sqllogictest

** the "sqlite3.h" header file at hand, you will find a copy in the first
** 6736 lines past this header comment.)  Additional code files may be
** needed if you want a wrapper to interface SQLite with your choice of
** programming language.  The code for the "sqlite3" command-line shell
** is also in a separate file.  This file contains only code for the core
** SQLite library.
**
** This amalgamation was generated on 2009-02-03 16:30:31 UTC.
*/
#define SQLITE_CORE 1
#define SQLITE_AMALGAMATION 1
#ifndef SQLITE_PRIVATE
# define SQLITE_PRIVATE static
#endif
#ifndef SQLITE_API

**    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.
**
*************************************************************************
** Internal interface definitions for SQLite.
**
** @(#) $Id: sqliteInt.h,v 1.830 2009/01/24 11:30:43 drh Exp $
*/
#ifndef _SQLITEINT_H_
#define _SQLITEINT_H_

/*
** Include the configuration header output by 'configure' if we're using the
** autoconf-based build

** on how SQLite interfaces are suppose to operate.
**
** The name of this file under configuration management is "sqlite.h.in".
** The makefile makes some minor changes to this file (such as inserting
** the version number) and changes its name to "sqlite3.h" as
** part of the build process.
**
** @(#) $Id: sqlite.h.in,v 1.422 2009/01/23 16:45:01 danielk1977 Exp $
*/
#ifndef _SQLITE3_H_
#define _SQLITE3_H_
#include <stdarg.h>     /* Needed for the definition of va_list */

/*
** Make sure we can call this stuff from C++.

**                SQLite is holding pinned pages that can be unpinned
**                by writing their contents to the database file (a
**                relatively expensive operation). In this situation the
**                cache implementation has two choices: it can return NULL,
**                in which case SQLite will attempt to unpin one or more 
**                pages before re-requesting the same page, or it can
**                allocate a new page and return a pointer to it. If a new
**                page is allocated, then the first sizeof(void*) bytes of
**                it (at least) must be zeroed before it is returned.
**   <tr><td>2<td>If createFlag is set to 2, then SQLite is not holding any
**                pinned pages associated with the specific cache passed
**                as the first argument to xFetch() that can be unpinned. The
**                cache implementation should attempt to allocate a new
**                cache entry and return a pointer to it. Again, the first
**                sizeof(void*) bytes of the page should be zeroed before 
**                it is returned. If the xFetch() method returns NULL when 
**                createFlag==2, SQLite assumes that a memory allocation 
**                failed and returns SQLITE_NOMEM to the user.
** </table>
**
** xUnpin() is called by SQLite with a pointer to a currently pinned page
** as its second argument. If the third parameter, discard, is non-zero,
** then the page should be evicted from the cache. In this case SQLite 
** assumes that the next time the page is retrieved from the cache using
** the xFetch() method, it will be zeroed. If the discard parameter is

#define TK_ID                             26
#define TK_ABORT                          27
#define TK_AFTER                          28
#define TK_ANALYZE                        29
#define TK_ASC                            30
#define TK_ATTACH                         31
#define TK_BEFORE                         32
#define TK_BY                             33
#define TK_CASCADE                        34
#define TK_CAST                           35
#define TK_COLUMNKW                       36
#define TK_CONFLICT                       37
#define TK_DATABASE                       38
#define TK_DESC                           39
#define TK_DETACH                         40
#define TK_EACH                           41
#define TK_FAIL                           42
#define TK_FOR                            43
#define TK_IGNORE                         44
#define TK_INITIALLY                      45
#define TK_INSTEAD                        46
#define TK_LIKE_KW                        47
#define TK_MATCH                          48
#define TK_KEY                            49
#define TK_OF                             50
#define TK_OFFSET                         51
#define TK_PRAGMA                         52
#define TK_RAISE                          53
#define TK_REPLACE                        54
#define TK_RESTRICT                       55
#define TK_ROW                            56
#define TK_TRIGGER                        57
#define TK_VACUUM                         58
#define TK_VIEW                           59
#define TK_VIRTUAL                        60
#define TK_REINDEX                        61
#define TK_RENAME                         62
#define TK_CTIME_KW                       63
#define TK_ANY                            64
#define TK_OR                             65
#define TK_AND                            66
#define TK_IS                             67
#define TK_BETWEEN                        68
#define TK_IN                             69
#define TK_ISNULL                         70
#define TK_NOTNULL                        71
#define TK_NE                             72
#define TK_EQ                             73
#define TK_GT                             74
#define TK_LE                             75
#define TK_LT                             76
#define TK_GE                             77
#define TK_ESCAPE                         78
#define TK_BITAND                         79
#define TK_BITOR                          80
#define TK_LSHIFT                         81
#define TK_RSHIFT                         82
#define TK_PLUS                           83
#define TK_MINUS                          84
#define TK_STAR                           85
#define TK_SLASH                          86
#define TK_REM                            87
#define TK_CONCAT                         88
#define TK_COLLATE                        89
#define TK_UMINUS                         90
#define TK_UPLUS                          91
#define TK_BITNOT                         92
#define TK_STRING                         93
#define TK_JOIN_KW                        94
#define TK_CONSTRAINT                     95
#define TK_DEFAULT                        96
#define TK_NULL                           97
#define TK_PRIMARY                        98
#define TK_UNIQUE                         99
#define TK_CHECK                          100
#define TK_REFERENCES                     101
#define TK_AUTOINCR                       102
#define TK_ON                             103
#define TK_DELETE                         104
#define TK_UPDATE                         105
#define TK_INSERT                         106
#define TK_SET                            107
#define TK_DEFERRABLE                     108
#define TK_FOREIGN                        109
#define TK_DROP                           110
#define TK_UNION                          111
#define TK_ALL                            112
#define TK_EXCEPT                         113
#define TK_INTERSECT                      114
#define TK_SELECT                         115
#define TK_DISTINCT                       116
#define TK_DOT                            117
#define TK_FROM                           118
#define TK_JOIN                           119
#define TK_INDEXED                        120

#define TK_USING                          121
#define TK_ORDER                          122
#define TK_GROUP                          123
#define TK_HAVING                         124
#define TK_LIMIT                          125
#define TK_WHERE                          126
#define TK_INTO                           127
#define TK_VALUES                         128
#define TK_INTEGER                        129
#define TK_FLOAT                          130
#define TK_BLOB                           131
#define TK_REGISTER                       132
#define TK_VARIABLE                       133
#define TK_CASE                           134
#define TK_WHEN                           135
#define TK_THEN                           136
#define TK_ELSE                           137
#define TK_INDEX                          138
#define TK_TO_TEXT                        139
#define TK_TO_BLOB                        140
#define TK_TO_NUMERIC                     141
#define TK_TO_INT                         142
#define TK_TO_REAL                        143
#define TK_END_OF_FILE                    144
#define TK_ILLEGAL                        145
#define TK_SPACE                          146
#define TK_UNCLOSED_STRING                147
#define TK_FUNCTION                       148
#define TK_COLUMN                         149
#define TK_AGG_FUNCTION                   150
#define TK_AGG_COLUMN                     151
#define TK_CONST_FUNC                     152

/************** End of parse.h ***********************************************/
/************** Continuing where we left off in sqliteInt.h ******************/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>

**    May you share freely, never taking more than you give.
**
*************************************************************************
** This header file defines the interface that the sqlite B-Tree file
** subsystem.  See comments in the source code for a detailed description
** of what each interface routine does.
**
** @(#) $Id: btree.h,v 1.107 2009/01/24 11:30:43 drh Exp $
*/
#ifndef _BTREE_H_
#define _BTREE_H_

/* TODO: This definition is just included so other modules compile. It
** needs to be revisited.
*/

SQLITE_PRIVATE int sqlite3BtreeIsInReadTrans(Btree*);
SQLITE_PRIVATE void *sqlite3BtreeSchema(Btree *, int, void(*)(void *));
SQLITE_PRIVATE int sqlite3BtreeSchemaLocked(Btree *);
SQLITE_PRIVATE int sqlite3BtreeLockTable(Btree *, int, u8);
SQLITE_PRIVATE int sqlite3BtreeSavepoint(Btree *, int, int);

SQLITE_PRIVATE const char *sqlite3BtreeGetFilename(Btree *);

SQLITE_PRIVATE const char *sqlite3BtreeGetJournalname(Btree *);
SQLITE_PRIVATE int sqlite3BtreeCopyFile(Btree *, Btree *);

SQLITE_PRIVATE int sqlite3BtreeIncrVacuum(Btree *);

/* The flags parameter to sqlite3BtreeCreateTable can be the bitwise OR
** of the following flags:

/* Automatically generated.  Do not edit */
/* See the mkopcodeh.awk script for details */
#define OP_VNext                                1
#define OP_Affinity                             2
#define OP_Column                               3
#define OP_SetCookie                            4
#define OP_Seek                                 5
#define OP_Real                               130   /* same as TK_FLOAT    */
#define OP_Sequence                             6
#define OP_Savepoint                            7
#define OP_Ge                                  77   /* same as TK_GE       */
#define OP_RowKey                               8
#define OP_SCopy                                9
#define OP_Eq                                  73   /* same as TK_EQ       */
#define OP_OpenWrite                           10
#define OP_NotNull                             71   /* same as TK_NOTNULL  */
#define OP_If                                  11
#define OP_ToInt                              142   /* same as TK_TO_INT   */
#define OP_String8                             93   /* same as TK_STRING   */
#define OP_VRowid                              12
#define OP_CollSeq                             13
#define OP_OpenRead                            14
#define OP_Expire                              15
#define OP_AutoCommit                          16
#define OP_Gt                                  74   /* same as TK_GT       */
#define OP_Pagecount                           17
#define OP_IntegrityCk                         18
#define OP_Sort                                20
#define OP_Copy                                21
#define OP_Trace                               22
#define OP_Function                            23
#define OP_IfNeg                               24
#define OP_And                                 66   /* same as TK_AND      */
#define OP_Subtract                            84   /* same as TK_MINUS    */
#define OP_Noop                                25
#define OP_Return                              26
#define OP_Remainder                           87   /* same as TK_REM      */
#define OP_NewRowid                            27
#define OP_Multiply                            85   /* same as TK_STAR     */
#define OP_Variable                            28
#define OP_String                              29
#define OP_RealAffinity                        30
#define OP_VRename                             31
#define OP_ParseSchema                         32
#define OP_VOpen                               33
#define OP_Close                               34
#define OP_CreateIndex                         35
#define OP_IsUnique                            36
#define OP_NotFound                            37
#define OP_Int64                               38
#define OP_MustBeInt                           39
#define OP_Halt                                40
#define OP_Rowid                               41
#define OP_IdxLT                               42
#define OP_AddImm                              43
#define OP_Statement                           44
#define OP_RowData                             45
#define OP_MemMax                              46
#define OP_Or                                  65   /* same as TK_OR       */
#define OP_NotExists                           47
#define OP_Gosub                               48
#define OP_Divide                              86   /* same as TK_SLASH    */
#define OP_Integer                             49
#define OP_ToNumeric                          141   /* same as TK_TO_NUMERIC*/
#define OP_Prev                                50
#define OP_RowSetRead                          51
#define OP_Concat                              88   /* same as TK_CONCAT   */
#define OP_RowSetAdd                           52
#define OP_BitAnd                              79   /* same as TK_BITAND   */
#define OP_VColumn                             53
#define OP_CreateTable                         54
#define OP_Last                                55
#define OP_SeekLe                              56
#define OP_IsNull                              70   /* same as TK_ISNULL   */
#define OP_IncrVacuum                          57
#define OP_IdxRowid                            58
#define OP_ShiftRight                          82   /* same as TK_RSHIFT   */
#define OP_ResetCount                          59
#define OP_ContextPush                         60
#define OP_Yield                               61
#define OP_DropTrigger                         62
#define OP_DropIndex                           63
#define OP_IdxGE                               64
#define OP_IdxDelete                           67
#define OP_Vacuum                              68
#define OP_IfNot                               69
#define OP_DropTable                           78
#define OP_SeekLt                              89
#define OP_MakeRecord                          90
#define OP_ToBlob                             140   /* same as TK_TO_BLOB  */
#define OP_ResultRow                           91
#define OP_Delete                              94
#define OP_AggFinal                            95
#define OP_Compare                             96
#define OP_ShiftLeft                           81   /* same as TK_LSHIFT   */
#define OP_Goto                                97
#define OP_TableLock                           98
#define OP_Clear                               99
#define OP_Le                                  75   /* same as TK_LE       */
#define OP_VerifyCookie                       100
#define OP_AggStep                            101
#define OP_ToText                             139   /* same as TK_TO_TEXT  */
#define OP_Not                                 19   /* same as TK_NOT      */
#define OP_ToReal                             143   /* same as TK_TO_REAL  */
#define OP_SetNumColumns                      102
#define OP_Transaction                        103
#define OP_VFilter                            104
#define OP_Ne                                  72   /* same as TK_NE       */
#define OP_VDestroy                           105
#define OP_ContextPop                         106
#define OP_BitOr                               80   /* same as TK_BITOR    */
#define OP_Next                               107
#define OP_IdxInsert                          108
#define OP_Lt                                  76   /* same as TK_LT       */
#define OP_SeekGe                             109
#define OP_Insert                             110
#define OP_Destroy                            111
#define OP_ReadCookie                         112
#define OP_LoadAnalysis                       113
#define OP_Explain                            114
#define OP_OpenPseudo                         115
#define OP_OpenEphemeral                      116
#define OP_Null                               117
#define OP_Move                               118
#define OP_Blob                               119
#define OP_Add                                 83   /* same as TK_PLUS     */
#define OP_Rewind                             120
#define OP_SeekGt                             121
#define OP_VBegin                             122
#define OP_VUpdate                            123
#define OP_IfZero                             124
#define OP_BitNot                              92   /* same as TK_BITNOT   */
#define OP_VCreate                            125
#define OP_Found                              126
#define OP_IfPos                              127
#define OP_NullRow                            128
#define OP_Jump                               129
#define OP_Permutation                        131

/* The following opcode values are never used */
#define OP_NotUsed_132                        132
#define OP_NotUsed_133                        133
#define OP_NotUsed_134                        134
#define OP_NotUsed_135                        135
#define OP_NotUsed_136                        136
#define OP_NotUsed_137                        137
#define OP_NotUsed_138                        138


/* Properties such as "out2" or "jump" that are specified in
** comments following the "case" for each opcode in the vdbe.c
** are encoded into bitvectors as follows:
*/
#define OPFLG_JUMP            0x0001  /* jump:  P2 holds jmp target */
#define OPFLG_OUT2_PRERELEASE 0x0002  /* out2-prerelease: */
#define OPFLG_IN1             0x0004  /* in1:   P1 is an input */
#define OPFLG_IN2             0x0008  /* in2:   P2 is an input */
#define OPFLG_IN3             0x0010  /* in3:   P3 is an input */
#define OPFLG_OUT3            0x0020  /* out3:  P3 is an output */
#define OPFLG_INITIALIZER {\
/*   0 */ 0x00, 0x01, 0x00, 0x00, 0x10, 0x08, 0x02, 0x00,\
/*   8 */ 0x00, 0x04, 0x00, 0x05, 0x02, 0x00, 0x00, 0x00,\
/*  16 */ 0x00, 0x02, 0x00, 0x04, 0x01, 0x04, 0x00, 0x00,\
/*  24 */ 0x05, 0x00, 0x04, 0x02, 0x02, 0x02, 0x04, 0x00,\
/*  32 */ 0x00, 0x00, 0x00, 0x02, 0x11, 0x11, 0x02, 0x05,\
/*  40 */ 0x00, 0x02, 0x11, 0x04, 0x00, 0x00, 0x0c, 0x11,\
/*  48 */ 0x01, 0x02, 0x01, 0x21, 0x08, 0x00, 0x02, 0x01,\
/*  56 */ 0x11, 0x01, 0x02, 0x00, 0x00, 0x04, 0x00, 0x00,\
/*  64 */ 0x11, 0x2c, 0x2c, 0x00, 0x00, 0x05, 0x05, 0x05,\
/*  72 */ 0x15, 0x15, 0x15, 0x15, 0x15, 0x15, 0x00, 0x2c,\
/*  80 */ 0x2c, 0x2c, 0x2c, 0x2c, 0x2c, 0x2c, 0x2c, 0x2c,\
/*  88 */ 0x2c, 0x11, 0x00, 0x00, 0x04, 0x02, 0x00, 0x00,\
/*  96 */ 0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,\
/* 104 */ 0x01, 0x00, 0x00, 0x01, 0x08, 0x11, 0x00, 0x02,\
/* 112 */ 0x02, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x02,\
/* 120 */ 0x01, 0x11, 0x00, 0x00, 0x05, 0x00, 0x11, 0x05,\
/* 128 */ 0x00, 0x01, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00,\
/* 136 */ 0x00, 0x00, 0x00, 0x04, 0x04, 0x04, 0x04, 0x04,\
}

/************** End of opcodes.h *********************************************/
/************** Continuing where we left off in vdbe.h ***********************/

/*
** Prototypes for the VDBE interface.  See comments on the implementation
** for a description of what each of these routines does.

