/ Check-in [876162c7]
Login

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

Overview
Comment:Further enhancements to comments in pager.c.
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | experimental
Files: files | file ages | folders
SHA1:876162c7e036af1cb447409b685afc72c0061a32
User & Date: dan 2010-08-06 17:18:01
Context
2010-08-07
05:15
Add test case 'cgt_pager_1', intended for use with callgrind to detect performance regression in the pager module, to threadtest3.c. check-in: b5d46f1e user: dan tags: experimental
2010-08-06
17:18
Further enhancements to comments in pager.c. check-in: 876162c7 user: dan tags: experimental
14:37
Fix some problems with running test scripts with the inmemory_journal permutation. check-in: 7bd8ba08 user: dan tags: experimental
Changes
Hide Diffs Unified Diffs Ignore Whitespace Patch

Changes to src/pager.c.

493
494
495
496
497
498
499
500
501
502










503
504
505
506
507
508
509
510
511
512
513
....
1746
1747
1748
1749
1750
1751
1752
1753
1754

1755
1756
1757
1758
1759
1760
1761
....
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
....
1799
1800
1801
1802
1803
1804
1805
1806












1807
1808
1809
1810
1811
1812
1813
....
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
....
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
....
2010
2011
2012
2013
2014
2015
2016











2017
2018
2019
2020
2021
2022
2023
....
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
....
2859
2860
2861
2862
2863
2864
2865
2866
2867

2868





2869
2870
2871
2872






2873

2874







2875
2876
2877
2878
2879
2880
2881
....
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
....
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
....
4990
4991
4992
4993
4994
4995
4996
4997
4998
4999
5000
5001
5002
5003
5004
5005
5006
5007
5008
5009
5010
5011
....
5043
5044
5045
5046
5047
5048
5049
5050
5051
5052
5053
5054
5055
5056
5057
**   The flag is cleared as soon as the journal file is finalized (either
**   by PagerCommitPhaseTwo or PagerRollback). If an IO error prevents the
**   journal file from being successfully finalized, the setMaster flag
**   is cleared anyway (and the pager will move to ERROR state).
**
** doNotSpill, doNotSyncSpill
**
**   When enabled, cache spills are prohibited.  The doNotSpill variable
**   inhibits all cache spill and doNotSyncSpill inhibits those spills that
**   would require a journal sync.  The doNotSyncSpill is set and cleared 










**   by sqlite3PagerWrite() in order to prevent a journal sync from happening 
**   in between the journalling of two pages on the same sector.  The
**   doNotSpill value set to prevent pagerStress() from trying to use
**   the journal during a rollback.
**
** subjInMemory
**
**   This is a boolean variable. If true, then any required sub-journal
**   is opened as an in-memory journal file. If false, then in-memory
**   sub-journals are only used for in-memory pager files.
**
................................................................................
/*
** This routine ends a transaction. A transaction is usually ended by 
** either a COMMIT or a ROLLBACK operation. This routine may be called 
** after rollback of a hot-journal, or if an error occurs while opening
** the journal file or writing the very first journal-header of a
** database transaction.
** 
** If the pager is in PAGER_SHARED or PAGER_UNLOCK state when this
** routine is called, it is a no-op (returns SQLITE_OK).

**
** Otherwise, any active savepoints are released.
**
** If the journal file is open, then it is "finalized". Once a journal 
** file has been finalized it is not possible to use it to roll back a 
** transaction. Nor will it be considered to be a hot-journal by this
** or any other database connection. Exactly how a journal is finalized
................................................................................
**     The journal file is closed and deleted using sqlite3OsDelete().
**
**     If the pager is running in exclusive mode, this method of finalizing
**     the journal file is never used. Instead, if the journalMode is
**     DELETE and the pager is in exclusive mode, the method described under
**     journalMode==PERSIST is used instead.
**
** After the journal is finalized, if running in non-exclusive mode, the
** pager moves to PAGER_SHARED state (and downgrades the lock on the
** database file accordingly).
**
** If the pager is running in exclusive mode and is in PAGER_SYNCED state,
** it moves to PAGER_EXCLUSIVE. No locks are downgraded when running in
** exclusive mode.
**
** SQLITE_OK is returned if no error occurs. If an error occurs during
** any of the IO operations to finalize the journal file or unlock the
** database then the IO error code is returned to the user. If the 
** operation to finalize the journal file fails, then the code still
** tries to unlock the database file if not in exclusive mode. If the
** unlock operation fails as well, then the first error code related
................................................................................
** to the first error encountered (the journal finalization one) is
** returned.
*/
static int pager_end_transaction(Pager *pPager, int hasMaster){
  int rc = SQLITE_OK;      /* Error code from journal finalization operation */
  int rc2 = SQLITE_OK;     /* Error code from db file unlock operation */

