SQLite

/*
** 2004 April 6
**
** 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.
**
*************************************************************************
** $Id: btree.c,v 1.332 2006/12/18 18:34:51 drh Exp $
**
** This file implements a external (disk-based) database using BTrees.
** For a detailed discussion of BTrees, refer to
**
**     Donald E. Knuth, THE ART OF COMPUTER PROGRAMMING, Volume 3:
**     "Sorting And Searching", pages 473-480. Addison-Wesley
**     Publishing Company, Reading, Massachusetts.

/*
** The TRACE macro will print high-level status information about the
** btree operation when the global variable sqlite3_btree_trace is
** enabled.
*/
#if SQLITE_TEST
# define TRACE(X)   if( sqlite3_btree_trace )\
/*                        { sqlite3DebugPrintf X; fflush(stdout); } */ \
{ printf X; fflush(stdout); }
int sqlite3_btree_trace=0;  /* True to enable tracing */
#else
# define TRACE(X)
#endif

/*
** Forward declaration

  Pgno nearby,
  u8 exact
){
  MemPage *pPage1;
  int rc;
  int n;     /* Number of pages on the freelist */
  int k;     /* Number of leaves on the trunk of the freelist */
  MemPage *pTrunk = 0;
  MemPage *pPrevTrunk = 0;

  pPage1 = pBt->pPage1;
  n = get4byte(&pPage1->aData[36]);
  if( n>0 ){
    /* There are pages on the freelist.  Reuse one of those pages. */

    Pgno iTrunk;

    u8 searchList = 0; /* If the free-list must be searched for 'nearby' */
    
    /* If the 'exact' parameter was true and a query of the pointer-map
    ** shows that the page 'nearby' is somewhere on the free-list, then
    ** the entire-list will be searched for that page.
    */
#ifndef SQLITE_OMIT_AUTOVACUUM

      if( pPrevTrunk ){
        iTrunk = get4byte(&pPrevTrunk->aData[0]);
      }else{
        iTrunk = get4byte(&pPage1->aData[32]);
      }
      rc = getPage(pBt, iTrunk, &pTrunk);
      if( rc ){
        pTrunk = 0;


        goto end_allocate_page;






      }

      k = get4byte(&pTrunk->aData[4]);
      if( k==0 && !searchList ){
        /* The trunk has no leaves and the list is not being searched. 
        ** So extract the trunk page itself and use it as the newly 
        ** allocated page */
        assert( pPrevTrunk==0 );
        rc = sqlite3pager_write(pTrunk->aData);
        if( rc ){
          goto end_allocate_page;
        }
        *pPgno = iTrunk;
        memcpy(&pPage1->aData[32], &pTrunk->aData[0], 4);
        *ppPage = pTrunk;
        pTrunk = 0;
        TRACE(("ALLOCATE: %d trunk - %d free pages left\n", *pPgno, n-1));
      }else if( k>pBt->usableSize/4 - 8 ){
        /* Value of k is out of range.  Database corruption */
        rc = SQLITE_CORRUPT_BKPT;
        goto end_allocate_page;
#ifndef SQLITE_OMIT_AUTOVACUUM
      }else if( searchList && nearby==iTrunk ){
        /* The list is being searched and this trunk page is the page
        ** to allocate, regardless of whether it has leaves.
        */
        assert( *pPgno==iTrunk );
        *ppPage = pTrunk;
        searchList = 0;
        rc = sqlite3pager_write(pTrunk->aData);
        if( rc ){
          goto end_allocate_page;
        }
        if( k==0 ){
          if( !pPrevTrunk ){
            memcpy(&pPage1->aData[32], &pTrunk->aData[0], 4);
          }else{
            memcpy(&pPrevTrunk->aData[0], &pTrunk->aData[0], 4);
          }
        }else{
          /* The trunk page is required by the caller but it contains 
          ** pointers to free-list leaves. The first leaf becomes a trunk
          ** page in this case.
          */
          MemPage *pNewTrunk;
          Pgno iNewTrunk = get4byte(&pTrunk->aData[8]);
          rc = getPage(pBt, iNewTrunk, &pNewTrunk);
          if( rc!=SQLITE_OK ){
            goto end_allocate_page;


          }
          rc = sqlite3pager_write(pNewTrunk->aData);
          if( rc!=SQLITE_OK ){
            releasePage(pNewTrunk);
            goto end_allocate_page;


