SQLite

Check-in [11bd66e090]
Login

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

Overview
Comment:Merge the btreeNext() assertion bug fix from trunk.
Downloads: Tarball | ZIP archive
Timelines: family | ancestors | descendants | both | in-scan-vs-index
Files: files | file ages | folders
SHA3-256: 11bd66e09035039fbfc4c5631196962dd3f2e6634f5123a7b2fec17aeaaab07b
User & Date: drh 2018-06-08 19:54:07.050
Context
2018-06-08
21:21
Only choose to scan an IN operator rather than use an index if we have real STAT1 data to suggest it is advantageous. (Closed-Leaf check-in: 30e874661d user: drh tags: in-scan-vs-index)
19:54
Merge the btreeNext() assertion bug fix from trunk. (check-in: 11bd66e090 user: drh tags: in-scan-vs-index)
19:13
Fix an assert() that can be false for a corrupt database and a strange query that uses a recursive SQL function to delete content from a corrupt database file while it is being queried. (check-in: 99057383ac user: drh tags: trunk)
18:22
Consider doing a partial table scan to fulfill an IN operator rather than using an index. Try to pick the plan with the lowest cost. (check-in: 1fa40a78fe user: drh tags: in-scan-vs-index)
Changes
Unified Diff Ignore Whitespace Patch
Changes to src/btree.c.
5585
5586
5587
5588
5589
5590
5591
5592









5593
5594
5595
5596
5597
5598
5599
      }
      pCur->skipNext = 0;
    }
  }

  pPage = pCur->pPage;
  idx = ++pCur->ix;
  assert( pPage->isInit );










  /* If the database file is corrupt, it is possible for the value of idx 
  ** to be invalid here. This can only occur if a second cursor modifies
  ** the page while cursor pCur is holding a reference to it. Which can
  ** only happen if the database is corrupt in such a way as to link the
  ** page into more than one b-tree structure. */
  testcase( idx>pPage->nCell );







|
>
>
>
>
>
>
>
>
>







5585
5586
5587
5588
5589
5590
5591
5592
5593
5594
5595
5596
5597
5598
5599
5600
5601
5602
5603
5604
5605
5606
5607
5608
      }
      pCur->skipNext = 0;
    }
  }

  pPage = pCur->pPage;
  idx = ++pCur->ix;
  if( !pPage->isInit ){
    /* The only known way for this to happen is for there to be a
    ** recursive SQL function that does a DELETE operation as part of a
    ** SELECT which deletes content out from under an active cursor
    ** in a corrupt database file where the table being DELETE-ed from
    ** has pages in common with the table being queried.  See TH3
    ** module cov1/btree78.test testcase 220 (2018-06-08) for an
    ** example. */
    return SQLITE_CORRUPT_BKPT;
  }

  /* If the database file is corrupt, it is possible for the value of idx 
  ** to be invalid here. This can only occur if a second cursor modifies
  ** the page while cursor pCur is holding a reference to it. Which can
  ** only happen if the database is corrupt in such a way as to link the
  ** page into more than one b-tree structure. */
  testcase( idx>pPage->nCell );