**    May you share freely, never taking more than you give.
**
*************************************************************************
** This header file defines the interface that the sqlite page cache
** subsystem.  The page cache subsystem reads and writes a file a page
** at a time and provides a journal for rollback.
**
** @(#) $Id: pager.h,v 1.99 2009/01/31 14:54:07 danielk1977 Exp $
*/

#ifndef _PAGER_H_
#define _PAGER_H_

/*
** Default maximum size for persistent journal files. A negative 
** value means no limit. This value may be overridden using the 
** sqlite3PagerJournalSizeLimit() API. See also "PRAGMA journal_size_limit".
*/
#ifndef SQLITE_DEFAULT_JOURNAL_SIZE_LIMIT
  #define SQLITE_DEFAULT_JOURNAL_SIZE_LIMIT -1
#endif

/*
** The type used to represent a page number.  The first page in a file

typedef struct Pager Pager;

/*
** Handle type for pages.
*/
typedef struct PgHdr DbPage;

/*
** Page number PAGER_MJ_PGNO is never used in an SQLite database (it is
** reserved for working around a windows/posix incompatibility). It is
** used in the journal to signify that the remainder of the journal file 
** is devoted to storing a master journal name - there are no more pages to
** roll back. See comments for function writeMasterJournal() in pager.c 
** for details.
*/
#define PAGER_MJ_PGNO(x) ((Pgno)((PENDING_BYTE/((x)->pageSize))+1))

/*
** Allowed values for the flags parameter to sqlite3PagerOpen().
**
** NOTE: These values must match the corresponding BTREE_ values in btree.h.
*/
#define PAGER_OMIT_JOURNAL  0x0001    /* Do not use a rollback journal */
#define PAGER_NO_READLOCK   0x0002    /* Omit readlocks on readonly files */

/*
** Valid values for the second argument to sqlite3PagerLockingMode().
*/

#define PAGER_JOURNALMODE_DELETE      0   /* Commit by deleting journal file */
#define PAGER_JOURNALMODE_PERSIST     1   /* Commit by zeroing journal header */
#define PAGER_JOURNALMODE_OFF         2   /* Journal omitted.  */
#define PAGER_JOURNALMODE_TRUNCATE    3   /* Commit by truncating journal */
#define PAGER_JOURNALMODE_MEMORY      4   /* In-memory journal file */

/*
** The remainder of this file contains the declarations of the functions
** that make up the Pager sub-system API. See source code comments for 
** a detailed description of each routine.
*/

/* Open and close a Pager connection. */ 
SQLITE_PRIVATE int sqlite3PagerOpen(sqlite3_vfs *, Pager **ppPager, const char*, int,int,int);
SQLITE_PRIVATE int sqlite3PagerClose(Pager *pPager);
SQLITE_PRIVATE int sqlite3PagerReadFileheader(Pager*, int, unsigned char*);

/* Functions used to configure a Pager object. */
SQLITE_PRIVATE void sqlite3PagerSetBusyhandler(Pager*, int(*)(void *), void *);
SQLITE_PRIVATE void sqlite3PagerSetReiniter(Pager*, void(*)(DbPage*));
SQLITE_PRIVATE int sqlite3PagerSetPagesize(Pager*, u16*);
SQLITE_PRIVATE int sqlite3PagerMaxPageCount(Pager*, int);
SQLITE_PRIVATE void sqlite3PagerSetCachesize(Pager*, int);
SQLITE_PRIVATE void sqlite3PagerSetSafetyLevel(Pager*,int,int);
SQLITE_PRIVATE int sqlite3PagerLockingMode(Pager *, int);
SQLITE_PRIVATE int sqlite3PagerJournalMode(Pager *, int);
SQLITE_PRIVATE i64 sqlite3PagerJournalSizeLimit(Pager *, i64);

/* Functions used to obtain and release page references. */ 
SQLITE_PRIVATE int sqlite3PagerAcquire(Pager *pPager, Pgno pgno, DbPage **ppPage, int clrFlag);
#define sqlite3PagerGet(A,B,C) sqlite3PagerAcquire(A,B,C,0)
SQLITE_PRIVATE DbPage *sqlite3PagerLookup(Pager *pPager, Pgno pgno);
SQLITE_PRIVATE void sqlite3PagerRef(DbPage*);
SQLITE_PRIVATE void sqlite3PagerUnref(DbPage*);

/* Operations on page references. */
SQLITE_PRIVATE int sqlite3PagerWrite(DbPage*);
SQLITE_PRIVATE void sqlite3PagerDontWrite(DbPage*);
SQLITE_PRIVATE int sqlite3PagerMovepage(Pager*,DbPage*,Pgno,int);
SQLITE_PRIVATE int sqlite3PagerPageRefcount(DbPage*);
SQLITE_PRIVATE void *sqlite3PagerGetData(DbPage *); 
SQLITE_PRIVATE void *sqlite3PagerGetExtra(DbPage *); 

/* Functions used to manage pager transactions and savepoints. */
SQLITE_PRIVATE int sqlite3PagerPagecount(Pager*, int*);
SQLITE_PRIVATE int sqlite3PagerBegin(Pager*, int exFlag);
SQLITE_PRIVATE int sqlite3PagerCommitPhaseOne(Pager*,const char *zMaster, int);
SQLITE_PRIVATE int sqlite3PagerSync(Pager *pPager);
SQLITE_PRIVATE int sqlite3PagerCommitPhaseTwo(Pager*);
SQLITE_PRIVATE int sqlite3PagerRollback(Pager*);
SQLITE_PRIVATE int sqlite3PagerOpenSavepoint(Pager *pPager, int n);
SQLITE_PRIVATE int sqlite3PagerSavepoint(Pager *pPager, int op, int iSavepoint);

/* Functions used to query pager state and configuration. */
SQLITE_PRIVATE u8 sqlite3PagerIsreadonly(Pager*);


SQLITE_PRIVATE int sqlite3PagerRefcount(Pager*);

SQLITE_PRIVATE const char *sqlite3PagerFilename(Pager*);
SQLITE_PRIVATE const sqlite3_vfs *sqlite3PagerVfs(Pager*);
SQLITE_PRIVATE sqlite3_file *sqlite3PagerFile(Pager*);

SQLITE_PRIVATE const char *sqlite3PagerJournalname(Pager*);
SQLITE_PRIVATE int sqlite3PagerNosync(Pager*);






SQLITE_PRIVATE void *sqlite3PagerTempSpace(Pager*);




/* Functions used to truncate the database file. */

SQLITE_PRIVATE void sqlite3PagerTruncateImage(Pager*,Pgno);



/* Used by encryption extensions. */
#ifdef SQLITE_HAS_CODEC
SQLITE_PRIVATE   void sqlite3PagerSetCodec(Pager*,void*(*)(void*,void*,Pgno,int),void*);
#endif

/* Functions to support testing and debugging. */
#if !defined(NDEBUG) || defined(SQLITE_TEST)
SQLITE_PRIVATE   Pgno sqlite3PagerPagenumber(DbPage*);
SQLITE_PRIVATE   int sqlite3PagerIswriteable(DbPage*);
#endif

#ifdef SQLITE_TEST
SQLITE_PRIVATE   int *sqlite3PagerStats(Pager*);
SQLITE_PRIVATE   void sqlite3PagerRefdump(Pager*);
SQLITE_PRIVATE   int sqlite3PagerIsMemdb(Pager*);



  void disable_simulated_io_errors(void);
  void enable_simulated_io_errors(void);
#else
# define disable_simulated_io_errors()
# define enable_simulated_io_errors()
#endif

#endif /* _PAGER_H_ */

**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
** This header file defines the interface that the sqlite page cache
** subsystem. 
**
** @(#) $Id: pcache.h,v 1.19 2009/01/20 17:06:27 danielk1977 Exp $
*/

#ifndef _PCACHE_H_

typedef struct PgHdr PgHdr;
typedef struct PCache PCache;

/* Reset and close the cache object */
SQLITE_PRIVATE void sqlite3PcacheClose(PCache*);

/* Clear flags from pages of the page cache */
SQLITE_PRIVATE void sqlite3PcacheClearSyncFlags(PCache *);

/* Discard the contents of the cache */
SQLITE_PRIVATE void sqlite3PcacheClear(PCache*);

/* Return the total number of outstanding page references */
SQLITE_PRIVATE int sqlite3PcacheRefCount(PCache*);

/* Increment the reference count of an existing page */
SQLITE_PRIVATE void sqlite3PcacheRef(PgHdr*);

SQLITE_PRIVATE int sqlite3OsOpen(sqlite3_vfs *, const char *, sqlite3_file*, int, int *);
SQLITE_PRIVATE int sqlite3OsDelete(sqlite3_vfs *, const char *, int);
SQLITE_PRIVATE int sqlite3OsAccess(sqlite3_vfs *, const char *, int, int *pResOut);
SQLITE_PRIVATE int sqlite3OsFullPathname(sqlite3_vfs *, const char *, int, char *);
#ifndef SQLITE_OMIT_LOAD_EXTENSION
SQLITE_PRIVATE void *sqlite3OsDlOpen(sqlite3_vfs *, const char *);
SQLITE_PRIVATE void sqlite3OsDlError(sqlite3_vfs *, int, char *);
SQLITE_PRIVATE void (*sqlite3OsDlSym(sqlite3_vfs *, void *, const char *))(void);
SQLITE_PRIVATE void sqlite3OsDlClose(sqlite3_vfs *, void *);
#endif /* SQLITE_OMIT_LOAD_EXTENSION */
SQLITE_PRIVATE int sqlite3OsRandomness(sqlite3_vfs *, int, char *);
SQLITE_PRIVATE int sqlite3OsSleep(sqlite3_vfs *, int);
SQLITE_PRIVATE int sqlite3OsCurrentTime(sqlite3_vfs *, double*);

/*

#ifdef SQLITE_DEBUG
SQLITE_PRIVATE   int sqlite3Corrupt(void);
# define SQLITE_CORRUPT_BKPT sqlite3Corrupt()
#else
# define SQLITE_CORRUPT_BKPT SQLITE_CORRUPT
#endif

/*
** The following macros mimic the standard library functions toupper(),
** isspace(), isalnum(), isdigit() and isxdigit(), respectively. The
** sqlite versions only work for ASCII characters, regardless of locale.
*/
#ifdef SQLITE_ASCII
# define sqlite3Toupper(x)  ((x)&~(sqlite3CtypeMap[(unsigned char)(x)]&0x20))
# define sqlite3Isspace(x)   (sqlite3CtypeMap[(unsigned char)(x)]&0x01)
# define sqlite3Isalnum(x)   (sqlite3CtypeMap[(unsigned char)(x)]&0x06)
# define sqlite3Isalpha(x)   (sqlite3CtypeMap[(unsigned char)(x)]&0x02)
# define sqlite3Isdigit(x)   (sqlite3CtypeMap[(unsigned char)(x)]&0x04)
# define sqlite3Isxdigit(x)  (sqlite3CtypeMap[(unsigned char)(x)]&0x08)
# define sqlite3Tolower(x)   (sqlite3UpperToLower[(unsigned char)(x)])
#else
# include <ctype.h>
# define sqlite3Toupper(x)   toupper((unsigned char)(x))
# define sqlite3Isspace(x)   isspace((unsigned char)(x))
# define sqlite3Isalnum(x)   isalnum((unsigned char)(x))
# define sqlite3Isalpha(x)   isalpha((unsigned char)(x))
# define sqlite3Isdigit(x)   isdigit((unsigned char)(x))
# define sqlite3Isxdigit(x)  isxdigit((unsigned char)(x))
# define sqlite3Tolower(x)   tolower((unsigned char)(x))
#endif

/*
** Internal function prototypes
*/
SQLITE_PRIVATE int sqlite3StrICmp(const char *, const char *);
SQLITE_PRIVATE int sqlite3StrNICmp(const char *, const char *, int);
SQLITE_PRIVATE int sqlite3IsNumber(const char*, int*, u8);
SQLITE_PRIVATE int sqlite3Strlen(sqlite3*, const char*);

SQLITE_PRIVATE void sqlite3EndTable(Parse*,Token*,Token*,Select*);

SQLITE_PRIVATE Bitvec *sqlite3BitvecCreate(u32);
SQLITE_PRIVATE int sqlite3BitvecTest(Bitvec*, u32);
SQLITE_PRIVATE int sqlite3BitvecSet(Bitvec*, u32);
SQLITE_PRIVATE void sqlite3BitvecClear(Bitvec*, u32);
SQLITE_PRIVATE void sqlite3BitvecDestroy(Bitvec*);
SQLITE_PRIVATE u32 sqlite3BitvecSize(Bitvec*);
SQLITE_PRIVATE int sqlite3BitvecBuiltinTest(int,int*);

SQLITE_PRIVATE RowSet *sqlite3RowSetInit(sqlite3*, void*, unsigned int);
SQLITE_PRIVATE void sqlite3RowSetClear(RowSet*);
SQLITE_PRIVATE void sqlite3RowSetInsert(RowSet*, i64);
SQLITE_PRIVATE int sqlite3RowSetNext(RowSet*, i64*);

SQLITE_PRIVATE void sqlite3ValueFree(sqlite3_value*);
SQLITE_PRIVATE sqlite3_value *sqlite3ValueNew(sqlite3 *);
SQLITE_PRIVATE char *sqlite3Utf16to8(sqlite3 *, const void*, int);
SQLITE_PRIVATE int sqlite3ValueFromExpr(sqlite3 *, Expr *, u8, u8, sqlite3_value **);
SQLITE_PRIVATE void sqlite3ValueApplyAffinity(sqlite3_value *, u8, u8);
#ifndef SQLITE_AMALGAMATION
SQLITE_PRIVATE const unsigned char sqlite3UpperToLower[];
SQLITE_PRIVATE const unsigned char sqlite3CtypeMap[];
SQLITE_PRIVATE SQLITE_WSD struct Sqlite3Config sqlite3Config;
SQLITE_PRIVATE SQLITE_WSD FuncDefHash sqlite3GlobalFunctions;
#endif
SQLITE_PRIVATE void sqlite3RootPageMoved(Db*, int, int);
SQLITE_PRIVATE void sqlite3Reindex(Parse*, Token*, Token*);
SQLITE_PRIVATE void sqlite3AlterFunctions(sqlite3*);
SQLITE_PRIVATE void sqlite3AlterRenameTable(Parse*, SrcList*, Token*);

**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
**
** This file contains definitions of global variables and contants.
**
** $Id: global.c,v 1.11 2009/01/24 11:30:43 drh Exp $
*/


/* An array to map all upper-case characters into their corresponding
** lower-case character. 
**
** SQLite only considers US-ASCII (or EBCDIC) characters.  We do not

    176,177,178,179,180,181,182,183,184,185,186,187,188,189,190,191, /* Bx */
    192,129,130,131,132,133,134,135,136,137,202,203,204,205,206,207, /* Cx */
    208,145,146,147,148,149,150,151,152,153,218,219,220,221,222,223, /* Dx */
    224,225,162,163,164,165,166,167,168,169,232,203,204,205,206,207, /* Ex */
    239,240,241,242,243,244,245,246,247,248,249,219,220,221,222,255, /* Fx */
#endif
};

/*
** The following 256 byte lookup table is used to support SQLites built-in
** equivalents to the following standard library functions:
**
**   isspace()                        0x01
**   isalpha()                        0x02
**   isdigit()                        0x04
**   isalnum()                        0x06
**   isxdigit()                       0x08
**   toupper()                        0x20
**
** Bit 0x20 is set if the mapped character requires translation to upper
** case. i.e. if the character is a lower-case ASCII character.
** If x is a lower-case ASCII character, then its upper-case equivalent
** is (x - 0x20). Therefore toupper() can be implemented as:
**
**   (x & ~(map[x]&0x20))
**
** Standard function tolower() is implemented using the sqlite3UpperToLower[]
** array. tolower() is used more often than toupper() by SQLite.
**
** SQLite's versions are identical to the standard versions assuming a
** locale of "C". They are implemented as macros in sqliteInt.h.
*/
#ifdef SQLITE_ASCII
SQLITE_PRIVATE const unsigned char sqlite3CtypeMap[256] = {
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* 00..07    ........ */
  0x00, 0x01, 0x01, 0x01, 0x01, 0x01, 0x00, 0x00,  /* 08..0f    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* 10..17    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* 18..1f    ........ */
  0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* 20..27     !"#$%&' */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* 28..2f    ()*+,-./ */
  0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c, 0x0c,  /* 30..37    01234567 */
  0x0c, 0x0c, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* 38..3f    89:;<=>? */

