerRevert(BtPager *p, int iLevel){ |
︙ | ︙ | |||
621 622 623 624 625 626 627 628 629 630 631 632 | pRet->nRef++; } *ppPg = pRet; return rc; } int sqlite4BtPageWrite(BtPage *pPg){ if( (pPg->flags & BT_PAGE_DIRTY)==0 ){ pPg->flags |= BT_PAGE_DIRTY; pPg->pNextDirty = pPg->pPager->pDirty; pPg->pPager->pDirty = pPg; } | > > > > > > > > | | 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 | pRet->nRef++; } *ppPg = pRet; return rc; } int sqlite4BtPageWrite(BtPage *pPg){ int rc = SQLITE4_OK; BtPager *p = pPg->pPager; /* If there are savepoints open, add this page to the innermost savepoint */ if( p->nSavepoint>0 ){ rc = btAddToSavepoint(p, pPg); } if( (pPg->flags & BT_PAGE_DIRTY)==0 ){ pPg->flags |= BT_PAGE_DIRTY; pPg->pNextDirty = pPg->pPager->pDirty; pPg->pPager->pDirty = pPg; } return rc; } /* ** Decrement the refcount on page pPg. Also, indicate that page pPg is ** no longer in use. */ int sqlite4BtPageTrim(BtPage *pPg){ |
︙ | ︙ |
Changes to src/kvbt.c.
︙ | ︙ | |||
73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 | pKVStore->iTransLevel = sqlite4BtTransactionLevel(p->pDb); return rc; } /* ** Revert a transaction back to what it was when it started. */ static int btRevert(KVStore *pKVStore, int iLevel){ KVBt *p = (KVBt *)pKVStore; int rc; rc = sqlite4BtRevert(p->pDb, iLevel); pKVStore->iTransLevel = sqlite4BtTransactionLevel(p->pDb); return rc; } /* ** Implementation of the xReplace(X, aKey, nKey, aData, nData) method. */ static int btReplace( KVStore *pKVStore, const KVByteArray *aKey, KVSize nKey, | > > | 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 | pKVStore->iTransLevel = sqlite4BtTransactionLevel(p->pDb); return rc; } /* ** Revert a transaction back to what it was when it started. */ #if 0 static int btRevert(KVStore *pKVStore, int iLevel){ KVBt *p = (KVBt *)pKVStore; int rc; rc = sqlite4BtRevert(p->pDb, iLevel); pKVStore->iTransLevel = sqlite4BtTransactionLevel(p->pDb); return rc; } #endif /* ** Implementation of the xReplace(X, aKey, nKey, aData, nData) method. */ static int btReplace( KVStore *pKVStore, const KVByteArray *aKey, KVSize nKey, |
︙ | ︙ | |||
262 263 264 265 266 267 268 | btData, /* xData */ btReset, /* xReset */ btCloseCursor, /* xCloseCursor */ btBegin, /* xBegin */ btCommitPhaseOne, /* xCommitPhaseOne */ btCommitPhaseTwo, /* xCommitPhaseTwo */ btRollback, /* xRollback */ | | | 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 | btData, /* xData */ btReset, /* xReset */ btCloseCursor, /* xCloseCursor */ btBegin, /* xBegin */ btCommitPhaseOne, /* xCommitPhaseOne */ btCommitPhaseTwo, /* xCommitPhaseTwo */ btRollback, /* xRollback */ 0, /* xRevert */ btClose, /* xClose */ btControl, /* xControl */ btGetMeta, /* xGetMeta */ btPutMeta, /* xPutMeta */ btGetMethod /* xGetMethod */ }; |
︙ | ︙ |
Added test/simple4.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 | # 2013 September 24 # # 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 were used while developing the SQLite 4 code. # set testdir [file dirname $argv0] source $testdir/tester.tcl set testprefix simple3 db close forcedelete test.db sqlite4 db file:test.db?kv=bt do_execsql_test 1.0 { CREATE TABLE t1(a, b); INSERT INTO t1 VALUES(1, 2); INSERT INTO t1 VALUES(3, 4); } do_execsql_test 1.1 { BEGIN; INSERT INTO t1 VALUES(5, 6); SAVEPOINT one; INSERT INTO t1 VALUES(7, 8); } do_execsql_test 1.2 { SELECT * FROM t1 } {1 2 3 4 5 6 7 8} do_execsql_test 1.3 { ROLLBACK TO one; SELECT * FROM t1; } {1 2 3 4 5 6} do_execsql_test 1.4 { ROLLBACK; SELECT * FROM t1; } {1 2 3 4} finish_test |