  /* Do nothing if the pager does not have an open write transaction. */












  assert( assert_pager_state(pPager) );
  assert( pPager->eState!=PAGER_ERROR );
  if( pPager->eState<PAGER_WRITER_LOCKED && pPager->eLock<RESERVED_LOCK ){
    return SQLITE_OK;
  }

  releaseAllSavepoints(pPager);
................................................................................
  return (rc==SQLITE_OK?rc2:rc);
}

/*
** Execute a rollback if a transaction is active and unlock the 
** database file. 
**
** If the pager has already entered the error state, do not attempt 
** the rollback at this time. Instead, pager_unlock() is called. The
** call to pager_unlock() will discard all in-memory pages, unlock
** the database file and clear the error state. If this means that
** there is a hot-journal left in the file-system, the next connection
** to obtain a shared lock on the pager (which may be this one) will
** roll it back.
**
** If the pager has not already entered the error state, but an IO or
** malloc error occurs during a rollback, then this will itself cause 
** the pager to enter the error state. Which will be cleared by the
** call to pager_unlock(), as described above.
*/
static void pagerUnlockAndRollback(Pager *pPager){
  if( pPager->eState!=PAGER_ERROR && pPager->eState!=PAGER_OPEN ){
    assert( assert_pager_state(pPager) );
    if( pPager->eState>=PAGER_WRITER_LOCKED ){
      sqlite3BeginBenignMalloc();
      sqlite3PagerRollback(pPager);
      sqlite3EndBenignMalloc();
    }else if( pPager->eLock>=RESERVED_LOCK && !pPager->exclusiveMode ){
      assert( pPager->eState==PAGER_READER );
      pager_end_transaction(pPager, 0);
    }
  }
  pager_unlock(pPager);
}

................................................................................

/*
** Read a single page from either the journal file (if isMainJrnl==1) or
** from the sub-journal (if isMainJrnl==0) and playback that page.
** The page begins at offset *pOffset into the file. The *pOffset
** value is increased to the start of the next page in the journal.
**
** The isMainJrnl flag is true if this is the main rollback journal and
** false for the statement journal.  The main rollback journal uses
** checksums - the statement journal does not.
**
** If the page number of the page record read from the (sub-)journal file
** is greater than the current value of Pager.dbSize, then playback is
** skipped and SQLITE_OK is returned.
**
** If pDone is not NULL, then it is a record of pages that have already
** been played back.  If the page at *pOffset has already been played back
................................................................................
  assert( (isSavepnt&~1)==0 );       /* isSavepnt is 0 or 1 */
  assert( isMainJrnl || pDone );     /* pDone always used on sub-journals */
  assert( isSavepnt || pDone==0 );   /* pDone never used on non-savepoint */

  aData = pPager->pTmpSpace;
  assert( aData );         /* Temp storage must have already been allocated */
  assert( pagerUseWal(pPager)==0 || (!isMainJrnl && isSavepnt) );












  /* Read the page number and page data from the journal or sub-journal
  ** file. Return an error code to the caller if an IO error occurs.
  */
  jfd = isMainJrnl ? pPager->jfd : pPager->sjfd;
  rc = read32bits(jfd, *pOffset, &pgno);
  if( rc!=SQLITE_OK ) return rc;
................................................................................

