SQLite

*/
#ifdef SQLITE_DEBUG
static int cursorHoldsMutex(BtCursor *p){
  return sqlite3_mutex_held(p->pBt->mutex);
}
#endif


#ifndef SQLITE_OMIT_INCRBLOB
/*
** Invalidate the overflow page-list cache for cursor pCur, if any.

*/
static void invalidateOverflowCache(BtCursor *pCur){
  assert( cursorHoldsMutex(pCur) );
  sqlite3_free(pCur->aOverflow);
  pCur->aOverflow = 0;
}

/*
** Invalidate the overflow page-list cache for all cursors opened
** on the shared btree structure pBt.
*/
static void invalidateAllOverflowCache(BtShared *pBt){
  BtCursor *p;
  assert( sqlite3_mutex_held(pBt->mutex) );
  for(p=pBt->pCursor; p; p=p->pNext){
    invalidateOverflowCache(p);
  }
}


/*
** This function is called before modifying the contents of a table
** to invalidate any incrblob cursors that are open on the
** row or one of the rows being modified.
**
** If argument isClearTable is true, then the entire contents of the
** table is about to be deleted. In this case invalidate all incrblob

    if( p->isIncrblobHandle && (isClearTable || p->info.nKey==iRow) ){
      p->eState = CURSOR_INVALID;
    }
  }
}

#else
  /* Stub functions when INCRBLOB is omitted */
  #define invalidateOverflowCache(x)
  #define invalidateAllOverflowCache(x)
  #define invalidateIncrblobCursors(x,y,z)
#endif /* SQLITE_OMIT_INCRBLOB */

/*
** Set bit pgno of the BtShared.pHasContent bitvec. This is called 
** when a page that previously contained data becomes a free-list leaf 
** page.

    if( pCur->pNext ){
      pCur->pNext->pPrev = pCur->pPrev;
    }
    for(i=0; i<=pCur->iPage; i++){
      releasePage(pCur->apPage[i]);
    }
    unlockBtreeIfUnused(pBt);
    invalidateOverflowCache(pCur);
    /* sqlite3_free(pCur); */
    sqlite3BtreeLeave(pBtree);
  }
  return SQLITE_OK;
}

/*

    memcpy(pBuf, pPayload, nByte);
  }
  return SQLITE_OK;
}

/*
** This function is used to read or overwrite payload information
** for the entry that the pCur cursor is pointing to. If the eOp
** parameter is 0, this is a read operation (data copied into
** buffer pBuf). If it is non-zero, a write (data copied from
** buffer pBuf).



**
** A total of "amt" bytes are read or written beginning at "offset".
** Data is read to or from the buffer pBuf.
**
** The content being read or written might appear on the main page
** or be scattered out on multiple overflow pages.
**
** If the BtCursor.isIncrblobHandle flag is set, and the current
** cursor entry uses one or more overflow pages, this function

** allocates space for and lazily popluates the overflow page-list 
** cache array (BtCursor.aOverflow). Subsequent calls use this
** cache to make seeking to the supplied offset more efficient.
**
** Once an overflow page-list cache has been allocated, it may be
** invalidated if some other cursor writes to the same table, or if
** the cursor is moved to a different row. Additionally, in auto-vacuum
** mode, the following events may invalidate an overflow page-list cache.
**
**   * An incremental vacuum,

  /* Check if data must be read/written to/from the btree page itself. */
  if( offset<pCur->info.nLocal ){
    int a = amt;
    if( a+offset>pCur->info.nLocal ){
      a = pCur->info.nLocal - offset;
    }
    rc = copyPayload(&aPayload[offset], pBuf, a, eOp, pPage->pDbPage);
    offset = 0;
    pBuf += a;
    amt -= a;
  }else{
    offset -= pCur->info.nLocal;
  }

  if( rc==SQLITE_OK && amt>0 ){
    const u32 ovflSize = pBt->usableSize - 4;  /* Bytes content per ovfl page */
    Pgno nextPage;

    nextPage = get4byte(&aPayload[pCur->info.nLocal]);

#ifndef SQLITE_OMIT_INCRBLOB
    /* If the isIncrblobHandle flag is set and the BtCursor.aOverflow[]
    ** has not been allocated, allocate it now. The array is sized at
    ** one entry for each overflow page in the overflow chain. The
    ** page number of the first overflow page is stored in aOverflow[0],
    ** etc. A value of 0 in the aOverflow[] array means "not yet known"
    ** (the cache is lazily populated).
    */
    if( pCur->isIncrblobHandle && !pCur->aOverflow ){
      int nOvfl = (pCur->info.nPayload-pCur->info.nLocal+ovflSize-1)/ovflSize;

      pCur->aOverflow = (Pgno *)sqlite3MallocZero(sizeof(Pgno)*nOvfl);
      /* nOvfl is always positive.  If it were zero, fetchPayload would have
      ** been used instead of this routine. */

      if( ALWAYS(nOvfl) && !pCur->aOverflow ){
        rc = SQLITE_NOMEM;