  0x00, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x0a, 0x02,  /* 40..47    @ABCDEFG */
  0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,  /* 48..4f    HIJKLMNO */
  0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02, 0x02,  /* 50..57    PQRSTUVW */
  0x02, 0x02, 0x02, 0x00, 0x00, 0x00, 0x00, 0x00,  /* 58..5f    XYZ[\]^_ */
  0x00, 0x2a, 0x2a, 0x2a, 0x2a, 0x2a, 0x2a, 0x22,  /* 60..67    `abcdefg */
  0x22, 0x22, 0x22, 0x22, 0x22, 0x22, 0x22, 0x22,  /* 68..6f    hijklmno */
  0x22, 0x22, 0x22, 0x22, 0x22, 0x22, 0x22, 0x22,  /* 70..77    pqrstuvw */
  0x22, 0x22, 0x22, 0x00, 0x00, 0x00, 0x00, 0x00,  /* 78..7f    xyz{|}~. */

  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* 80..87    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* 88..8f    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* 90..97    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* 98..9f    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* a0..a7    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* a8..af    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* b0..b7    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* b8..bf    ........ */

  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* c0..c7    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* c8..cf    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* d0..d7    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* d8..df    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* e0..e7    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* e8..ef    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,  /* f0..f7    ........ */
  0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00   /* f8..ff    ........ */
};
#endif



/*
** The following singleton contains the global configuration for
** the SQLite library.
*/
SQLITE_PRIVATE SQLITE_WSD struct Sqlite3Config sqlite3Config = {
   SQLITE_DEFAULT_MEMSTATUS,  /* bMemstat */

** This file contains the C functions that implement date and time
** functions for SQLite.  
**
** There is only one exported symbol in this file - the function
** sqlite3RegisterDateTimeFunctions() found at the bottom of the file.
** All other code has file scope.
**
** $Id: date.c,v 1.102 2009/01/30 17:27:44 drh Exp $
**
** SQLite processes all times and dates as Julian Day numbers.  The
** dates and times are stored as the number of days since noon
** in Greenwich on November 24, 4714 B.C. according to the Gregorian
** calendar system. 
**
** 1970-01-01 00:00:00 is JD 2440587.5

**
**      Jean Meeus
**      Astronomical Algorithms, 2nd Edition, 1998
**      ISBM 0-943396-61-1
**      Willmann-Bell, Inc
**      Richmond, Virginia (USA)
*/

#include <time.h>

#ifndef SQLITE_OMIT_DATETIME_FUNCS

/*
** On recent Windows platforms, the localtime_s() function is available
** as part of the "Secure CRT". It is essentially equivalent to 

    N = va_arg(ap, int);
    min = va_arg(ap, int);
    max = va_arg(ap, int);
    nextC = va_arg(ap, int);
    pVal = va_arg(ap, int*);
    val = 0;
    while( N-- ){
      if( !sqlite3Isdigit(*zDate) ){
        goto end_getDigits;
      }
      val = val*10 + *zDate - '0';
      zDate++;
    }
    if( val<min || val>max || (nextC!=0 && nextC!=*zDate) ){
      goto end_getDigits;

**
** A missing specifier is not considered an error.
*/
static int parseTimezone(const char *zDate, DateTime *p){
  int sgn = 0;
  int nHr, nMn;
  int c;
  while( sqlite3Isspace(*zDate) ){ zDate++; }
  p->tz = 0;
  c = *zDate;
  if( c=='-' ){
    sgn = -1;
  }else if( c=='+' ){
    sgn = +1;
  }else if( c=='Z' || c=='z' ){
    zDate++;
    goto zulu_time;
  }else{
    return c!=0;
  }
  zDate++;
  if( getDigits(zDate, 2, 0, 14, ':', &nHr, 2, 0, 59, 0, &nMn)!=2 ){
    return 1;
  }
  zDate += 5;
  p->tz = sgn*(nMn + nHr*60);
zulu_time:
  while( sqlite3Isspace(*zDate) ){ zDate++; }
  return *zDate!=0;
}

/*
** Parse times of the form HH:MM or HH:MM:SS or HH:MM:SS.FFFF.
** The HH, MM, and SS must each be exactly 2 digits.  The
** fractional seconds FFFF can be one or more digits.

  zDate += 5;
  if( *zDate==':' ){
    zDate++;
    if( getDigits(zDate, 2, 0, 59, 0, &s)!=1 ){
      return 1;
    }
    zDate += 2;
    if( *zDate=='.' && sqlite3Isdigit(zDate[1]) ){
      double rScale = 1.0;
      zDate++;
      while( sqlite3Isdigit(*zDate) ){
        ms = ms*10.0 + *zDate - '0';
        rScale *= 10.0;
        zDate++;
      }
      ms /= rScale;
    }
  }else{

  }else{
    neg = 0;
  }
  if( getDigits(zDate,4,0,9999,'-',&Y,2,1,12,'-',&M,2,1,31,0,&D)!=3 ){
    return 1;
  }
  zDate += 10;
  while( sqlite3Isspace(*zDate) || 'T'==*(u8*)zDate ){ zDate++; }
  if( parseHhMmSs(zDate, p)==0 ){
    /* We got the time */
  }else if( *zDate==0 ){
    p->validHMS = 0;
  }else{
    return 1;
  }

    case '3':
    case '4':
    case '5':
    case '6':
    case '7':
    case '8':
    case '9': {
      double rRounder;
      n = getValue(z, &r);
      assert( n>=1 );
      if( z[n]==':' ){
        /* A modifier of the form (+|-)HH:MM:SS.FFF adds (or subtracts) the
        ** specified number of hours, minutes, seconds, and fractional seconds
        ** to the time.  The ".FFF" may be omitted.  The ":SS.FFF" may be
        ** omitted.
        */
        const char *z2 = z;
        DateTime tx;
        sqlite3_int64 day;
        if( !sqlite3Isdigit(*z2) ) z2++;
        memset(&tx, 0, sizeof(tx));
        if( parseHhMmSs(z2, &tx) ) break;
        computeJD(&tx);
        tx.iJD -= 43200000;
        day = tx.iJD/86400000;
        tx.iJD -= day*86400000;
        if( z[0]=='-' ) tx.iJD = -tx.iJD;
        computeJD(p);
        clearYMD_HMS_TZ(p);
        p->iJD += tx.iJD;
        rc = 0;
        break;
      }
      z += n;
      while( sqlite3Isspace(*z) ) z++;
      n = sqlite3Strlen30(z);
      if( n>10 || n<3 ) break;
      if( z[n-1]=='s' ){ z[n-1] = 0; n--; }
      computeJD(p);
      rc = 0;
      rRounder = r<0 ? -0.5 : +0.5;
      if( n==3 && strcmp(z,"day")==0 ){
        p->iJD += (sqlite3_int64)(r*86400000.0 + rRounder);
      }else if( n==4 && strcmp(z,"hour")==0 ){
        p->iJD += (sqlite3_int64)(r*(86400000.0/24.0) + rRounder);
      }else if( n==6 && strcmp(z,"minute")==0 ){
        p->iJD += (sqlite3_int64)(r*(86400000.0/(24.0*60.0)) + rRounder);
      }else if( n==6 && strcmp(z,"second")==0 ){
        p->iJD += (sqlite3_int64)(r*(86400000.0/(24.0*60.0*60.0)) + rRounder);
      }else if( n==5 && strcmp(z,"month")==0 ){
        int x, y;
        computeYMD_HMS(p);
        p->M += (int)r;
        x = p->M>0 ? (p->M-1)/12 : (p->M-12)/12;
        p->Y += x;
        p->M -= x*12;
        p->validJD = 0;
        computeJD(p);
        y = (int)r;
        if( y!=r ){
          p->iJD += (sqlite3_int64)((r - y)*30.0*86400000.0 + rRounder);
        }
      }else if( n==4 && strcmp(z,"year")==0 ){
        int y = (int)r;
        computeYMD_HMS(p);
        p->Y += y;
        p->validJD = 0;
        computeJD(p);
        if( y!=r ){
          p->iJD += (sqlite3_int64)((r - y)*365.0*86400000.0 + rRounder);
        }
      }else{
        rc = 1;
      }
      clearYMD_HMS_TZ(p);
      break;
    }
    default: {

          break;
        default:
          return;  /* ERROR.  return a NULL */
      }
      i++;
    }
  }
  testcase( n==sizeof(zBuf)-1 );
  testcase( n==sizeof(zBuf) );
  testcase( n==(u64)db->aLimit[SQLITE_LIMIT_LENGTH]+1 );
  testcase( n==(u64)db->aLimit[SQLITE_LIMIT_LENGTH] );
  if( n<sizeof(zBuf) ){
    z = zBuf;
  }else if( n>(u64)db->aLimit[SQLITE_LIMIT_LENGTH] ){
    sqlite3_result_error_toobig(context);
    return;
  }else{
    z = sqlite3DbMallocRaw(db, (int)n);

#ifndef SQLITE_OMIT_LOAD_EXTENSION
SQLITE_PRIVATE void *sqlite3OsDlOpen(sqlite3_vfs *pVfs, const char *zPath){
  return pVfs->xDlOpen(pVfs, zPath);
}
SQLITE_PRIVATE void sqlite3OsDlError(sqlite3_vfs *pVfs, int nByte, char *zBufOut){
  pVfs->xDlError(pVfs, nByte, zBufOut);
}
SQLITE_PRIVATE void (*sqlite3OsDlSym(sqlite3_vfs *pVfs, void *pHdle, const char *zSym))(void){
  return pVfs->xDlSym(pVfs, pHdle, zSym);
}
SQLITE_PRIVATE void sqlite3OsDlClose(sqlite3_vfs *pVfs, void *pHandle){
  pVfs->xDlClose(pVfs, pHandle);
}
#endif /* SQLITE_OMIT_LOAD_EXTENSION */
SQLITE_PRIVATE int sqlite3OsRandomness(sqlite3_vfs *pVfs, int nByte, char *zBufOut){

**    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.
**
*************************************************************************
** This file contains the C functions that implement mutexes for win32
**
** $Id: mutex_w32.c,v 1.15 2009/01/30 16:09:23 shane Exp $
*/

/*
** The code in this file is only used if we are compiling multithreaded
** on a win32 system.
*/
#ifdef SQLITE_MUTEX_W32

  */
#if 0
  if( mutexIsNT() && TryEnterCriticalSection(&p->mutex) ){
    p->owner = GetCurrentThreadId();
    p->nRef++;
    rc = SQLITE_OK;
  }
#else
  UNUSED_PARAMETER(p);
#endif
  return rc;
}

/*
** The sqlite3_mutex_leave() routine exits a mutex that was
** previously entered by the same thread.  The behavior

**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
*************************************************************************
**
** Memory allocation functions used throughout sqlite.
**
** $Id: malloc.c,v 1.54 2009/01/20 16:53:41 danielk1977 Exp $
*/

/*
** This routine runs when the memory allocator sees that the
** total memory allocation is about to exceed the soft heap
** limit.
*/

*************************************************************************
** This is the header file for information that is private to the
** VDBE.  This information used to all be at the top of the single
** source code file "vdbe.c".  When that file became too big (over
** 6000 lines long) it was split up into several smaller files and
** this header information was factored out.
**
** $Id: vdbeInt.h,v 1.162 2009/02/03 15:39:01 drh Exp $
*/
#ifndef _VDBEINT_H_
#define _VDBEINT_H_

/*
** intToKey() and keyToInt() used to transform the rowid.  But with
** the latest versions of the design they are no-ops.

SQLITE_PRIVATE const char *sqlite3OpcodeName(int);
SQLITE_PRIVATE int sqlite3VdbeOpcodeHasProperty(int, int);
SQLITE_PRIVATE int sqlite3VdbeMemGrow(Mem *pMem, int n, int preserve);
#ifdef SQLITE_ENABLE_MEMORY_MANAGEMENT
SQLITE_PRIVATE int sqlite3VdbeReleaseBuffers(Vdbe *p);
#endif




SQLITE_PRIVATE int sqlite3VdbeMemTranslate(Mem*, u8);
#ifdef SQLITE_DEBUG
SQLITE_PRIVATE   void sqlite3VdbePrintSql(Vdbe*);
SQLITE_PRIVATE   void sqlite3VdbeMemPrettyPrint(Mem *pMem, char *zBuf);
#endif
SQLITE_PRIVATE int sqlite3VdbeMemHandleBom(Mem *pMem);

**
*************************************************************************
** Utility functions used throughout sqlite.
**
** This file contains functions for allocating memory, comparing
** strings, and stuff like that.
**
** $Id: util.c,v 1.247 2009/01/20 16:53:41 danielk1977 Exp $
*/


/*
** Routine needed to support the testcase() macro.
*/
#ifdef SQLITE_COVERAGE_TEST

**
** An empty string is considered non-numeric.
*/
SQLITE_PRIVATE int sqlite3IsNumber(const char *z, int *realnum, u8 enc){
  int incr = (enc==SQLITE_UTF8?1:2);
  if( enc==SQLITE_UTF16BE ) z++;
  if( *z=='-' || *z=='+' ) z += incr;
  if( !sqlite3Isdigit(*z) ){
    return 0;
  }
  z += incr;
  if( realnum ) *realnum = 0;
  while( sqlite3Isdigit(*z) ){ z += incr; }
  if( *z=='.' ){
    z += incr;
    if( !sqlite3Isdigit(*z) ) return 0;
    while( sqlite3Isdigit(*z) ){ z += incr; }
    if( realnum ) *realnum = 1;
  }
  if( *z=='e' || *z=='E' ){
    z += incr;
    if( *z=='+' || *z=='-' ) z += incr;
    if( !sqlite3Isdigit(*z) ) return 0;
    while( sqlite3Isdigit(*z) ){ z += incr; }
    if( realnum ) *realnum = 1;
  }
  return *z==0;
}

/*
** The string z[] is an ascii representation of a real number.