  /* If this page has already been played by before during the current
  ** rollback, then don't bother to play it back again.
  */
  if( pDone && (rc = sqlite3BitvecSet(pDone, pgno))!=SQLITE_OK ){
    return rc;
  }
  assert( pPager->eState>=PAGER_WRITER_CACHEMOD 
       || (pPager->eState==PAGER_OPEN && isMainJrnl)
  );

  /* When playing back page 1, restore the nReserve setting
  */
  if( pgno==1 && pPager->nReserve!=((u8*)aData)[20] ){
    pPager->nReserve = ((u8*)aData)[20];
    pagerReportSize(pPager);
  }
................................................................................
  if( rc==SQLITE_OK && changed ){
    pager_reset(pPager);
  }

  return rc;
}


/*

** TODO: Description here.





*/
static int pagerPagecount(Pager *pPager, Pgno *pnPage){
  Pgno nPage;                     /* Value to return via *pnPage */







  assert( pPager->eState==PAGER_OPEN );

  nPage = sqlite3WalDbsize(pPager->pWal);







  if( nPage==0 ){
    i64 n = 0;                    /* Size of db file in bytes */
    assert( isOpen(pPager->fd) || pPager->tempFile );
    if( isOpen(pPager->fd) ){
      int rc = sqlite3OsFileSize(pPager->fd, &n);
      if( rc!=SQLITE_OK ){
        return rc;
................................................................................
    }
    nPage = (Pgno)(n / pPager->pageSize);
    if( nPage==0 && n>0 ){
      nPage = 1;
    }
  }

  /* If the current number of pages in the file is greater than the 
  ** configured maximum pager number, increase the allowed limit so
  ** that the file can be read.
  */
  if( nPage>pPager->mxPgno ){
    pPager->mxPgno = (Pgno)nPage;
  }

................................................................................
** If the database is not empty and the *-wal file exists, open the pager
** in WAL mode.  If the database is empty or if no *-wal file exists and
** if no error occurs, make sure Pager.journalMode is not set to
** PAGER_JOURNALMODE_WAL.
**
** Return SQLITE_OK or an error code.
**
** If the WAL file is opened, also open a snapshot (read transaction).
**
** The caller must hold a SHARED lock on the database file to call this
** function. Because an EXCLUSIVE lock on the db file is required to delete 
** a WAL on a none-empty database, this ensures there is no race condition 
** between the xAccess() below and an xDelete() being executed by some 
** other connection.
*/
static int pagerOpenWalIfPresent(Pager *pPager){
  int rc = SQLITE_OK;



  if( !pPager->tempFile ){
    int isWal;                    /* True if WAL file exists */
    Pgno nPage;                   /* Size of the database file */

    assert( pPager->eState==PAGER_OPEN );
    assert( pPager->eLock>=SHARED_LOCK || pPager->noReadlock );
    rc = pagerPagecount(pPager, &nPage);
    if( rc ) return rc;
    if( nPage==0 ){
      rc = sqlite3OsDelete(pPager->pVfs, pPager->zWal, 0);
      isWal = 0;
    }else{
      rc = sqlite3OsAccess(
................................................................................
          pPager->pVfs, pPager->zWal, SQLITE_ACCESS_EXISTS, &isWal
      );
    }
    if( rc==SQLITE_OK ){
      if( isWal ){
        testcase( sqlite3PcachePagecount(pPager->pPCache)==0 );
        rc = sqlite3PagerOpenWal(pPager, 0);
        if( rc==SQLITE_OK ){
          rc = pagerBeginReadTransaction(pPager);
        }
      }else if( pPager->journalMode==PAGER_JOURNALMODE_WAL ){
        pPager->journalMode = PAGER_JOURNALMODE_DELETE;
      }
    }
  }
  return rc;
}
................................................................................
** write-transaction has already been opened, this function is a no-op.
**
** If the exFlag argument is false, then acquire at least a RESERVED
** lock on the database file. If exFlag is true, then acquire at least
** an EXCLUSIVE lock. If such a lock is already held, no locking 
** functions need be called.
**
** If this is not a temporary or in-memory file and, the journal file is 
** opened if it has not been already. For a temporary file, the opening 
** of the journal file is deferred until there is an actual need to 
** write to the journal. TODO: Why handle temporary files differently?
**
** If the journal file is opened (or if it is already open), then a
** journal-header is written to the start of it.
**
** If the subjInMemory argument is non-zero, then any sub-journal opened
** within this transaction will be opened as an in-memory file. This
** has no effect if the sub-journal is already opened (as it may be when
** running in exclusive mode) or if the transaction does not require a
** sub-journal. If the subjInMemory argument is zero, then any required
** sub-journal is implemented in-memory if pPager is an in-memory database, 
** or using a temporary file otherwise.
................................................................................
      ** lock, but not when obtaining the RESERVED lock.
      */
      rc = pagerLockDb(pPager, RESERVED_LOCK);
      if( rc==SQLITE_OK && exFlag ){
        rc = pager_wait_on_lock(pPager, EXCLUSIVE_LOCK);
      }
    }