          }
          memcpy(&pNewTrunk->aData[0], &pTrunk->aData[0], 4);
          put4byte(&pNewTrunk->aData[4], k-1);
          memcpy(&pNewTrunk->aData[8], &pTrunk->aData[12], (k-1)*4);
          releasePage(pNewTrunk);
          if( !pPrevTrunk ){
            put4byte(&pPage1->aData[32], iNewTrunk);
          }else{
            rc = sqlite3pager_write(pPrevTrunk->aData);
            if( rc ){
              goto end_allocate_page;
            }
            put4byte(&pPrevTrunk->aData[0], iNewTrunk);
          }

        }
        pTrunk = 0;
        TRACE(("ALLOCATE: %d trunk - %d free pages left\n", *pPgno, n-1));
#endif
      }else{
        /* Extract a leaf from the trunk */
        int closest;
        Pgno iPage;
        unsigned char *aData = pTrunk->aData;
        rc = sqlite3pager_write(aData);
        if( rc ){
          goto end_allocate_page;
        }
        if( nearby>0 ){
          int i, dist;
          closest = 0;
          dist = get4byte(&aData[8]) - nearby;
          if( dist<0 ) dist = -dist;
          for(i=1; i<k; i++){
            int d2 = get4byte(&aData[8+i*4]) - nearby;

              releasePage(*ppPage);
            }
          }
          searchList = 0;
        }
      }
      releasePage(pPrevTrunk);
      pPrevTrunk = 0;
    }while( searchList );

  }else{
    /* There are no pages on the freelist, so create a new page at the
    ** end of the file */
    *pPgno = sqlite3pager_pagecount(pBt->pPager) + 1;

#ifndef SQLITE_OMIT_AUTOVACUUM
    if( pBt->autoVacuum && PTRMAP_ISPAGE(pBt, *pPgno) ){

    if( rc!=SQLITE_OK ){
      releasePage(*ppPage);
    }
    TRACE(("ALLOCATE: %d from end of file\n", *pPgno));
  }

  assert( *pPgno!=PENDING_BYTE_PAGE(pBt) );

end_allocate_page:
  releasePage(pTrunk);
  releasePage(pPrevTrunk);
  return rc;
}

/*
** Add a page of the database file to the freelist.
**
** sqlite3pager_unref() is NOT called for pPage.

** The pager is used to access a database disk file.  It implements
** atomic commit and rollback through the use of a journal file that
** is separate from the database file.  The pager also implements file
** locking to prevent two processes from writing the same database
** file simultaneously, or one process from reading the database while
** another is writing.
**
** @(#) $Id: pager.c,v 1.277 2006/12/18 18:34:51 drh Exp $
*/
#ifndef SQLITE_OMIT_DISKIO
#include "sqliteInt.h"
#include "os.h"
#include "pager.h"
#include <assert.h>
#include <string.h>

/*
** Macros for troubleshooting.  Normally turned off
*/
#if 0
#define sqlite3DebugPrintf printf
#define TRACE1(X)       sqlite3DebugPrintf(X)
#define TRACE2(X,Y)     sqlite3DebugPrintf(X,Y)
#define TRACE3(X,Y,Z)   sqlite3DebugPrintf(X,Y,Z)
#define TRACE4(X,Y,Z,W) sqlite3DebugPrintf(X,Y,Z,W)
#define TRACE5(X,Y,Z,W,V) sqlite3DebugPrintf(X,Y,Z,W,V)
#else
#define TRACE1(X)

** means that the pager does not have to record the given page in the
** rollback journal.
*/
void sqlite3pager_dont_rollback(void *pData){
  PgHdr *pPg = DATA_TO_PGHDR(pData);
  Pager *pPager = pPg->pPager;

  assert( pPager->state>=PAGER_RESERVED );
  if( pPager->journalOpen==0 ) return;
  if( pPg->alwaysRollback || pPager->alwaysRollback || MEMDB ) return;
  if( !pPg->inJournal && (int)pPg->pgno <= pPager->origDbSize ){
    assert( pPager->aInJournal!=0 );
    pPager->aInJournal[pPg->pgno/8] |= 1<<(pPg->pgno&7);
    pPg->inJournal = 1;
    if( pPager->stmtInUse ){
      pPager->aInStmt[pPg->pgno/8] |= 1<<(pPg->pgno&7);