*/
SQLITE_PRIVATE int sqlite3AtoF(const char *z, double *pResult){
#ifndef SQLITE_OMIT_FLOATING_POINT
  int sign = 1;
  const char *zBegin = z;
  LONGDOUBLE_TYPE v1 = 0.0;
  int nSignificant = 0;
  while( sqlite3Isspace(*z) ) z++;
  if( *z=='-' ){
    sign = -1;
    z++;
  }else if( *z=='+' ){
    z++;
  }
  while( z[0]=='0' ){
    z++;
  }
  while( sqlite3Isdigit(*z) ){
    v1 = v1*10.0 + (*z - '0');
    z++;
    nSignificant++;
  }
  if( *z=='.' ){
    LONGDOUBLE_TYPE divisor = 1.0;
    z++;
    if( nSignificant==0 ){
      while( z[0]=='0' ){
        divisor *= 10.0;
        z++;
      }
    }
    while( sqlite3Isdigit(*z) ){
      if( nSignificant<18 ){
        v1 = v1*10.0 + (*z - '0');
        divisor *= 10.0;
        nSignificant++;
      }
      z++;
    }
    v1 /= divisor;
  }
  if( *z=='e' || *z=='E' ){
    int esign = 1;
    int eval = 0;
    LONGDOUBLE_TYPE scale = 1.0;
    z++;
    if( *z=='-' ){
      esign = -1;
      z++;
    }else if( *z=='+' ){
      z++;
    }
    while( sqlite3Isdigit(*z) ){
      eval = eval*10 + *z - '0';
      z++;
    }
    while( eval>=64 ){ scale *= 1.0e+64; eval -= 64; }
    while( eval>=16 ){ scale *= 1.0e+16; eval -= 16; }
    while( eval>=4 ){ scale *= 1.0e+4; eval -= 4; }
    while( eval>=1 ){ scale *= 1.0e+1; eval -= 1; }

** atoi() library routine in RedHat 7.2.
*/
SQLITE_PRIVATE int sqlite3Atoi64(const char *zNum, i64 *pNum){
  i64 v = 0;
  int neg;
  int i, c;
  const char *zStart;
  while( sqlite3Isspace(*zNum) ) zNum++;
  if( *zNum=='-' ){
    neg = 1;
    zNum++;
  }else if( *zNum=='+' ){
    neg = 0;
    zNum++;
  }else{

     /*  56 */ "SeekLe",
     /*  57 */ "IncrVacuum",
     /*  58 */ "IdxRowid",
     /*  59 */ "ResetCount",
     /*  60 */ "ContextPush",
     /*  61 */ "Yield",
     /*  62 */ "DropTrigger",
     /*  63 */ "DropIndex",
     /*  64 */ "IdxGE",
     /*  65 */ "Or",
     /*  66 */ "And",
     /*  67 */ "IdxDelete",
     /*  68 */ "Vacuum",
     /*  69 */ "IfNot",
     /*  70 */ "IsNull",
     /*  71 */ "NotNull",
     /*  72 */ "Ne",
     /*  73 */ "Eq",
     /*  74 */ "Gt",
     /*  75 */ "Le",
     /*  76 */ "Lt",
     /*  77 */ "Ge",
     /*  78 */ "DropTable",
     /*  79 */ "BitAnd",
     /*  80 */ "BitOr",
     /*  81 */ "ShiftLeft",
     /*  82 */ "ShiftRight",
     /*  83 */ "Add",
     /*  84 */ "Subtract",
     /*  85 */ "Multiply",
     /*  86 */ "Divide",
     /*  87 */ "Remainder",
     /*  88 */ "Concat",
     /*  89 */ "SeekLt",
     /*  90 */ "MakeRecord",
     /*  91 */ "ResultRow",
     /*  92 */ "BitNot",
     /*  93 */ "String8",
     /*  94 */ "Delete",
     /*  95 */ "AggFinal",
     /*  96 */ "Compare",
     /*  97 */ "Goto",
     /*  98 */ "TableLock",
     /*  99 */ "Clear",
     /* 100 */ "VerifyCookie",

     /* 122 */ "VBegin",
     /* 123 */ "VUpdate",
     /* 124 */ "IfZero",
     /* 125 */ "VCreate",
     /* 126 */ "Found",
     /* 127 */ "IfPos",
     /* 128 */ "NullRow",
     /* 129 */ "Jump",
     /* 130 */ "Real",
     /* 131 */ "Permutation",
     /* 132 */ "NotUsed_132",
     /* 133 */ "NotUsed_133",
     /* 134 */ "NotUsed_134",
     /* 135 */ "NotUsed_135",
     /* 136 */ "NotUsed_136",
     /* 137 */ "NotUsed_137",
     /* 138 */ "NotUsed_138",
     /* 139 */ "ToText",
     /* 140 */ "ToBlob",
     /* 141 */ "ToNumeric",
     /* 142 */ "ToInt",
     /* 143 */ "ToReal",
  };
  return azName[i];
}
#endif

/************** End of opcodes.c *********************************************/
/************** Begin file os_os2.c ******************************************/

**   *  Definitions of sqlite3_io_methods objects for all locking
**      methods plus "finder" functions for each locking method.
**   *  sqlite3_vfs method implementations.
**   *  Locking primitives for the proxy uber-locking-method. (MacOSX only)
**   *  Definitions of sqlite3_vfs objects for all locking methods
**      plus implementations of sqlite3_os_init() and sqlite3_os_end().
**
** $Id: os_unix.c,v 1.239 2009/02/03 15:27:02 drh Exp $
*/
#if SQLITE_OS_UNIX              /* This file is used on unix only */

/*
** There are various methods for file locking used for concurrency
** control:
**

    pFile->lastErrno = errno;
#ifdef EOVERFLOW
    if( pFile->lastErrno==EOVERFLOW ) return SQLITE_NOLFS;
#endif
    return SQLITE_IOERR;
  }

#ifdef __APPLE__
  /* On OS X on an msdos filesystem, the inode number is reported
  ** incorrectly for zero-size files.  See ticket #3260.  To work
  ** around this problem (we consider it a bug in OS X, not SQLite)
  ** we always increase the file size to 1 by writing a single byte
  ** prior to accessing the inode number.  The one byte written is
  ** an ASCII 'S' character which also happens to be the first byte
  ** in the header of every SQLite database.  In this way, if there
  ** is a race condition such that another thread has already populated
  ** the first page of the database, no damage is done.
  */
  if( statbuf.st_size==0 ){
    rc = write(fd, "S", 1);
    if( rc!=1 ){
      return SQLITE_IOERR;
    }
    rc = fstat(fd, &statbuf);
    if( rc!=0 ){
      pFile->lastErrno = errno;
      return SQLITE_IOERR;
    }
  }
#endif

  memset(&lockKey, 0, sizeof(lockKey));
  lockKey.fid.dev = statbuf.st_dev;
#if OS_VXWORKS
  lockKey.fid.pId = pFile->pId;
#else
  lockKey.fid.ino = statbuf.st_ino;

  ** has changed.  If the transaction counter is modified, record that
  ** fact too.
  */
  if( ((unixFile*)id)->inNormalWrite ){
    unixFile *pFile = (unixFile*)id;
    pFile->dbUpdate = 1;  /* The database has been modified */
    if( offset<=24 && offset+amt>=27 ){
      int rc;
      char oldCntr[4];
      SimulateIOErrorBenign(1);
      rc = seekAndRead(pFile, 24, oldCntr, 4);
      SimulateIOErrorBenign(0);
      if( rc!=4 || memcmp(oldCntr, &((char*)pBuf)[24-offset], 4)!=0 ){
        pFile->transCntrChng = 1;  /* The transaction counter has changed */
      }
    }
  }
#endif

  while( amt>0 && (wrote = seekAndWrite((unixFile*)id, offset, pBuf, amt))>0 ){

**    May you find forgiveness for yourself and forgive others.
**    May you share freely, never taking more than you give.
**
******************************************************************************
**
** This file contains code that is specific to windows.
**
** $Id: os_win.c,v 1.146 2009/01/30 05:59:11 shane Exp $
*/
#if SQLITE_OS_WIN               /* This file is used for windows only */


/*
** A Note About Memory Allocation:
**

*/
typedef struct winFile winFile;
struct winFile {
  const sqlite3_io_methods *pMethod;/* Must be first */
  HANDLE h;               /* Handle for accessing the file */
  unsigned char locktype; /* Type of lock currently held on this file */
  short sharedLockByte;   /* Randomly chosen byte used as a shared lock */
  DWORD lastErrno;        /* The Windows errno from the last I/O error */
#if SQLITE_OS_WINCE
  WCHAR *zDeleteOnClose;  /* Name of file to delete when closing */
  HANDLE hMutex;          /* Mutex used to control access to shared lock */  
  HANDLE hShared;         /* Shared memory segment used for locking */
  winceLock local;        /* Locks obtained by this instance of winFile */
  winceLock *shared;      /* Global shared lock memory for the file  */
#endif

  for (;*zTok;zTok++){
    if (*zTok == '\\') *zTok = '_';
  }

  /* Create/open the named mutex */
  pFile->hMutex = CreateMutexW(NULL, FALSE, zName);
  if (!pFile->hMutex){
    pFile->lastErrno = GetLastError();
    free(zName);
    return FALSE;
  }

  /* Acquire the mutex before continuing */
  winceMutexAcquire(pFile->hMutex);
  

  /* If we succeeded in making the shared memory handle, map it. */
  if (pFile->hShared){
    pFile->shared = (winceLock*)MapViewOfFile(pFile->hShared, 
             FILE_MAP_READ|FILE_MAP_WRITE, 0, 0, sizeof(winceLock));
    /* If mapping failed, close the shared memory handle and erase it */
    if (!pFile->shared){
      pFile->lastErrno = GetLastError();
      CloseHandle(pFile->hShared);
      pFile->hShared = NULL;
    }
  }

  /* If shared memory could not be created, then close the mutex and fail */
  if (pFile->hShared == NULL){

  sqlite3_int64 offset       /* Begin reading at this offset */
){
  LONG upperBits = (LONG)((offset>>32) & 0x7fffffff);
  LONG lowerBits = (LONG)(offset & 0xffffffff);
  DWORD rc;
  DWORD got;
  winFile *pFile = (winFile*)id;
  DWORD error;
  assert( id!=0 );
  SimulateIOError(return SQLITE_IOERR_READ);
  OSTRACE3("READ %d lock=%d\n", pFile->h, pFile->locktype);
  rc = SetFilePointer(pFile->h, lowerBits, &upperBits, FILE_BEGIN);
  if( rc==INVALID_SET_FILE_POINTER && (error=GetLastError())!=NO_ERROR ){
    pFile->lastErrno = error;
    return SQLITE_FULL;
  }
  if( !ReadFile(pFile->h, pBuf, amt, &got, 0) ){
    pFile->lastErrno = GetLastError();
    return SQLITE_IOERR_READ;
  }
  if( got==(DWORD)amt ){
    return SQLITE_OK;
  }else{
    /* Unread parts of the buffer must be zero-filled */
    memset(&((char*)pBuf)[got], 0, amt-got);

  sqlite3_int64 offset      /* Offset into the file to begin writing at */
){
  LONG upperBits = (LONG)((offset>>32) & 0x7fffffff);
  LONG lowerBits = (LONG)(offset & 0xffffffff);
  DWORD rc;
  DWORD wrote = 0;
  winFile *pFile = (winFile*)id;
  DWORD error;
  assert( id!=0 );
  SimulateIOError(return SQLITE_IOERR_WRITE);
  SimulateDiskfullError(return SQLITE_FULL);
  OSTRACE3("WRITE %d lock=%d\n", pFile->h, pFile->locktype);
  rc = SetFilePointer(pFile->h, lowerBits, &upperBits, FILE_BEGIN);
  if( rc==INVALID_SET_FILE_POINTER && (error=GetLastError())!=NO_ERROR ){
    pFile->lastErrno = error;
    return SQLITE_FULL;
  }
  assert( amt>0 );
  while(
     amt>0
     && (rc = WriteFile(pFile->h, pBuf, amt, &wrote, 0))!=0
     && wrote>0
  ){
    amt -= wrote;
    pBuf = &((char*)pBuf)[wrote];
  }
  if( !rc || amt>(int)wrote ){
    pFile->lastErrno = GetLastError();
    return SQLITE_FULL;
  }
  return SQLITE_OK;
}

/*
** Truncate an open file to a specified size

  rc = SetFilePointer(pFile->h, lowerBits, &upperBits, FILE_BEGIN);
  if( INVALID_SET_FILE_POINTER != rc ){
    /* SetEndOfFile will fail if nByte is negative */
    if( SetEndOfFile(pFile->h) ){
      return SQLITE_OK;
    }
  }
  pFile->lastErrno = GetLastError();
  return SQLITE_IOERR_TRUNCATE;
}

#ifdef SQLITE_TEST
/*
** Count the number of fullsyncs and normal syncs.  This is used to test
** that syncs and fullsyncs are occuring at the right times.

  */
#ifdef SQLITE_NO_SYNC
    return SQLITE_OK;
#else
  if( FlushFileBuffers(pFile->h) ){
    return SQLITE_OK;
  }else{
    pFile->lastErrno = GetLastError();
    return SQLITE_IOERR;
  }
#endif
}

/*
** Determine the current size of a file in bytes
*/
static int winFileSize(sqlite3_file *id, sqlite3_int64 *pSize){
  winFile *pFile = (winFile*)id;
  DWORD upperBits, lowerBits;
  DWORD error;
  SimulateIOError(return SQLITE_IOERR_FSTAT);
  lowerBits = GetFileSize(pFile->h, &upperBits);
  if(   (lowerBits == INVALID_FILE_SIZE)
     && ((error = GetLastError()) != NO_ERROR) )
  {
    pFile->lastErrno = error;
    return SQLITE_IOERR_FSTAT;
  }
  *pSize = (((sqlite3_int64)upperBits)<<32) + lowerBits;
  return SQLITE_OK;
}

/*
** LOCKFILE_FAIL_IMMEDIATELY is undefined on some Windows systems.
*/

  }else{
    int lk;
    sqlite3_randomness(sizeof(lk), &lk);
    pFile->sharedLockByte = (short)((lk & 0x7fffffff)%(SHARED_SIZE - 1));
    res = LockFile(pFile->h, SHARED_FIRST+pFile->sharedLockByte, 0, 1, 0);
#endif
  }
  if( res == 0 ){
    pFile->lastErrno = GetLastError();
  }
  return res;
}

/*
** Undo a readlock
*/
static int unlockReadLock(winFile *pFile){
  int res;
  if( isNT() ){
    res = UnlockFile(pFile->h, SHARED_FIRST, 0, SHARED_SIZE, 0);
/* isNT() is 1 if SQLITE_OS_WINCE==1, so this else is never executed. 
*/
#if SQLITE_OS_WINCE==0
  }else{
    res = UnlockFile(pFile->h, SHARED_FIRST + pFile->sharedLockByte, 0, 1, 0);
#endif
  }
  if( res == 0 ){
    pFile->lastErrno = GetLastError();
  }
  return res;
}

/*
** Lock the file with the lock specified by parameter locktype - one
** of the following:
**

*/
static int winLock(sqlite3_file *id, int locktype){
  int rc = SQLITE_OK;    /* Return code from subroutines */
  int res = 1;           /* Result of a windows lock call */
  int newLocktype;       /* Set pFile->locktype to this value before exiting */
  int gotPendingLock = 0;/* True if we acquired a PENDING lock this time */
  winFile *pFile = (winFile*)id;
  DWORD error = NO_ERROR;

  assert( pFile!=0 );
  OSTRACE5("LOCK %d %d was %d(%d)\n",
          pFile->h, locktype, pFile->locktype, pFile->sharedLockByte);

  /* If there is already a lock of this type or more restrictive on the
  ** OsFile, do nothing. Don't use the end_lock: exit path, as

  assert( locktype!=RESERVED_LOCK || pFile->locktype==SHARED_LOCK );

  /* Lock the PENDING_LOCK byte if we need to acquire a PENDING lock or
  ** a SHARED lock.  If we are acquiring a SHARED lock, the acquisition of
  ** the PENDING_LOCK byte is temporary.
  */
  newLocktype = pFile->locktype;
  if(   (pFile->locktype==NO_LOCK)
     || (   (locktype==EXCLUSIVE_LOCK)
         && (pFile->locktype==RESERVED_LOCK))
  ){
    int cnt = 3;
    while( cnt-->0 && (res = LockFile(pFile->h, PENDING_BYTE, 0, 1, 0))==0 ){
      /* Try 3 times to get the pending lock.  The pending lock might be
      ** held by another reader process who will release it momentarily.
      */
      OSTRACE2("could not get a PENDING lock. cnt=%d\n", cnt);
      Sleep(1);
    }
    gotPendingLock = res;
    if( !res ){
      error = GetLastError();
    }
  }

  /* Acquire a shared lock
  */
  if( locktype==SHARED_LOCK && res ){
    assert( pFile->locktype==NO_LOCK );
    res = getReadLock(pFile);
    if( res ){
      newLocktype = SHARED_LOCK;
    }else{
      error = GetLastError();
    }
  }

  /* Acquire a RESERVED lock
  */
  if( locktype==RESERVED_LOCK && res ){
    assert( pFile->locktype==SHARED_LOCK );
    res = LockFile(pFile->h, RESERVED_BYTE, 0, 1, 0);
    if( res ){
      newLocktype = RESERVED_LOCK;
    }else{
      error = GetLastError();
    }
  }

  /* Acquire a PENDING lock
  */
  if( locktype==EXCLUSIVE_LOCK && res ){
    newLocktype = PENDING_LOCK;
    gotPendingLock = 0;
  }

  /* Acquire an EXCLUSIVE lock
  */
  if( locktype==EXCLUSIVE_LOCK && res ){
    assert( pFile->locktype>=SHARED_LOCK );
    res = unlockReadLock(pFile);
    OSTRACE2("unreadlock = %d\n", res);
    res = LockFile(pFile->h, SHARED_FIRST, 0, SHARED_SIZE, 0);
    if( res ){
      newLocktype = EXCLUSIVE_LOCK;
    }else{
      error = GetLastError();
      OSTRACE2("error-code = %d\n", error);
      getReadLock(pFile);
    }
  }

  /* If we are holding a PENDING lock that ought to be released, then
  ** release it now.
  */
  if( gotPendingLock && locktype==SHARED_LOCK ){
    UnlockFile(pFile->h, PENDING_BYTE, 0, 1, 0);
  }

  /* Update the state of the lock has held in the file descriptor then
  ** return the appropriate result code.
  */
  if( res ){
    rc = SQLITE_OK;
  }else{
    OSTRACE4("LOCK FAILED %d trying for %d but got %d\n", pFile->h,
           locktype, newLocktype);
    pFile->lastErrno = error;
    rc = SQLITE_BUSY;
  }
  pFile->locktype = (u8)newLocktype;
  return rc;
}