    if( rc==SQLITE_OK ){
      /* Change to WRITER_LOCKED state.
      **
      ** WAL mode sets Pager.eState to PAGER_WRITER_LOCKED or CACHEMOD
      ** when it has an open transaction, but never to DBMOD or FINISHED.
      ** This is because in those states the code to roll back savepoint 







|
|
|
>
>
>
>
>
>
>
>
>
>
|
|
<
<







 







|
|
>







 







|
|
|
<
<
<
<







 







|
>
>
>
>
>
>
>
>
>
>
>
>







 







|


|
|
|
|

|

|









|







 







|
|
<







 







>
>
>
>
>
>
>
>
>
>
>







 







<
<
<







 







<

>
|
>
>
>
>
>




>
>
>
>
>
>

>

>
>
>
>
>
>
>







 







|







 







<
<








>
>
>




<
<







 







<
<
<







 







<
<
<
<
<
<
<
<







 







<







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
520
521
....
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
....
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796




1797
1798
1799
1800
1801
1802
1803
....
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
....
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
....
1972
1973
1974
1975
1976
1977
1978
1979
1980

1981
1982
1983
1984
1985
1986
1987
....
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
....
2074
2075
2076
2077
2078
2079
2080



2081
2082
2083
2084
2085
2086
2087
....
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
....
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
....
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
2983
2984
2985
2986



2987
2988
2989
2990
2991
2992
2993
....
5029
5030
5031
5032
5033
5034
5035








5036
5037
5038
5039
5040
5041
5042
....
5074
5075
5076
5077
5078
5079
5080

5081
5082
5083
5084
5085
5086
5087
**   The flag is cleared as soon as the journal file is finalized (either
**   by PagerCommitPhaseTwo or PagerRollback). If an IO error prevents the
**   journal file from being successfully finalized, the setMaster flag
**   is cleared anyway (and the pager will move to ERROR state).
**
** doNotSpill, doNotSyncSpill
**
**   These two boolean variables control the behaviour of cache-spills
**   (calls made by the pcache module to the pagerStress() routine to
**   write cached data to the file-system in order to free up memory).
**
**   When doNotSpill is non-zero, writing to the database from pagerStress()
**   is disabled altogether. This is done in a very obscure case that
**   comes up during savepoint rollback that requires the pcache module
**   to allocate a new page to prevent the journal file from being written
**   while it is being traversed by code in pager_playback().
** 
**   If doNotSyncSpill is non-zero, writing to the database from pagerStress()
**   is permitted, but syncing the journal file is not. This flag is set
**   by sqlite3PagerWrite() when the file-system sector-size is larger than
**   the database page-size in order to prevent a journal sync from happening 
**   in between the journalling of two pages on the same sector. 