      }
    }

    /* If the overflow page-list cache has been allocated and the
    ** entry for the first required overflow page is valid, skip
    ** directly to it.
    */
    if( pCur->aOverflow && pCur->aOverflow[offset/ovflSize] ){
      iIdx = (offset/ovflSize);
      nextPage = pCur->aOverflow[iIdx];
      offset = (offset%ovflSize);
    }
#endif

    for( ; rc==SQLITE_OK && amt>0 && nextPage; iIdx++){

#ifndef SQLITE_OMIT_INCRBLOB
      /* If required, populate the overflow page-list cache. */
      if( pCur->aOverflow ){
        assert(!pCur->aOverflow[iIdx] || pCur->aOverflow[iIdx]==nextPage);
        pCur->aOverflow[iIdx] = nextPage;
      }
#endif

      if( offset>=ovflSize ){
        /* The only reason to read this page is to obtain the page
        ** number for the next page in the overflow chain. The page
        ** data is not required. So first try to lookup the overflow
        ** page-list cache, if any, then fall back to the getOverflowPage()
        ** function.
        */
#ifndef SQLITE_OMIT_INCRBLOB
        if( pCur->aOverflow && pCur->aOverflow[iIdx+1] ){
          nextPage = pCur->aOverflow[iIdx+1];
        } else 
#endif
          rc = getOverflowPage(pBt, nextPage, 0, &nextPage);
        offset -= ovflSize;
      }else{
        /* Need to read this page properly. It contains some of the
        ** range of data that is being read (eOp==0) or written (eOp!=0).
        */
#ifdef SQLITE_DIRECT_OVERFLOW_READ

        **   4) there is no open write-transaction, and
        **   5) the database is not a WAL database,
        **
        ** then data can be read directly from the database file into the
        ** output buffer, bypassing the page-cache altogether. This speeds
        ** up loading large records that span many overflow pages.
        */
        if( eOp==0                                             /* (1) */
         && offset==0                                          /* (2) */
         && pBt->inTransaction==TRANS_READ                     /* (4) */
         && (fd = sqlite3PagerFile(pBt->pPager))->pMethods     /* (3) */
         && pBt->pPage1->aData[19]==0x01                       /* (5) */
        ){
          u8 aSave[4];
          u8 *aWrite = &pBuf[-4];
          memcpy(aSave, aWrite, 4);
          rc = sqlite3OsRead(fd, aWrite, a+4, (i64)pBt->pageSize*(nextPage-1));
          nextPage = get4byte(aWrite);
          memcpy(aWrite, aSave, 4);
        }else
#endif

        {
          DbPage *pDbPage;
          rc = sqlite3PagerAcquire(pBt->pPager, nextPage, &pDbPage,
              (eOp==0 ? PAGER_GET_READONLY : 0)
          );
          if( rc==SQLITE_OK ){
            aPayload = sqlite3PagerGetData(pDbPage);
            nextPage = get4byte(aPayload);
            rc = copyPayload(&aPayload[offset+4], pBuf, a, eOp, pDbPage);
            sqlite3PagerUnref(pDbPage);
            offset = 0;
          }
        }
        amt -= a;
        pBuf += a;
      }

  MemPage *pRoot;
  int rc = SQLITE_OK;

  assert( cursorHoldsMutex(pCur) );
  assert( CURSOR_INVALID < CURSOR_REQUIRESEEK );
  assert( CURSOR_VALID   < CURSOR_REQUIRESEEK );
  assert( CURSOR_FAULT   > CURSOR_REQUIRESEEK );

  if( pCur->eState>=CURSOR_REQUIRESEEK ){
    if( pCur->eState==CURSOR_FAULT ){
      assert( pCur->skipNext!=SQLITE_OK );
      return pCur->skipNext;
    }
    sqlite3BtreeClearCursor(pCur);
  }

          nCell = (int)pCur->info.nKey;
          pCellKey = sqlite3Malloc( nCell );
          if( pCellKey==0 ){
            rc = SQLITE_NOMEM;
            goto moveto_finish;
          }
          pCur->aiIdx[pCur->iPage] = (u16)idx;
          rc = accessPayload(pCur, 0, nCell, (unsigned char*)pCellKey, 0);
          if( rc ){
            sqlite3_free(pCellKey);
            goto moveto_finish;
          }
          c = xRecordCompare(nCell, pCellKey, pIdxKey, 0);
          sqlite3_free(pCellKey);
        }