/*

** Control and query of the open file handle.
*/
static int winFileControl(sqlite3_file *id, int op, void *pArg){
  switch( op ){
    case SQLITE_FCNTL_LOCKSTATE: {
      *(int*)pArg = ((winFile*)id)->locktype;
      return SQLITE_OK;
    }
    case SQLITE_LAST_ERRNO: {
      *(int*)pArg = (int)((winFile*)id)->lastErrno;
      return SQLITE_OK;
    }
  }
  return SQLITE_ERROR;
}

/*
** Return the sector size in bytes of the underlying block device for

    }else{
      *pOutFlags = SQLITE_OPEN_READONLY;
    }
  }
  memset(pFile, 0, sizeof(*pFile));
  pFile->pMethod = &winIoMethod;
  pFile->h = h;
  pFile->lastErrno = NO_ERROR;
#if SQLITE_OS_WINCE
  if( (flags & (SQLITE_OPEN_READWRITE|SQLITE_OPEN_MAIN_DB)) ==
               (SQLITE_OPEN_READWRITE|SQLITE_OPEN_MAIN_DB)
       && !winceCreateLock(zName, pFile)
  ){
    CloseHandle(h);
    free(zConverted);

** Clear operations are exceedingly rare.  There are usually between
** 5 and 500 set operations per Bitvec object, though the number of sets can
** sometimes grow into tens of thousands or larger.  The size of the
** Bitvec object is the number of pages in the database file at the
** start of a transaction, and is thus usually less than a few thousand,
** but can be as large as 2 billion for a really big database.
**
** @(#) $Id: bitvec.c,v 1.13 2009/01/20 17:06:27 danielk1977 Exp $
*/

/* Size of the Bitvec structure in bytes. */
#define BITVEC_SZ        512

/* Round the union size down to the nearest pointer boundary, since that's how 
** it will be aligned within the Bitvec struct. */

    unsigned int i;
    for(i=0; i<BITVEC_NPTR; i++){
      sqlite3BitvecDestroy(p->u.apSub[i]);
    }
  }
  sqlite3_free(p);
}

/*
** Return the value of the iSize parameter specified when Bitvec *p
** was created.
*/
SQLITE_PRIVATE u32 sqlite3BitvecSize(Bitvec *p){
  return p->iSize;
}

#ifndef SQLITE_OMIT_BUILTIN_TEST
/*
** Let V[] be an array of unsigned characters sufficient to hold
** up to N bits.  Let I be an integer between 0 and N.  0<=I<N.
** Then the following macros can be used to set, clear, or test
** individual bits within V.

**    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.
**
*************************************************************************
** This file implements that page cache.
**
** @(#) $Id: pcache.c,v 1.43 2009/01/23 16:45:01 danielk1977 Exp $
*/

/*
** A complete page cache is an instance of this structure.
*/
struct PCache {
  PgHdr *pDirty, *pDirtyTail;         /* List of dirty pages in LRU order */
  PgHdr *pSynced;                     /* Last synced page in dirty page list */
  int nRef;                           /* Number of referenced pages */
  int nMax;                           /* Configured cache size */

  int szPage;                         /* Size of every page in this cache */
  int szExtra;                        /* Size of extra space for each page */
  int bPurgeable;                     /* True if pages are on backing store */
  int (*xStress)(void*,PgHdr*);       /* Call to try make a page clean */
  void *pStress;                      /* Argument to xStress */
  sqlite3_pcache *pCache;             /* Pluggable cache module */
  PgHdr *pPage1;                      /* Reference to page 1 */
};

/*
** Some of the assert() macros in this code are too expensive to run
** even during normal debugging.  Use them only rarely on long-running
** tests.  Enable the expensive asserts using the
** -DSQLITE_ENABLE_EXPENSIVE_ASSERT=1 compile-time option.

  memset(p, 0, sizeof(PCache));
  p->szPage = szPage;
  p->szExtra = szExtra;
  p->bPurgeable = bPurgeable;
  p->xStress = xStress;
  p->pStress = pStress;
  p->nMax = 100;

}

/*
** Change the page size for PCache object. The caller must ensure that there
** are no outstanding page references when this function is called.
*/
SQLITE_PRIVATE void sqlite3PcacheSetPageSize(PCache *pCache, int szPage){

      }
    }

    pPage = sqlite3GlobalConfig.pcache.xFetch(pCache->pCache, pgno, 2);
  }

  if( pPage ){
    if( !pPage->pData ){
      memset(pPage, 0, sizeof(PgHdr) + pCache->szExtra);
      pPage->pExtra = (void*)&pPage[1];
      pPage->pData = (void *)&((char *)pPage)[sizeof(PgHdr) + pCache->szExtra];
      pPage->pCache = pCache;
      pPage->pgno = pgno;
    }
    assert( pPage->pCache==pCache );
    assert( pPage->pgno==pgno );
    assert( pPage->pExtra==(void *)&pPage[1] );

    if( 0==pPage->nRef ){
      pCache->nRef++;
    }
    pPage->nRef++;




    if( pgno==1 ){
      pCache->pPage1 = pPage;
    }
  }
  *ppPage = pPage;
  return (pPage==0 && eCreate) ? SQLITE_NOMEM : SQLITE_OK;
}

    sqlite3GlobalConfig.pcache.xDestroy(pCache->pCache);
  }
}

/* 
** Discard the contents of the cache.
*/
SQLITE_PRIVATE void sqlite3PcacheClear(PCache *pCache){
  sqlite3PcacheTruncate(pCache, 0);

}

/*
** Merge two lists of pages connected by pDirty and in pgno order.
** Do not both fixing the pDirtyPrev pointers.
*/
static PgHdr *pcacheMergeDirtyList(PgHdr *pA, PgHdr *pB){

**
** This file implements the default page cache implementation (the
** sqlite3_pcache interface). It also contains part of the implementation
** of the SQLITE_CONFIG_PAGECACHE and sqlite3_release_memory() features.
** If the default page cache implementation is overriden, then neither of
** these two features are available.
**
** @(#) $Id: pcache1.c,v 1.8 2009/01/23 16:45:01 danielk1977 Exp $
*/


typedef struct PCache1 PCache1;
typedef struct PgHdr1 PgHdr1;
typedef struct PgFreeslot PgFreeslot;

/*
** Allocate a new page object initially associated with cache pCache.
*/
static PgHdr1 *pcache1AllocPage(PCache1 *pCache){
  int nByte = sizeof(PgHdr1) + pCache->szPage;
  PgHdr1 *p = (PgHdr1 *)pcache1Alloc(nByte);
  if( p ){

    if( pCache->bPurgeable ){
      pcache1.nCurrentPage++;
    }
  }
  return p;
}

  */
  if( !pPage ){
    pPage = pcache1AllocPage(pCache);
  }

  if( pPage ){
    unsigned int h = iKey % pCache->nHash;
    *(void **)(PGHDR1_TO_PAGE(pPage)) = 0;
    pCache->nPage++;
    pPage->iKey = iKey;
    pPage->pNext = pCache->apHash[h];
    pPage->pCache = pCache;
    pPage->pLruPrev = 0;
    pPage->pLruNext = 0;
    pCache->apHash[h] = pPage;
  }

fetch_out:
  if( pPage && iKey>pCache->iMaxKey ){
    pCache->iMaxKey = iKey;
  }

** 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 1.561 2009/01/31 14:54:07 danielk1977 Exp $
*/
#ifndef SQLITE_OMIT_DISKIO

/*
** Macros for troubleshooting.  Normally turned off
*/
#if 0

  Pgno nOrig;                  /* Original number of pages in file */
  Pgno iSubRec;                /* Index of first record in sub-journal */
};

/*
** A open page cache is an instance of the following structure.
**
** errCode






**
**   Pager.errCode may be set to SQLITE_IOERR, SQLITE_CORRUPT, or
**   or SQLITE_FULL. Once one of the first three errors occurs, it persists
**   and is returned as the result of every major pager API call.  The
**   SQLITE_FULL return code is slightly different. It persists only until the
**   next successful rollback is performed on the pager cache. Also,
**   SQLITE_FULL does not affect the sqlite3PagerGet() and sqlite3PagerLookup()
**   APIs, they may still be used successfully.
**
** dbSizeValid, dbSize, dbOrigSize, dbFileSize
**
**   Managing the size of the database file in pages is a little complicated.
**   The variable Pager.dbSize contains the number of pages that the database
**   image currently contains. As the database image grows or shrinks this
**   variable is updated. The variable Pager.dbFileSize contains the number
**   of pages in the database file. This may be different from Pager.dbSize
**   if some pages have been appended to the database image but not yet written
**   out from the cache to the actual file on disk. Or if the image has been
**   truncated by an incremental-vacuum operation. The Pager.dbOrigSize variable
**   contains the number of pages in the database image when the current
**   transaction was opened. The contents of all three of these variables is
**   only guaranteed to be correct if the boolean Pager.dbSizeValid is true.
**
**   TODO: Under what conditions is dbSizeValid set? Cleared?
**
** changeCountDone
**
**   This boolean variable is used to make sure that the change-counter 
**   (the 4-byte header field at byte offset 24 of the database file) is 
**   not updated more often than necessary. 
**
**   It is set to true when the change-counter field is updated, which 
**   can only happen if an exclusive lock is held on the database file.
**   It is cleared (set to false) whenever an exclusive lock is 
**   relinquished on the database file. Each time a transaction is committed,
**   The changeCountDone flag is inspected. If it is true, the work of
**   updating the change-counter is omitted for the current transaction.
**
**   This mechanism means that when running in exclusive mode, a connection 
**   need only update the change-counter once, for the first transaction
**   committed.
**
** dbModified
**
**   The dbModified flag is set whenever a database page is dirtied.
**   It is cleared at the end of each transaction.
**
**   It is used when committing or otherwise ending a transaction. If
**   the dbModified flag is clear then less work has to be done.
**
** journalStarted
**
**   This flag is set whenever the the main journal is synced. 
**
**   The point of this flag is that it must be set after the 
**   first journal header in a journal file has been synced to disk.
**   After this has happened, new pages appended to the database 
**   do not need the PGHDR_NEED_SYNC flag set, as they do not need
**   to wait for a journal sync before they can be written out to
**   the database file (see function pager_write()).
**   
** setMaster
**
**   This variable is used to ensure that the master journal file name
**   (if any) is only written into the journal file once.
**
**   When committing a transaction, the master journal file name (if any)
**   may be written into the journal file while the pager is still in
**   PAGER_RESERVED state (see CommitPhaseOne() for the action). It
**   then attempts to upgrade to an exclusive lock. If this attempt
**   fails, then SQLITE_BUSY may be returned to the user and the user
**   may attempt to commit the transaction again later (calling
**   CommitPhaseOne() again). This flag is used to ensure that the 
**   master journal name is only written to the journal file the first
**   time CommitPhaseOne() is called.
**
** doNotSync
**
**   This variable is set and cleared by sqlite3PagerWrite().
**
** needSync
**
**   TODO: It might be easier to set this variable in writeJournalHdr()
**   and writeMasterJournal() only. Change its meaning to "unsynced data
**   has been written to the journal".
*/
struct Pager {
  sqlite3_vfs *pVfs;          /* OS functions to use for IO */
  u8 exclusiveMode;           /* Boolean. True if locking_mode==EXCLUSIVE */
  u8 journalMode;             /* On of the PAGER_JOURNALMODE_* values */
  u8 useJournal;              /* Use a rollback journal on this file */
  u8 noReadlock;              /* Do not bother to obtain readlocks */
  u8 noSync;                  /* Do not sync the journal if true */
  u8 fullSync;                /* Do extra syncs of the journal for robustness */
  u8 sync_flags;              /* One of SYNC_NORMAL or SYNC_FULL */

  u8 tempFile;                /* zFilename is a temporary file */
  u8 readOnly;                /* True for a read-only database */


  u8 memDb;                   /* True to inhibit all file I/O */

  /* The following block contains those class members that are dynamically
  ** modified during normal operations. The other variables in this structure
  ** are either constant throughout the lifetime of the pager, or else
  ** used to store configuration parameters that affect the way the pager 
  ** operates.
  **
  ** The 'state' variable is described in more detail along with the
  ** descriptions of the values it may take - PAGER_UNLOCK etc. Many of the
  ** other variables in this block are described in the comment directly 
  ** above this class definition.
  */
  u8 state;                   /* PAGER_UNLOCK, _SHARED, _RESERVED, etc. */
  u8 dbModified;              /* True if there are any changes to the Db */
  u8 needSync;                /* True if an fsync() is needed on the journal */
  u8 journalStarted;          /* True if header of journal is synced */
  u8 changeCountDone;         /* Set after incrementing the change-counter */
  u8 setMaster;               /* True if a m-j name has been written to jrnl */
  u8 doNotSync;               /* Boolean. While true, do not spill the cache */
  u8 dbSizeValid;             /* Set when dbSize is correct */
  Pgno dbSize;                /* Number of pages in the database */
  Pgno dbOrigSize;            /* dbSize before the current transaction */
  Pgno dbFileSize;            /* Number of pages in the database file */

  int errCode;                /* One of several kinds of errors */
  int nRec;                   /* Pages journalled since last j-header written */
  u32 cksumInit;              /* Quasi-random value added to every checksum */
  u32 nSubRec;                /* Number of records written to sub-journal */
  Bitvec *pInJournal;         /* One bit for each page in the database file */
  sqlite3_file *fd;           /* File descriptor for database */
  sqlite3_file *jfd;          /* File descriptor for main journal */
  sqlite3_file *sjfd;         /* File descriptor for sub-journal */
  i64 journalOff;             /* Current write offset in the journal file */
  i64 journalHdr;             /* Byte offset to previous journal header */
  PagerSavepoint *aSavepoint; /* Array of active savepoints */
  int nSavepoint;             /* Number of elements in aSavepoint[] */
  char dbFileVers[16];        /* Changes whenever database file changes */
  u32 sectorSize;             /* Assumed sector size during rollback */

  int nExtra;                 /* Add this many bytes to each in-memory page */
  u32 vfsFlags;               /* Flags for sqlite3_vfs.xOpen() */
  int pageSize;               /* Number of bytes in a page */


  Pgno mxPgno;                /* Maximum allowed size of the database */


  char *zFilename;            /* Name of the database file */
  char *zJournal;             /* Name of the journal file */



  int (*xBusyHandler)(void*); /* Function to call when busy */
  void *pBusyHandlerArg;      /* Context argument for xBusyHandler */



#ifdef SQLITE_TEST
  int nHit, nMiss;            /* Cache hits and missing */
  int nRead, nWrite;          /* Database pages read/written */
#endif
  void (*xReiniter)(DbPage*); /* Call this routine when reloading pages */
#ifdef SQLITE_HAS_CODEC
  void *(*xCodec)(void*,void*,Pgno,int); /* Routine for en/decoding data */
  void *pCodecArg;            /* First argument to xCodec() */
#endif
  char *pTmpSpace;            /* Pager.pageSize bytes of space for tmp use */

  i64 journalSizeLimit;       /* Size limit for persistent journal files */
  PCache *pPCache;            /* Pointer to page cache object */


};

/*
** The following global variables hold counters used for
** testing purposes only.  These variables do not exist in
** a non-testing build.  These variables are not thread-safe.
*/

** is different for every journal, we minimize that risk.
*/
static const unsigned char aJournalMagic[] = {
  0xd9, 0xd5, 0x05, 0xf9, 0x20, 0xa1, 0x63, 0xd7,
};

/*
** The size of the of each page record in the journal is given by
** the following macro.
*/
#define JOURNAL_PG_SZ(pPager)  ((pPager->pageSize) + 8)

/*
** The journal header size for this pager. This is usually the same 

** size as a single disk sector. See also setSectorSize().
*/
#define JOURNAL_HDR_SZ(pPager) (pPager->sectorSize)

/*
** The macro MEMDB is true if we are dealing with an in-memory database.
** We do this as a macro so that if the SQLITE_OMIT_MEMORYDB macro is set,
** the value of MEMDB will be a constant and the compiler will optimize
** out code that would never execute.
*/
#ifdef SQLITE_OMIT_MEMORYDB
# define MEMDB 0
#else
# define MEMDB pPager->memDb
#endif











/*
** The maximum legal page number is (2^31 - 1).
*/
#define PAGER_MAX_PGNO 2147483647

#ifndef NDEBUG 
/*
** Usage:
**
**   assert( assert_pager_state(pPager) );
*/
static int assert_pager_state(Pager *pPager){

  /* A temp-file is always in PAGER_EXCLUSIVE or PAGER_SYNCED state. */
  assert( pPager->tempFile==0 || pPager->state>=PAGER_EXCLUSIVE );