**
** subjInMemory
**
**   This is a boolean variable. If true, then any required sub-journal
**   is opened as an in-memory journal file. If false, then in-memory
**   sub-journals are only used for in-memory pager files.
**
................................................................................
/*
** This routine ends a transaction. A transaction is usually ended by 
** either a COMMIT or a ROLLBACK operation. This routine may be called 
** after rollback of a hot-journal, or if an error occurs while opening
** the journal file or writing the very first journal-header of a
** database transaction.
** 
** This routine is never called in PAGER_ERROR state. If it is called
** in PAGER_NONE or PAGER_SHARED state and the lock held is less
** exclusive than a RESERVED lock, it is a no-op.
**
** Otherwise, any active savepoints are released.
**
** If the journal file is open, then it is "finalized". Once a journal 
** file has been finalized it is not possible to use it to roll back a 
** transaction. Nor will it be considered to be a hot-journal by this
** or any other database connection. Exactly how a journal is finalized
................................................................................
**     The journal file is closed and deleted using sqlite3OsDelete().
**
**     If the pager is running in exclusive mode, this method of finalizing
**     the journal file is never used. Instead, if the journalMode is
**     DELETE and the pager is in exclusive mode, the method described under
**     journalMode==PERSIST is used instead.
**
** After the journal is finalized, the pager moves to PAGER_READER state.
** If running in non-exclusive rollback mode, the lock on the file is 
** downgraded to a SHARED_LOCK.




**
** SQLITE_OK is returned if no error occurs. If an error occurs during
** any of the IO operations to finalize the journal file or unlock the
** database then the IO error code is returned to the user. If the 
** operation to finalize the journal file fails, then the code still
** tries to unlock the database file if not in exclusive mode. If the
** unlock operation fails as well, then the first error code related
................................................................................
** to the first error encountered (the journal finalization one) is
** returned.
*/
static int pager_end_transaction(Pager *pPager, int hasMaster){
  int rc = SQLITE_OK;      /* Error code from journal finalization operation */
  int rc2 = SQLITE_OK;     /* Error code from db file unlock operation */

  /* Do nothing if the pager does not have an open write transaction
  ** or at least a RESERVED lock. This function may be called when there
  ** is no write-transaction active but a RESERVED or greater lock is
  ** held under two circumstances:
  **
  **   1. After a successful hot-journal rollback, it is called with
  **      eState==PAGER_NONE and eLock==EXCLUSIVE_LOCK.
  **
  **   2. If a connection with locking_mode=exclusive holding an EXCLUSIVE 
  **      lock switches back to locking_mode=normal and then executes a
  **      read-transaction, this function is called with eState==PAGER_READER 
  **      and eLock==EXCLUSIVE_LOCK when the read-transaction is closed.
  */
  assert( assert_pager_state(pPager) );
  assert( pPager->eState!=PAGER_ERROR );
  if( pPager->eState<PAGER_WRITER_LOCKED && pPager->eLock<RESERVED_LOCK ){
    return SQLITE_OK;
  }

  releaseAllSavepoints(pPager);
................................................................................
  return (rc==SQLITE_OK?rc2:rc);
}

/*
** Execute a rollback if a transaction is active and unlock the 
** database file. 
**
** If the pager has already entered the ERROR state, do not attempt 
** the rollback at this time. Instead, pager_unlock() is called. The
** call to pager_unlock() will discard all in-memory pages, unlock
** the database file and move the pager back to OPEN state. If this 
** means that there is a hot-journal left in the file-system, the next 
** connection to obtain a shared lock on the pager (which may be this one) 
** will roll it back.
**
** If the pager has not already entered the ERROR state, but an IO or
** malloc error occurs during a rollback, then this will itself cause 
** the pager to enter the ERROR state. Which will be cleared by the
** call to pager_unlock(), as described above.
*/
static void pagerUnlockAndRollback(Pager *pPager){
  if( pPager->eState!=PAGER_ERROR && pPager->eState!=PAGER_OPEN ){
    assert( assert_pager_state(pPager) );
    if( pPager->eState>=PAGER_WRITER_LOCKED ){
      sqlite3BeginBenignMalloc();
      sqlite3PagerRollback(pPager);
      sqlite3EndBenignMalloc();
    }else if( !pPager->exclusiveMode ){
      assert( pPager->eState==PAGER_READER );
      pager_end_transaction(pPager, 0);
    }
  }
  pager_unlock(pPager);
}

................................................................................