  int idx;
  MemPage *pPage;

  assert( cursorHoldsMutex(pCur) );
  assert( pRes!=0 );
  assert( *pRes==0 || *pRes==1 );
  assert( pCur->skipNext==0 || pCur->eState!=CURSOR_VALID );

  if( pCur->eState!=CURSOR_VALID ){
    rc = restoreCursorPosition(pCur);
    if( rc!=SQLITE_OK ){
      *pRes = 0;
      return rc;
    }
    if( CURSOR_INVALID==pCur->eState ){

  int rc;
  MemPage *pPage;

  assert( cursorHoldsMutex(pCur) );
  assert( pRes!=0 );
  assert( *pRes==0 || *pRes==1 );
  assert( pCur->skipNext==0 || pCur->eState!=CURSOR_VALID );

  pCur->atLast = 0;
  if( pCur->eState!=CURSOR_VALID ){
    if( ALWAYS(pCur->eState>=CURSOR_REQUIRESEEK) ){
      rc = btreeRestoreCursorPosition(pCur);
      if( rc!=SQLITE_OK ){
        *pRes = 0;
        return rc;

  assert( !hasReadConflicts(pCsr->pBtree, pCsr->pgnoRoot) );
  assert( pCsr->apPage[pCsr->iPage]->intKey );

  return accessPayload(pCsr, offset, amt, (unsigned char *)z, 1);
}

/* 
** Set a flag on this cursor to cache the locations of pages from the 
** overflow list for the current row. This is used by cursors opened
** for incremental blob IO only.
**
** This function sets a flag only. The actual page location cache
** (stored in BtCursor.aOverflow[]) is allocated and used by function
** accessPayload() (the worker function for sqlite3BtreeData() and
** sqlite3BtreePutData()).
*/
void sqlite3BtreeCacheOverflow(BtCursor *pCur){
  assert( cursorHoldsMutex(pCur) );
  assert( sqlite3_mutex_held(pCur->pBtree->db->mutex) );
  invalidateOverflowCache(pCur);
  pCur->isIncrblobHandle = 1;
}
#endif

/*
** Set both the "read version" (single byte at byte offset 18) and 
** "write version" (single byte at byte offset 19) fields in the database

int sqlite3BtreeDataSize(BtCursor*, u32 *pSize);
int sqlite3BtreeData(BtCursor*, u32 offset, u32 amt, void*);

char *sqlite3BtreeIntegrityCheck(Btree*, int *aRoot, int nRoot, int, int*);
struct Pager *sqlite3BtreePager(Btree*);

int sqlite3BtreePutData(BtCursor*, u32 offset, u32 amt, void*);
void sqlite3BtreeCacheOverflow(BtCursor *);
void sqlite3BtreeClearCursor(BtCursor *);
int sqlite3BtreeSetVersion(Btree *pBt, int iVersion);
void sqlite3BtreeCursorHints(BtCursor *, unsigned int mask);

#ifndef NDEBUG
int sqlite3BtreeCursorIsValid(BtCursor*);
#endif

*/
struct BtCursor {
  Btree *pBtree;            /* The Btree to which this cursor belongs */
  BtShared *pBt;            /* The BtShared this cursor points to */
  BtCursor *pNext, *pPrev;  /* Forms a linked list of all cursors */
  struct KeyInfo *pKeyInfo; /* Argument passed to comparison function */
#ifndef SQLITE_OMIT_INCRBLOB


  Pgno *aOverflow;          /* Cache of overflow page locations */
#endif
  Pgno pgnoRoot;            /* The root page of this tree */
  CellInfo info;            /* A parse of the cell we are pointing at */
  i64 nKey;        /* Size of pKey, or last integer key */
  void *pKey;      /* Saved key that was cursor's last known position */
  int skipNext;    /* Prev() is noop if negative. Next() is noop if positive */

      rc = SQLITE_ERROR;
      sqlite3_finalize(p->pStmt);
      p->pStmt = 0;
    }else{
      p->iOffset = pC->aType[p->iCol + pC->nField];
      p->nByte = sqlite3VdbeSerialTypeLen(type);
      p->pCsr =  pC->pCursor;
      sqlite3BtreeEnterCursor(p->pCsr);
      sqlite3BtreeCacheOverflow(p->pCsr);
      sqlite3BtreeLeaveCursor(p->pCsr);
    }
  }

  if( rc==SQLITE_ROW ){
    rc = SQLITE_OK;
  }else if( p->pStmt ){
    rc = sqlite3_finalize(p->pStmt);