  /* The changeCountDone flag is always set for temp-files */
  assert( pPager->tempFile==0 || pPager->changeCountDone );

  return 1;
}
#endif

/*
** Return true if it is necessary to write page *pPg into the sub-journal.
** A page needs to be written into the sub-journal if there exists one
** or more open savepoints for which:
**
**   * The page-number is less than or equal to PagerSavepoint.nOrig, and

*/
static int write32bits(sqlite3_file *fd, i64 offset, u32 val){
  char ac[4];
  put32bits(ac, val);
  return sqlite3OsWrite(fd, ac, 4, offset);
}

/*
** The argument to this macro is a file descriptor (type sqlite3_file*).
** Return 0 if it is not open, or non-zero (but not 1) if it is.
**
** This is so that expressions can be written as:
**
**   if( isOpen(pPager->jfd) ){ ...
**
** instead of
**
**   if( pPager->jfd->pMethods ){ ...
*/
#define isOpen(pFd) ((pFd)->pMethods)

/*
** If file pFd is open, call sqlite3OsUnlock() on it.
*/
static int osUnlock(sqlite3_file *pFd, int eLock){
  if( !isOpen(pFd) ){
    return SQLITE_OK;
  }
  return sqlite3OsUnlock(pFd, eLock);
}

/*
** This function determines whether or not the atomic-write optimization
** can be used with this pager. The optimization can be used if:
**
**  (a) the value returned by OsDeviceCharacteristics() indicates that
**      a database page may be written atomically, and
**  (b) the value returned by OsSectorSize() is less than or equal
**      to the page size.
**
** The optimization is also always enabled for temporary files. It is
** an error to call this function if pPager is opened on an in-memory
** database.
**
** If the optimization cannot be used, 0 is returned. If it can be used,
** then the value returned is the size of the journal file when it
** contains rollback data for exactly one page.
*/
#ifdef SQLITE_ENABLE_ATOMIC_WRITE
static int jrnlBufferSize(Pager *pPager){
  assert( !MEMDB );
  if( !pPager->tempFile ){
    int dc;                           /* Device characteristics */
    int nSector;                      /* Sector size */
    int szPage;                       /* Page size */



    assert( isOpen(pPager->fd) );
    dc = sqlite3OsDeviceCharacteristics(pPager->fd);
    nSector = pPager->sectorSize;
    szPage = pPager->pageSize;


    assert(SQLITE_IOCAP_ATOMIC512==(512>>8));
    assert(SQLITE_IOCAP_ATOMIC64K==(65536>>8));


    if( 0==(dc&(SQLITE_IOCAP_ATOMIC|(szPage>>8)) || nSector>szPage) ){


      return 0;
    }

  }









































  return JOURNAL_HDR_SZ(pPager) + JOURNAL_PG_SZ(pPager);
}
#endif

/*
** If SQLITE_CHECK_PAGES is defined then we do some sanity checking
** on the cache using a hash function.  This is used for testing
** and debugging only.
*/
#ifdef SQLITE_CHECK_PAGES

#define pager_datahash(X,Y)  0
#define pager_pagehash(X)  0
#define CHECK_PAGE(x)
#endif  /* SQLITE_CHECK_PAGES */

/*
** When this is called the journal file for pager pPager must be open.
** This function attempts to read a master journal file name from the 
** end of the file and, if successful, copies it into memory supplied 
** by the caller. See comments above writeMasterJournal() for the format
** used to store a master journal file name at the end of a journal file.
**
** zMaster must point to a buffer of at least nMaster bytes allocated by
** the caller. This should be sqlite3_vfs.mxPathname+1 (to ensure there is
** enough space to write the master journal name). If the master journal
** name in the journal is longer than nMaster bytes (including a
** nul-terminator), then this is handled as if no master journal name
** were present in the journal.
**
** If a master journal file name is present at the end of the journal
** file, then it is copied into the buffer pointed to by zMaster. A
** nul-terminator byte is appended to the buffer following the master
** journal file name.
**
** If it is determined that no master journal file name is present 
** zMaster[0] is set to 0 and SQLITE_OK returned.
**
** If an error occurs while reading from the journal file, an SQLite
** error code is returned.
*/
static int readMasterJournal(sqlite3_file *pJrnl, char *zMaster, u32 nMaster){
  int rc;                    /* Return code */
  u32 len;                   /* Length in bytes of master journal name */
  i64 szJ;                   /* Total size in bytes of journal file pJrnl */
  u32 cksum;                 /* MJ checksum value read from journal */
  u32 u;                     /* Unsigned loop counter */
  unsigned char aMagic[8];   /* A buffer to hold the magic header */

  zMaster[0] = '\0';

  if( SQLITE_OK!=(rc = sqlite3OsFileSize(pJrnl, &szJ))
   || szJ<16

   || SQLITE_OK!=(rc = read32bits(pJrnl, szJ-16, &len))


   || len>=nMaster 



   || SQLITE_OK!=(rc = read32bits(pJrnl, szJ-12, &cksum))


   || SQLITE_OK!=(rc = sqlite3OsRead(pJrnl, aMagic, 8, szJ-8))
   || memcmp(aMagic, aJournalMagic, 8)

   || SQLITE_OK!=(rc = sqlite3OsRead(pJrnl, zMaster, len, szJ-16-len))
  ){
    return rc;
  }


  /* See if the checksum matches the master journal name */
  for(u=0; u<len; u++){
    cksum -= zMaster[u];
  }
  if( cksum ){
    /* If the checksum doesn't add up, then one or more of the disk sectors
    ** containing the master journal filename is corrupted. This means
    ** definitely roll back, so just return SQLITE_OK and report a (nul)
    ** master-journal filename.
    */
    len = 0;
  }
  zMaster[len] = '\0';
   
  return SQLITE_OK;
}

/*
** Return the offset of the sector boundary at or immediately 
** following the value in pPager->journalOff, assuming a sector 
** size of pPager->sectorSize bytes.
**
** i.e for a sector size of 512:
**
**   Pager.journalOff          Return value
**   ---------------------------------------
**   0                         0
**   512                       512
**   100                       512
**   2000                      2048
** 
*/
static i64 journalHdrOffset(Pager *pPager){
  i64 offset = 0;
  i64 c = pPager->journalOff;
  if( c ){
    offset = ((c-1)/JOURNAL_HDR_SZ(pPager) + 1) * JOURNAL_HDR_SZ(pPager);
  }
  assert( offset%JOURNAL_HDR_SZ(pPager)==0 );
  assert( offset>=c );
  assert( (offset-c)<JOURNAL_HDR_SZ(pPager) );
  return offset;
}




/*
** The journal file must be open when this function is called.
**
** This function is a no-op if the journal file has not been written to
** within the current transaction (i.e. if Pager.journalOff==0).
**
** If doTruncate is non-zero or the Pager.journalSizeLimit variable is
** set to 0, then truncate the journal file to zero bytes in size. Otherwise,
** zero the 28-byte header at the start of the journal file. In either case, 
** if the pager is not in no-sync mode, sync the journal file immediately 
** after writing or truncating it.
**
** If Pager.journalSizeLimit is set to a positive, non-zero value, and
** following the truncation or zeroing described above the size of the 
** journal file in bytes is larger than this value, then truncate the
** journal file to Pager.journalSizeLimit bytes. The journal file does
** not need to be synced following this operation.
**
** If an IO error occurs, abandon processing and return the IO error code.
** Otherwise, return SQLITE_OK.
*/
static int zeroJournalHdr(Pager *pPager, int doTruncate){
  int rc = SQLITE_OK;                               /* Return code */

  assert( isOpen(pPager->jfd) );
  if( pPager->journalOff ){
    const i64 iLimit = pPager->journalSizeLimit;    /* Local cache of jsl */

    IOTRACE(("JZEROHDR %p\n", pPager))
    if( doTruncate || iLimit==0 ){
      rc = sqlite3OsTruncate(pPager->jfd, 0);
    }else{
      static const char zeroHdr[28] = {0};
      rc = sqlite3OsWrite(pPager->jfd, zeroHdr, sizeof(zeroHdr), 0);
    }
    if( rc==SQLITE_OK && !pPager->noSync ){
      rc = sqlite3OsSync(pPager->jfd, SQLITE_SYNC_DATAONLY|pPager->sync_flags);
    }

    /* At this point the transaction is committed but the write lock 

** - 4 bytes: Initial database page count.
** - 4 bytes: Sector size used by the process that wrote this journal.
** - 4 bytes: Database page size.
** 
** Followed by (JOURNAL_HDR_SZ - 28) bytes of unused space.
*/
static int writeJournalHdr(Pager *pPager){
  int rc = SQLITE_OK;                 /* Return code */
  char *zHeader = pPager->pTmpSpace;  /* Temporary space used to build header */
  u32 nHeader = pPager->pageSize;     /* Size of buffer pointed to by zHeader */
  u32 nWrite;                         /* Bytes of header sector written */
  int ii;                             /* Loop counter */

  assert( isOpen(pPager->jfd) );      /* Journal file must be open. */

  if( nHeader>JOURNAL_HDR_SZ(pPager) ){
    nHeader = JOURNAL_HDR_SZ(pPager);
  }

  /* If there are active savepoints and any of them were created 
  ** since the most recent journal header was written, update the 
  ** PagerSavepoint.iHdrOffset fields now.
  */
  for(ii=0; ii<pPager->nSavepoint; ii++){
    if( pPager->aSavepoint[ii].iHdrOffset==0 ){
      pPager->aSavepoint[ii].iHdrOffset = pPager->journalOff;
    }
  }


  pPager->journalHdr = pPager->journalOff = journalHdrOffset(pPager);

  memcpy(zHeader, aJournalMagic, sizeof(aJournalMagic));

  /* 
  ** Write the nRec Field - the number of page records that follow this
  ** journal header. Normally, zero is written to this value at this time.
  ** After the records are added to the journal (and the journal synced, 
  ** if in full-sync mode), the zero is overwritten with the true number

  **
  **   * When the pager is in no-sync mode. Corruption can follow a
  **     power failure in this case anyway.
  **
  **   * When the SQLITE_IOCAP_SAFE_APPEND flag is set. This guarantees
  **     that garbage data is never appended to the journal file.
  */
  assert( isOpen(pPager->fd) || pPager->noSync );
  if( (pPager->noSync) || (pPager->journalMode==PAGER_JOURNALMODE_MEMORY)
   || (sqlite3OsDeviceCharacteristics(pPager->fd)&SQLITE_IOCAP_SAFE_APPEND) 
  ){
    put32bits(&zHeader[sizeof(aJournalMagic)], 0xffffffff);
  }else{
    put32bits(&zHeader[sizeof(aJournalMagic)], 0);
  }

  /* The random check-hash initialiser */ 
  sqlite3_randomness(sizeof(pPager->cksumInit), &pPager->cksumInit);
  put32bits(&zHeader[sizeof(aJournalMagic)+4], pPager->cksumInit);
  /* The initial database size */
  put32bits(&zHeader[sizeof(aJournalMagic)+8], pPager->dbOrigSize);
  /* The assumed sector size for this process */
  put32bits(&zHeader[sizeof(aJournalMagic)+12], pPager->sectorSize);

  /* The page size */
  put32bits(&zHeader[sizeof(aJournalMagic)+16], pPager->pageSize);

  /* Initializing the tail of the buffer is not necessary.  Everything
  ** works find if the following memset() is omitted.  But initializing
  ** the memory prevents valgrind from complaining, so we are willing to
  ** take the performance hit.
  */
  memset(&zHeader[sizeof(aJournalMagic)+20], 0,
         nHeader-(sizeof(aJournalMagic)+20));

  /* In theory, it is only necessary to write the 28 bytes that the 
  ** journal header consumes to the journal file here. Then increment the 
  ** Pager.journalOff variable by JOURNAL_HDR_SZ so that the next 
  ** record is written to the following sector (leaving a gap in the file
  ** that will be implicitly filled in by the OS).
  **
  ** However it has been discovered that on some systems this pattern can 
  ** be significantly slower than contiguously writing data to the file,
  ** even if that means explicitly writing data to the block of 
  ** (JOURNAL_HDR_SZ - 28) bytes that will not be used. So that is what
  ** is done. 
  **
  ** The loop is required here in case the sector-size is larger than the 
  ** database page size. Since the zHeader buffer is only Pager.pageSize


  ** bytes in size, more than one call to sqlite3OsWrite() may be required
  ** to populate the entire journal header sector.
  */ 
  for(nWrite=0; rc==SQLITE_OK&&nWrite<JOURNAL_HDR_SZ(pPager); nWrite+=nHeader){
    IOTRACE(("JHDR %p %lld %d\n", pPager, pPager->journalHdr, nHeader))
    rc = sqlite3OsWrite(pPager->jfd, zHeader, nHeader, pPager->journalOff);
    pPager->journalOff += nHeader;
  }

  return rc;
}

/*
** The journal file must be open when this is called. A journal header file
** (JOURNAL_HDR_SZ bytes) is read from the current location in the journal
** file. The current location in the journal file is given by
** pPager->journalOff. See comments above function writeJournalHdr() for
** a description of the journal header format.
**
** If the header is read successfully, *pNRec is set to the number of
** page records following this header and *pDbSize is set to the size of the
** database before the transaction began, in pages. Also, pPager->cksumInit
** is set to the value read from the journal header. SQLITE_OK is returned
** in this case.
**
** If the journal header file appears to be corrupted, SQLITE_DONE is
** returned and *pNRec and *PDbSize are undefined.  If JOURNAL_HDR_SZ bytes
** cannot be read from the journal file an error code is returned.
*/
static int readJournalHdr(
  Pager *pPager,               /* Pager object */
  i64 journalSize,             /* Size of the open journal file in bytes */
  u32 *pNRec,                  /* OUT: Value read from the nRec field */
  u32 *pDbSize                 /* OUT: Value of original database size field */
){
  int rc;                      /* Return code */
  unsigned char aMagic[8];     /* A buffer to hold the magic header */
  i64 iHdrOff;                 /* Offset of journal header being read */

  assert( isOpen(pPager->jfd) );      /* Journal file must be open. */

  /* Advance Pager.journalOff to the start of the next sector. If the
  ** journal file is too small for there to be a header stored at this
  ** point, return SQLITE_DONE.
  */
  pPager->journalOff = journalHdrOffset(pPager);
  if( pPager->journalOff+JOURNAL_HDR_SZ(pPager) > journalSize ){
    return SQLITE_DONE;
  }
  iHdrOff = pPager->journalOff;

  /* Read in the first 8 bytes of the journal header. If they do not match
  ** the  magic string found at the start of each journal header, return
  ** SQLITE_DONE. If an IO error occurs, return an error code. Otherwise,
  ** proceed.
  */
  rc = sqlite3OsRead(pPager->jfd, aMagic, sizeof(aMagic), iHdrOff);
  if( rc ){
    return rc;
  }
  if( memcmp(aMagic, aJournalMagic, sizeof(aMagic))!=0 ){
    return SQLITE_DONE;
  }

  /* Read the first three 32-bit fields of the journal header: The nRec
  ** field, the checksum-initializer and the database size at the start
  ** of the transaction. Return an error code if anything goes wrong.
  */
  if( SQLITE_OK!=(rc = read32bits(pPager->jfd, iHdrOff+8, pNRec))

   || SQLITE_OK!=(rc = read32bits(pPager->jfd, iHdrOff+12, &pPager->cksumInit))
   || SQLITE_OK!=(rc = read32bits(pPager->jfd, iHdrOff+16, pDbSize))
  ){
    return rc;
  }



  if( pPager->journalOff==0 ){
    u32 iPageSize;               /* Page-size field of journal header */
    u32 iSectorSize;             /* Sector-size field of journal header */
    u16 iPageSize16;             /* Copy of iPageSize in 16-bit variable */

    /* Read the page-size and sector-size journal header fields. */
    if( SQLITE_OK!=(rc = read32bits(pPager->jfd, iHdrOff+20, &iSectorSize))
     || SQLITE_OK!=(rc = read32bits(pPager->jfd, iHdrOff+24, &iPageSize))
    ){
      return rc;
    }

    /* Check that the values read from the page-size and sector-size fields
    ** are within range. To be 'in range', both values need to be a power
    ** of two greater than or equal to 512, and not greater than their 
    ** respective compile time maximum limits.
    */
    if( iPageSize<512                  || iSectorSize<512
     || iPageSize>SQLITE_MAX_PAGE_SIZE || iSectorSize>MAX_SECTOR_SIZE
     || ((iPageSize-1)&iPageSize)!=0   || ((iSectorSize-1)&iSectorSize)!=0 
    ){
      /* If the either the page-size or sector-size in the journal-header is 
      ** invalid, then the process that wrote the journal-header must have 
      ** crashed before the header was synced. In this case stop reading 
      ** the journal file here.
      */
      return SQLITE_DONE;
    }