/*
** Read a single page from either the journal file (if isMainJrnl==1) or
** from the sub-journal (if isMainJrnl==0) and playback that page.
** The page begins at offset *pOffset into the file. The *pOffset
** value is increased to the start of the next page in the journal.
**
** The main rollback journal uses checksums - the statement journal does 
** not.

**
** If the page number of the page record read from the (sub-)journal file
** is greater than the current value of Pager.dbSize, then playback is
** skipped and SQLITE_OK is returned.
**
** If pDone is not NULL, then it is a record of pages that have already
** been played back.  If the page at *pOffset has already been played back
................................................................................
  assert( (isSavepnt&~1)==0 );       /* isSavepnt is 0 or 1 */
  assert( isMainJrnl || pDone );     /* pDone always used on sub-journals */
  assert( isSavepnt || pDone==0 );   /* pDone never used on non-savepoint */

  aData = pPager->pTmpSpace;
  assert( aData );         /* Temp storage must have already been allocated */
  assert( pagerUseWal(pPager)==0 || (!isMainJrnl && isSavepnt) );

  /* Either the state is greater than PAGER_WRITER_CACHEMOD (a transaction 
  ** or savepoint rollback done at the request of the caller) or this is
  ** a hot-journal rollback. If it is a hot-journal rollback, the pager
  ** is in state OPEN and holds an EXCLUSIVE lock. Hot-journal rollback
  ** only reads from the main journal, not the sub-journal.
  */
  assert( pPager->eState>=PAGER_WRITER_CACHEMOD
       || (pPager->eState==PAGER_OPEN && pPager->eLock==EXCLUSIVE_LOCK)
  );
  assert( pPager->eState>=PAGER_WRITER_CACHEMOD || isMainJrnl );

  /* Read the page number and page data from the journal or sub-journal
  ** file. Return an error code to the caller if an IO error occurs.
  */
  jfd = isMainJrnl ? pPager->jfd : pPager->sjfd;
  rc = read32bits(jfd, *pOffset, &pgno);
  if( rc!=SQLITE_OK ) return rc;
................................................................................

  /* If this page has already been played by before during the current
  ** rollback, then don't bother to play it back again.
  */
  if( pDone && (rc = sqlite3BitvecSet(pDone, pgno))!=SQLITE_OK ){
    return rc;
  }




  /* When playing back page 1, restore the nReserve setting
  */
  if( pgno==1 && pPager->nReserve!=((u8*)aData)[20] ){
    pPager->nReserve = ((u8*)aData)[20];
    pagerReportSize(pPager);
  }
................................................................................
  if( rc==SQLITE_OK && changed ){
    pager_reset(pPager);
  }

  return rc;
}


/*
** This function is called as part of the transition from PAGER_OPEN
** to PAGER_READER state to determine the size of the database file
** in pages (assuming the page size currently stored in Pager.pageSize).
**
** If no error occurs, SQLITE_OK is returned and the size of the database
** in pages is stored in *pnPage. Otherwise, an error code (perhaps
** SQLITE_IOERR_FSTAT) is returned and *pnPage is left unmodified.
*/
static int pagerPagecount(Pager *pPager, Pgno *pnPage){
  Pgno nPage;                     /* Value to return via *pnPage */

  /* Query the WAL sub-system for the database size. The WalDbsize()
  ** function returns zero if the WAL is not open (i.e. Pager.pWal==0), or
  ** if the database size is not available. The database size is not
  ** available from the WAL sub-system if the log file is empty or
  ** contains no valid committed transactions.
  */
  assert( pPager->eState==PAGER_OPEN );
  assert( pPager->eLock>=SHARED_LOCK || pPager->noReadlock );
  nPage = sqlite3WalDbsize(pPager->pWal);

  /* If the database size was not available from the WAL sub-system,
  ** determine it based on the size of the database file. If the size
  ** of the database file is not an integer multiple of the page-size,
  ** round down to the nearest page. Except, any file larger than 0
  ** bytes in size is considered to contain at least one page.
  */
  if( nPage==0 ){
    i64 n = 0;                    /* Size of db file in bytes */
    assert( isOpen(pPager->fd) || pPager->tempFile );
    if( isOpen(pPager->fd) ){
      int rc = sqlite3OsFileSize(pPager->fd, &n);
      if( rc!=SQLITE_OK ){
        return rc;
................................................................................
    }
    nPage = (Pgno)(n / pPager->pageSize);
    if( nPage==0 && n>0 ){
      nPage = 1;
    }
  }

  /* If the current number of pages in the file is greater than the
  ** configured maximum pager number, increase the allowed limit so
  ** that the file can be read.
  */
  if( nPage>pPager->mxPgno ){
    pPager->mxPgno = (Pgno)nPage;
  }

................................................................................
** If the database is not empty and the *-wal file exists, open the pager
** in WAL mode.  If the database is empty or if no *-wal file exists and
** if no error occurs, make sure Pager.journalMode is not set to
** PAGER_JOURNALMODE_WAL.
**
** Return SQLITE_OK or an error code.
**


** The caller must hold a SHARED lock on the database file to call this
** function. Because an EXCLUSIVE lock on the db file is required to delete 
** a WAL on a none-empty database, this ensures there is no race condition 
** between the xAccess() below and an xDelete() being executed by some 
** other connection.
*/
static int pagerOpenWalIfPresent(Pager *pPager){
  int rc = SQLITE_OK;
  assert( pPager->eState==PAGER_OPEN );
  assert( pPager->eLock>=SHARED_LOCK || pPager->noReadlock );

  if( !pPager->tempFile ){
    int isWal;                    /* True if WAL file exists */
    Pgno nPage;                   /* Size of the database file */



    rc = pagerPagecount(pPager, &nPage);
    if( rc ) return rc;
    if( nPage==0 ){
      rc = sqlite3OsDelete(pPager->pVfs, pPager->zWal, 0);
      isWal = 0;
    }else{
      rc = sqlite3OsAccess(
................................................................................
          pPager->pVfs, pPager->zWal, SQLITE_ACCESS_EXISTS, &isWal
      );
    }
    if( rc==SQLITE_OK ){
      if( isWal ){
        testcase( sqlite3PcachePagecount(pPager->pPCache)==0 );
        rc = sqlite3PagerOpenWal(pPager, 0);



      }else if( pPager->journalMode==PAGER_JOURNALMODE_WAL ){
        pPager->journalMode = PAGER_JOURNALMODE_DELETE;
      }
    }
  }
  return rc;
}
................................................................................
** write-transaction has already been opened, this function is a no-op.
**
** If the exFlag argument is false, then acquire at least a RESERVED
** lock on the database file. If exFlag is true, then acquire at least
** an EXCLUSIVE lock. If such a lock is already held, no locking 
** functions need be called.
**








** If the subjInMemory argument is non-zero, then any sub-journal opened
** within this transaction will be opened as an in-memory file. This
** has no effect if the sub-journal is already opened (as it may be when
** running in exclusive mode) or if the transaction does not require a
** sub-journal. If the subjInMemory argument is zero, then any required
** sub-journal is implemented in-memory if pPager is an in-memory database, 
** or using a temporary file otherwise.
................................................................................
      ** lock, but not when obtaining the RESERVED lock.
      */
      rc = pagerLockDb(pPager, RESERVED_LOCK);
      if( rc==SQLITE_OK && exFlag ){
        rc = pager_wait_on_lock(pPager, EXCLUSIVE_LOCK);
      }
    }


    if( rc==SQLITE_OK ){
      /* Change to WRITER_LOCKED state.
      **
      ** WAL mode sets Pager.eState to PAGER_WRITER_LOCKED or CACHEMOD
      ** when it has an open transaction, but never to DBMOD or FINISHED.
      ** This is because in those states the code to roll back savepoint