    /* Update the page-size to match the value read from the journal. 
    ** Use a testcase() macro to make sure that malloc failure within 
    ** PagerSetPagesize() is tested.
    */
    iPageSize16 = (u16)iPageSize;
    rc = sqlite3PagerSetPagesize(pPager, &iPageSize16);
    testcase( rc!=SQLITE_OK );

    assert( rc!=SQLITE_OK || iPageSize16==(u16)iPageSize );

    /* Update the assumed sector-size to match the value used by 
    ** the process that created this journal. If this journal was
    ** created by a process other than this one, then this routine
    ** is being called from within pager_playback(). The local value
    ** of Pager.sectorSize is restored at the end of that routine.
    */








    pPager->sectorSize = iSectorSize;
  }

  pPager->journalOff += JOURNAL_HDR_SZ(pPager);
  return rc;
}


/*
** Write the supplied master journal name into the journal file for pager
** pPager at the current location. The master journal name must be the last
** thing written to a journal file. If the pager is in full-sync mode, the
** journal file descriptor is advanced to the next sector boundary before
** anything is written. The format is:
**
**   + 4 bytes: PAGER_MJ_PGNO.
**   + N bytes: Master journal filename in utf-8.
**   + 4 bytes: N (length of master journal name in bytes, no nul-terminator).
**   + 4 bytes: Master journal name checksum.
**   + 8 bytes: aJournalMagic[].
**
** The master journal page checksum is the sum of the bytes in the master
** journal name, where each byte is interpreted as a signed 8-bit integer.
**
** If zMaster is a NULL pointer (occurs for a single database transaction), 
** this call is a no-op.
*/
static int writeMasterJournal(Pager *pPager, const char *zMaster){
  int rc;                          /* Return code */

  int nMaster;                     /* Length of string zMaster */
  i64 iHdrOff;                     /* Offset of header in journal file */
  i64 jrnlSize;                    /* Size of journal file on disk */
  u32 cksum = 0;                   /* Checksum of string zMaster */


  if( !zMaster || pPager->setMaster
   || pPager->journalMode==PAGER_JOURNALMODE_MEMORY 
   || pPager->journalMode==PAGER_JOURNALMODE_OFF 
  ){
    return SQLITE_OK;
  }
  pPager->setMaster = 1;
  assert( isOpen(pPager->jfd) );

  /* Calculate the length in bytes and the checksum of zMaster */
  for(nMaster=0; zMaster[nMaster]; nMaster++){
    cksum += zMaster[nMaster];
  }

  /* If in full-sync mode, advance to the next disk sector before writing
  ** the master journal name. This is in case the previous page written to
  ** the journal has already been synced.
  */
  if( pPager->fullSync ){
    pPager->journalOff = journalHdrOffset(pPager);
  }
  iHdrOff = pPager->journalOff;


  /* Write the master journal data to the end of the journal file. If
  ** an error occurs, return the error code to the caller.
  */
  if( (rc = write32bits(pPager->jfd, iHdrOff, PAGER_MJ_PGNO(pPager)))
   || (rc = sqlite3OsWrite(pPager->jfd, zMaster, nMaster, iHdrOff+4))
   || (rc = write32bits(pPager->jfd, iHdrOff+4+nMaster, nMaster))


   || (rc = write32bits(pPager->jfd, iHdrOff+4+nMaster+4, cksum))
   || (rc = sqlite3OsWrite(pPager->jfd, aJournalMagic, 8, iHdrOff+4+nMaster+8))
  ){
    return rc;

  }




  pPager->journalOff += (nMaster+20);
  pPager->needSync = !pPager->noSync;

  /* If the pager is in peristent-journal mode, then the physical 
  ** journal-file may extend past the end of the master-journal name
  ** and 8 bytes of magic data just written to the file. This is 
  ** dangerous because the code to rollback a hot-journal file
  ** will not be able to find the master-journal name to determine 
  ** whether or not the journal is hot. 
  **
  ** Easiest thing to do in this scenario is to truncate the journal 
  ** file to the required size.
  */ 

  if( SQLITE_OK==(rc = sqlite3OsFileSize(pPager->jfd, &jrnlSize))
   && jrnlSize>pPager->journalOff
  ){
    rc = sqlite3OsTruncate(pPager->jfd, pPager->journalOff);
  }
  return rc;
}

/*
** Find a page in the hash table given its page number. Return
** a pointer to the page or NULL if the requested page is not 
** already in memory.
*/
static PgHdr *pager_lookup(Pager *pPager, Pgno pgno){
  PgHdr *p;                         /* Return value */

  /* It is not possible for a call to PcacheFetch() with createFlag==0 to
  ** fail, since no attempt to allocate dynamic memory will be made.
  */
  (void)sqlite3PcacheFetch(pPager->pPCache, pgno, 0, &p);
  return p;
}

/*
** Unless the pager is in error-state, discard all in-memory pages. If
** the pager is in error-state, then this call is a no-op.


*/
static void pager_reset(Pager *pPager){
  if( SQLITE_OK==pPager->errCode ){
    sqlite3PcacheClear(pPager->pPCache);
  }
}

/*
** Free all structures in the Pager.aSavepoint[] array and set both
** Pager.aSavepoint and Pager.nSavepoint to zero. Close the sub-journal
** if it is open and the pager is not in exclusive mode.
*/
static void releaseAllSavepoints(Pager *pPager){
  int ii;               /* Iterator for looping through Pager.aSavepoint */
  for(ii=0; ii<pPager->nSavepoint; ii++){
    sqlite3BitvecDestroy(pPager->aSavepoint[ii].pInSavepoint);
  }
  if( !pPager->exclusiveMode ){
    sqlite3OsClose(pPager->sjfd);
  }
  sqlite3_free(pPager->aSavepoint);
  pPager->aSavepoint = 0;
  pPager->nSavepoint = 0;
  pPager->nSubRec = 0;
}

/*
** Set the bit number pgno in the PagerSavepoint.pInSavepoint 
** bitvecs of all open savepoints. Return SQLITE_OK if successful
** or SQLITE_NOMEM if a malloc failure occurs.
*/
static int addToSavepointBitvecs(Pager *pPager, Pgno pgno){
  int ii;                   /* Loop counter */
  int rc = SQLITE_OK;       /* Result code */

  for(ii=0; ii<pPager->nSavepoint; ii++){
    PagerSavepoint *p = &pPager->aSavepoint[ii];
    if( pgno<=p->nOrig ){
      rc |= sqlite3BitvecSet(p->pInSavepoint, pgno);
      testcase( rc==SQLITE_NOMEM );
      assert( rc==SQLITE_OK || rc==SQLITE_NOMEM );
    }
  }
  return rc;
}

/*
** Unlock the database file. This function is a no-op if the pager
** is in exclusive mode.
**
** If the pager is currently in error state, discard the contents of 
** the cache and reset the Pager structure internal state. If there is
** an open journal-file, then the next time a shared-lock is obtained
** on the pager file (by this or any other process), it will be
** treated as a hot-journal and rolled back.
*/
static void pager_unlock(Pager *pPager){
  if( !pPager->exclusiveMode ){
    int rc;                      /* Return code */

    /* Always close the journal file when dropping the database lock.
    ** Otherwise, another connection with journal_mode=delete might
    ** delete the file out from under us.
    */

    sqlite3OsClose(pPager->jfd);

    sqlite3BitvecDestroy(pPager->pInJournal);
    pPager->pInJournal = 0;

    releaseAllSavepoints(pPager);

    /* If the file is unlocked, somebody else might change it. The
    ** values stored in Pager.dbSize etc. might become invalid if
    ** this happens. TODO: Really, this doesn't need to be cleared
    ** until the change-counter check fails in pagerSharedLock().
    */
    pPager->dbSizeValid = 0;

    rc = osUnlock(pPager->fd, NO_LOCK);
    if( rc ){
      pPager->errCode = rc;
    }
    IOTRACE(("UNLOCK %p\n", pPager))

    /* If Pager.errCode is set, the contents of the pager cache cannot be
    ** trusted. Now that the pager file is unlocked, the contents of the
    ** cache can be discarded and the error code safely cleared.
    */
    if( pPager->errCode ){
      if( rc==SQLITE_OK ){
        pPager->errCode = SQLITE_OK;
      }
      pager_reset(pPager);
    }

    pPager->changeCountDone = 0;

    pPager->state = PAGER_UNLOCK;
  }
}

/*
** This function should be called when an IOERR, CORRUPT or FULL error
** may have occured. The first argument is a pointer to the pager 
** structure, the second the error-code about to be returned by a pager 
** API function. The value returned is a copy of the second argument 
** to this function. 
**
** If the second argument is SQLITE_IOERR, SQLITE_CORRUPT, or SQLITE_FULL
** the error becomes persistent. Until the persisten error is cleared,
** subsequent API calls on this Pager will immediately return the same 
** error code.
**
** A persistent error indicates that the contents of the pager-cache 
** cannot be trusted. This state can be cleared by completely discarding 
** the contents of the pager-cache. If a transaction was active when
** the persistent error occured, then the rollback journal may need
** to be replayed to restore the contents of the database file (as if
** it were a hot-journal).
*/
static int pager_error(Pager *pPager, int rc){
  int rc2 = rc & 0xff;
  assert(
       pPager->errCode==SQLITE_FULL ||
       pPager->errCode==SQLITE_OK ||
       (pPager->errCode & 0xff)==SQLITE_IOERR
  );
  if(
    rc2==SQLITE_FULL ||
    rc2==SQLITE_IOERR ||
    rc2==SQLITE_CORRUPT
  ){
    pPager->errCode = rc;
    if( pPager->state==PAGER_UNLOCK 
     && sqlite3PcacheRefCount(pPager->pPCache)==0 
    ){
      /* If the pager is already unlocked, call pager_unlock() now to
      ** clear the error state and ensure that the pager-cache is 
      ** completely empty.
      */
      pager_unlock(pPager);
    }
  }
  return 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->errCode==SQLITE_OK && pPager->state>=PAGER_RESERVED ){
    sqlite3BeginBenignMalloc();
    sqlite3PagerRollback(pPager);
    sqlite3EndBenignMalloc();
  }
  pager_unlock(pPager);
}

/*
** 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

** depends on whether or not the pager is running in exclusive mode and

** the current journal-mode (Pager.journalMode value), as follows:
**
**   journalMode==MEMORY
**     Journal file descriptor is simply closed. This destroys an 
**     in-memory journal.
**
**   journalMode==TRUNCATE
**     Journal file is truncated to zero bytes in size.
**
**   journalMode==PERSIST
**     The first 28 bytes of the journal file are zeroed. This invalidates
**     the first journal header in the file, and hence the entire journal
**     file. An invalid journal file cannot be rolled back.
**
**   journalMode==DELETE
**     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 */

  if( pPager->state<PAGER_RESERVED ){
    return SQLITE_OK;
  }
  releaseAllSavepoints(pPager);

  assert( isOpen(pPager->jfd) || pPager->pInJournal==0 );
  if( isOpen(pPager->jfd) ){

    /* TODO: There's a problem here if a journal-file was opened in MEMORY
    ** mode and then the journal-mode is changed to TRUNCATE or PERSIST
    ** during the transaction. This code should be changed to assume
    ** that the journal mode has not changed since the transaction was
    ** started. And the sqlite3PagerJournalMode() function should be
    ** changed to make sure that this is the case too.
    */

    /* Finalize the journal file. */
    if( pPager->journalMode==PAGER_JOURNALMODE_MEMORY ){
      int isMemoryJournal = sqlite3IsMemJournal(pPager->jfd);
      sqlite3OsClose(pPager->jfd);

      if( !isMemoryJournal ){
        rc = sqlite3OsDelete(pPager->pVfs, pPager->zJournal, 0);
      }
    }else if( pPager->journalMode==PAGER_JOURNALMODE_TRUNCATE
         && (rc = sqlite3OsTruncate(pPager->jfd, 0))==SQLITE_OK ){
      pPager->journalOff = 0;
      pPager->journalStarted = 0;
    }else if( pPager->exclusiveMode 
     || pPager->journalMode==PAGER_JOURNALMODE_PERSIST
    ){
      rc = zeroJournalHdr(pPager, hasMaster);
      pager_error(pPager, rc);
      pPager->journalOff = 0;
      pPager->journalStarted = 0;
    }else{
      assert( pPager->journalMode==PAGER_JOURNALMODE_DELETE || rc );
      sqlite3OsClose(pPager->jfd);

      if( rc==SQLITE_OK && !pPager->tempFile ){
        rc = sqlite3OsDelete(pPager->pVfs, pPager->zJournal, 0);
      }
    }




#ifdef SQLITE_CHECK_PAGES
    sqlite3PcacheIterateDirty(pPager->pPCache, pager_set_pagehash);
#endif

    sqlite3PcacheCleanAll(pPager->pPCache);
    sqlite3BitvecDestroy(pPager->pInJournal);
    pPager->pInJournal = 0;
    pPager->nRec = 0;


  }

  if( !pPager->exclusiveMode ){
    rc2 = osUnlock(pPager->fd, SHARED_LOCK);
    pPager->state = PAGER_SHARED;
    pPager->changeCountDone = 0;
  }else if( pPager->state==PAGER_SYNCED ){
    pPager->state = PAGER_EXCLUSIVE;
  }

  pPager->setMaster = 0;
  pPager->needSync = 0;
  pPager->dbModified = 0;

  /* TODO: Is this optimal? Why is the db size invalidated here 
  ** when the database file is not unlocked? */
  pPager->dbOrigSize = 0;
  sqlite3PcacheTruncate(pPager->pPCache, pPager->dbSize);
  if( !MEMDB ){
    pPager->dbSizeValid = 0;
  }


  return (rc==SQLITE_OK?rc2:rc);
}

/*
** Parameter aData must point to a buffer of pPager->pageSize bytes
** of data. Compute and return a checksum based ont the contents of the 
** page of data and the current value of pPager->cksumInit.
**
** This is not a real checksum. It is really just the sum of the 
** random initial value (pPager->cksumInit) and every 200th byte
** of the page data, starting with byte offset (pPager->pageSize%200).
** Each byte is interpreted as an 8-bit unsigned integer.
**

** Changing the formula used to compute this checksum results in an
** incompatible journal file format.
**
** If journal corruption occurs due to a power failure, the most likely 
** scenario is that one end or the other of the record will be changed. 
** It is much less likely that the two ends of the journal record will be
** correct and the middle be corrupt.  Thus, this "checksum" scheme,
** though fast and simple, catches the mostly likely kind of corruption.





*/
static u32 pager_cksum(Pager *pPager, const u8 *aData){
  u32 cksum = pPager->cksumInit;         /* Checksum value to return */
  int i = pPager->pageSize-200;          /* Loop counter */
  while( i>0 ){
    cksum += aData[i];
    i -= 200;
  }
  return cksum;
}

/*
** 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
** (if the corresponding pDone bit is set) then skip the playback.
** Make sure the pDone bit corresponding to the *pOffset page is set
** prior to returning.
**
** If the page record is successfully read from the (sub-)journal file
** and played back, then SQLITE_OK is returned. If an IO error occurs
** while reading the record from the (sub-)journal file or while writing
** to the database file, then the IO error code is returned. If data
** is successfully read from the (sub-)journal file but appears to be
** corrupted, SQLITE_DONE is returned. Data is considered corrupted in
** two circumstances:
** 
**   * If the record page-number is illegal (0 or PAGER_MJ_PGNO), or
**   * If the record is being rolled back from the main journal file
**     and the checksum field does not match the record content.
**
** Neither of these two scenarios are possible during a savepoint rollback.
**
** If this is a savepoint rollback, then memory may have to be dynamically
** allocated by this function. If this is the case and an allocation fails,
** SQLITE_NOMEM is returned.
*/
static int pager_playback_one_page(
  Pager *pPager,                /* The pager being played back */
  int isMainJrnl,               /* 1 -> main journal. 0 -> sub-journal. */
  i64 *pOffset,                 /* Offset of record to playback */
  int isSavepnt,                /* True for a savepoint rollback */
  Bitvec *pDone                 /* Bitvec of pages already 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 = (u8*)pPager->pTmpSpace;
  assert( aData );         /* Temp storage must have already been allocated */

  /* 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;
  rc = sqlite3OsRead(jfd, aData, pPager->pageSize, (*pOffset)+4);
  if( rc!=SQLITE_OK ) return rc;
  *pOffset += pPager->pageSize + 4 + isMainJrnl*4;

  /* Sanity checking on the page.  This is more important that I originally
  ** thought.  If a power failure occurs while the journal is being written,
  ** it could cause invalid data to be written into the journal.  We need to
  ** detect this invalid data (with high probability) and ignore it.
  */
  if( pgno==0 || pgno==PAGER_MJ_PGNO(pPager) ){
    assert( !isSavepnt );
    return SQLITE_DONE;
  }
  if( pgno>(Pgno)pPager->dbSize || sqlite3BitvecTest(pDone, pgno) ){
    return SQLITE_OK;
  }
  if( isMainJrnl ){
    rc = read32bits(jfd, (*pOffset)-4, &cksum);
    if( rc ) return rc;
    if( !isSavepnt && pager_cksum(pPager, aData)!=cksum ){
      return SQLITE_DONE;
    }
  }

  if( pDone && (rc = sqlite3BitvecSet(pDone, pgno))!=SQLITE_OK ){
    return rc;
  }

  assert( pPager->state==PAGER_RESERVED || pPager->state>=PAGER_EXCLUSIVE );

  /* If the pager is in RESERVED state, then there must be a copy of this
  ** page in the pager cache. In this case just update the pager cache,

  pPg = pager_lookup(pPager, pgno);
  PAGERTRACE(("PLAYBACK %d page %d hash(%08x) %s\n",
               PAGERID(pPager), pgno, pager_datahash(pPager->pageSize, aData),
               (isMainJrnl?"main-journal":"sub-journal")
  ));
  if( (pPager->state>=PAGER_EXCLUSIVE)
   && (pPg==0 || 0==(pPg->flags&PGHDR_NEED_SYNC))
   && isOpen(pPager->fd)
  ){
    i64 ofst = (pgno-1)*(i64)pPager->pageSize;
    rc = sqlite3OsWrite(pPager->fd, aData, pPager->pageSize, ofst);
    if( pgno>pPager->dbFileSize ){
      pPager->dbFileSize = pgno;
    }
  }else if( !isMainJrnl && pPg==0 ){

    **
    ** The solution is to add an in-memory page to the cache containing
    ** the data just read from the sub-journal. Mark the page as dirty 
    ** and if the pager requires a journal-sync, then mark the page as 
    ** requiring a journal-sync before it is written.
    */
    assert( isSavepnt );
    if( (rc = sqlite3PagerAcquire(pPager, pgno, &pPg, 1))!=SQLITE_OK ){
      return rc;
    }
    pPg->flags &= ~PGHDR_NEED_READ;
    sqlite3PcacheMakeDirty(pPg);
  }
  if( pPg ){
    /* No page should ever be explicitly rolled back that is in use, except
    ** for page 1 which is held in use in order to keep the lock on the
    ** database active. However such a page may be rolled back as a result
    ** of an internal error resulting in an automatic call to
    ** sqlite3PagerRollback().
    */
    void *pData;
    pData = pPg->pData;
    memcpy(pData, aData, pPager->pageSize);
    if( pPager->xReiniter ){
      pPager->xReiniter(pPg);
    }
    if( isMainJrnl && (!isSavepnt || *pOffset<=pPager->journalHdr) ){
      /* If the contents of this page were just restored from the main 
      ** journal file, then its content must be as they were when the 
      ** transaction was first opened. In this case we can mark the page
      ** as clean, since there will be no need to write it out to the.
      **
      ** There is one exception to this rule. If the page is being rolled
      ** back as part of a savepoint (or statement) rollback from an 

** file that referred to the master journal file has just been rolled back.
** This routine checks if it is possible to delete the master journal file,
** and does so if it is.
**
** Argument zMaster may point to Pager.pTmpSpace. So that buffer is not 
** available for use within this function.
**
** When a master journal file is created, it is populated with the names 
** of all of its child journals, one after another, formatted as utf-8 
** encoded text. The end of each child journal file is marked with a 
** nul-terminator byte (0x00). i.e. the entire contents of a master journal
** file for a transaction involving two databases might be:
**
**   "/home/bill/a.db-journal\x00/home/bill/b.db-journal\x00"
**
** A master journal file may only be deleted once all of its child 
** journals have been rolled back.
**
** This function reads the contents of the master-journal file into 
** memory and loops through each of the child journal names. For
** each child journal, it checks if:
**
**   * if the child journal exists, and if so
**   * if the child journal contains a reference to master journal 
**     file zMaster
**
** If a child journal can be found that matches both of the criteria

** above, this function returns without doing anything. Otherwise, if
** no such child journal can be found, file zMaster is deleted from
** the file-system using sqlite3OsDelete().
**
** If an IO error within this function, an error code is returned. This
** function allocates memory by calling sqlite3Malloc(). If an allocation
** fails, SQLITE_NOMEM is returned. Otherwise, if no IO or malloc errors 
** occur, SQLITE_OK is returned.
**
** TODO: This function allocates a single block of memory to load
** the entire contents of the master journal file. This could be
** a couple of kilobytes or so - potentially larger than the page 
** size.
*/
static int pager_delmaster(Pager *pPager, const char *zMaster){
  sqlite3_vfs *pVfs = pPager->pVfs;
  int rc;                   /* Return code */

  sqlite3_file *pMaster;    /* Malloc'd master-journal file descriptor */
  sqlite3_file *pJournal;   /* Malloc'd child-journal file descriptor */
  char *zMasterJournal = 0; /* Contents of master journal file */
  i64 nMasterJournal;       /* Size of master journal file */

  /* Allocate space for both the pJournal and pMaster file descriptors.
  ** If successful, open the master journal file for reading.

  */
  pMaster = (sqlite3_file *)sqlite3MallocZero(pVfs->szOsFile * 2);
  pJournal = (sqlite3_file *)(((u8 *)pMaster) + pVfs->szOsFile);
  if( !pMaster ){
    rc = SQLITE_NOMEM;
  }else{
    const int flags = (SQLITE_OPEN_READONLY|SQLITE_OPEN_MASTER_JOURNAL);
    rc = sqlite3OsOpen(pVfs, zMaster, pMaster, flags, 0);
  }
  if( rc!=SQLITE_OK ) goto delmaster_out;


  rc = sqlite3OsFileSize(pMaster, &nMasterJournal);
  if( rc!=SQLITE_OK ) goto delmaster_out;

  if( nMasterJournal>0 ){
    char *zJournal;
    char *zMasterPtr = 0;
    int nMasterPtr = pVfs->mxPathname+1;

    /* Load the entire master journal file into space obtained from
    ** sqlite3_malloc() and pointed to by zMasterJournal. 
    */
    zMasterJournal = (char *)sqlite3Malloc((int)nMasterJournal + nMasterPtr);
    if( !zMasterJournal ){
      rc = SQLITE_NOMEM;

  
  rc = sqlite3OsDelete(pVfs, zMaster, 0);

delmaster_out:
  if( zMasterJournal ){
    sqlite3_free(zMasterJournal);
  }  
  if( pMaster ){
    sqlite3OsClose(pMaster);
    assert( !isOpen(pJournal) );
  }
  sqlite3_free(pMaster);
  return rc;
}


/*
** This function is used to change the actual size of the database 
** file in the file-system. This only happens when committing a transaction,
** or rolling back a transaction (including rolling back a hot-journal).
**
** If the main database file is not open, or an exclusive lock is not
** held, this function is a no-op. Otherwise, the size of the file is
** changed to nPage pages (nPage*pPager->pageSize bytes). If the file
** on disk is currently larger than nPage pages, then use the VFS
** xTruncate() method to truncate it.

**
** Or, it might might be the case that the file on disk is smaller than 



** nPage pages. Some operating system implementations can get confused if 
** you try to truncate a file to some size that is larger than it 
** currently is, so detect this case and write a single zero byte to 
** the end of the new file instead.
**
** If successful, return SQLITE_OK. If an IO error occurs while modifying
** the database file, return the error code to the caller.
*/
static int pager_truncate(Pager *pPager, Pgno nPage){
  int rc = SQLITE_OK;
  if( pPager->state>=PAGER_EXCLUSIVE && isOpen(pPager->fd) ){
    i64 currentSize, newSize;
    /* TODO: Is it safe to use Pager.dbFileSize here? */
    rc = sqlite3OsFileSize(pPager->fd, &currentSize);
    newSize = pPager->pageSize*(i64)nPage;
    if( rc==SQLITE_OK && currentSize!=newSize ){
      if( currentSize>newSize ){
        rc = sqlite3OsTruncate(pPager->fd, newSize);
      }else{
        rc = sqlite3OsWrite(pPager->fd, "", 1, newSize-1);
      }
      if( rc==SQLITE_OK ){
        pPager->dbFileSize = nPage;
      }
    }
  }
  return rc;
}

/*
** Set the value of the Pager.sectorSize variable for the given
** pager based on the value returned by the xSectorSize method
** of the open database file. The sector size will be used used 
** to determine the size and alignment of journal header and 
** master journal pointers within created journal files.
**
** For temporary files the effective sector size is always 512 bytes.
**
** Otherwise, for non-temporary files, the effective sector size is
** the value returned by the xSectorSize() method rounded up to 512 if
** it is less than 512, or rounded down to MAX_SECTOR_SIZE if it
** is greater than MAX_SECTOR_SIZE.
*/
static void setSectorSize(Pager *pPager){
  assert( isOpen(pPager->fd) || pPager->tempFile );

  if( !pPager->tempFile ){
    /* Sector size doesn't matter for temporary files. Also, the file
    ** may not have been opened yet, in which case the OsSectorSize()
    ** call will segfault.
    */
    pPager->sectorSize = sqlite3OsSectorSize(pPager->fd);
  }
  if( pPager->sectorSize<512 ){
    pPager->sectorSize = 512;
  }
  if( pPager->sectorSize>MAX_SECTOR_SIZE ){
    assert( MAX_SECTOR_SIZE>=512 );
    pPager->sectorSize = MAX_SECTOR_SIZE;
  }
}

/*
** Playback the journal and thus restore the database file to
** the state it was in before we started making changes.  

  int rc;                  /* Result code of a subroutine */
  int res = 1;             /* Value returned by sqlite3OsAccess() */
  char *zMaster = 0;       /* Name of master journal file if any */

  /* Figure out how many records are in the journal.  Abort early if
  ** the journal is empty.
  */
  assert( isOpen(pPager->jfd) );
  rc = sqlite3OsFileSize(pPager->jfd, &szJ);
  if( rc!=SQLITE_OK || szJ==0 ){
    goto end_playback;
  }

  /* Read the master journal name from the journal, if it is present.
  ** If a master journal file name is specified, but the file is not
  ** present on disk, then the journal is not hot and does not need to be
  ** played back.
  **
  ** TODO: Technically the following is an error because it assumes that
  ** buffer Pager.pTmpSpace is (mxPathname+1) bytes or larger. i.e. that
  ** (pPager->pageSize >= pPager->pVfs->mxPathname+1). Using os_unix.c,
  **  mxPathname is 512, which is the same as the minimum allowable value
  ** for pageSize.
  */
  zMaster = pPager->pTmpSpace;
  rc = readMasterJournal(pPager->jfd, zMaster, pPager->pVfs->mxPathname+1);
  if( rc==SQLITE_OK && zMaster[0] ){
    rc = sqlite3OsAccess(pVfs, zMaster, SQLITE_ACCESS_EXISTS, &res);
  }
  zMaster = 0;
  if( rc!=SQLITE_OK || !res ){
    goto end_playback;
  }
  pPager->journalOff = 0;

  /* This loop terminates either when a readJournalHdr() or 
  ** pager_playback_one_page() call returns SQLITE_DONE or an IO error 
  ** occurs. 
  */
  while( 1 ){

    /* Read the next journal header from the journal file.  If there are
    ** not enough bytes left in the journal file for a complete header, or
    ** it is corrupted, then a process must of failed while writing it.
    ** This indicates nothing more needs to be rolled back.
    */

      rc = pager_truncate(pPager, mxPg);
      if( rc!=SQLITE_OK ){
        goto end_playback;
      }
      pPager->dbSize = mxPg;
    }

    /* Copy original pages out of the journal and back into the 
    ** database file and/or page cache.
    */
    for(u=0; u<nRec; u++){
      rc = pager_playback_one_page(pPager, 1, &pPager->journalOff, 0, 0);
      if( rc!=SQLITE_OK ){
        if( rc==SQLITE_DONE ){
          rc = SQLITE_OK;
          pPager->journalOff = szJ;

  ** assertion that the transaction counter was modified.
  */
  assert(
    pPager->fd->pMethods==0 ||
    sqlite3OsFileControl(pPager->fd,SQLITE_FCNTL_DB_UNCHANGED,0)>=SQLITE_OK
  );

  /* If this playback is happening automatically as a result of an IO or 
  ** malloc error that occured after the change-counter was updated but 
  ** before the transaction was committed, then the change-counter 
  ** modification may just have been reverted. If this happens in exclusive 
  ** mode, then subsequent transactions performed by the connection will not
  ** update the change-counter at all. This may lead to cache inconsistency
  ** problems for other processes at some point in the future. So, just
  ** in case this has happened, clear the changeCountDone flag now.
  */
  pPager->changeCountDone = pPager->tempFile;

  if( rc==SQLITE_OK ){
    zMaster = pPager->pTmpSpace;
    rc = readMasterJournal(pPager->jfd, zMaster, pPager->pVfs->mxPathname+1);
    testcase( rc!=SQLITE_OK );
  }
  if( rc==SQLITE_OK ){
    rc = pager_end_transaction(pPager, zMaster[0]!='\0');
    testcase( rc!=SQLITE_OK );
  }
  if( rc==SQLITE_OK && zMaster[0] && res ){
    /* If there was a master journal and this routine will return success,
    ** see if it is possible to delete the master journal.
    */
    rc = pager_delmaster(pPager, zMaster);
    testcase( rc!=SQLITE_OK );
  }

  /* The Pager.sectorSize variable may have been updated while rolling
  ** back a journal created by a process with a different sector size
  ** value. Reset it to the correct value for this process.
  */
  setSectorSize(pPager);
  return rc;
}

/*
** Playback savepoint pSavepoint. Or, if pSavepoint==NULL, then playback
** the entire master journal file. The case pSavepoint==NULL occurs when 
** a ROLLBACK TO command is invoked on a SAVEPOINT that is a transaction 
** savepoint.
**
** When pSavepoint is not NULL (meaning a non-transaction savepoint is 
** being rolled back), then the rollback consists of up to three stages,
** performed in the order specified:
**
**   * Pages are played back from the main journal starting at byte
**     offset PagerSavepoint.iOffset and continuing to 
**     PagerSavepoint.iHdrOffset, or to the end of the main journal
**     file if PagerSavepoint.iHdrOffset is zero.
**
**   * If PagerSavepoint.iHdrOffset is not zero, then pages are played
**     back starting from the journal header immediately following 
**     PagerSavepoint.iHdrOffset to the end of the main journal file.
**
**   * Pages are then played back from the sub-journal file, starting
**     with the PagerSavepoint.iSubRec and continuing to the end of
**     the journal file.
**
** Throughout the rollback process, each time a page is rolled back, the
** corresponding bit is set in a bitvec structure (variable pDone in the
** implementation below). This is used to ensure that a page is only
** rolled back the first time it is encountered in either journal.
**
** If pSavepoint is NULL, then pages are only played back from the main
** journal file. There is no need for a bitvec in this case.
**
** In either case, before playback commences the Pager.dbSize variable
** is reset to the value that it held at the start of the savepoint 
** (or transaction). No page with a page-number greater than this value
** is played back. If one is encountered it is simply skipped.
*/
static int pagerPlaybackSavepoint(Pager *pPager, PagerSavepoint *pSavepoint){
  i64 szJ;                 /* Effective size of the main journal */
  i64 iHdrOff;             /* End of first segment of main-journal records */

  int rc = SQLITE_OK;      /* Return code */
  Bitvec *pDone = 0;       /* Bitvec to ensure pages played back only once */

  assert( pPager->state>=PAGER_SHARED );

  /* Allocate a bitvec to use to store the set of pages rolled back */
  if( pSavepoint ){
    pDone = sqlite3BitvecCreate(pSavepoint->nOrig);
    if( !pDone ){
      return SQLITE_NOMEM;
    }
  }

  /* Set the database size back to the value it was before the savepoint 
  ** being reverted was opened.
  */
  pPager->dbSize = pSavepoint ? pSavepoint->nOrig : pPager->dbOrigSize;


  /* Use pPager->journalOff as the effective size of the main rollback
  ** journal.  The actual file might be larger than this in
  ** PAGER_JOURNALMODE_TRUNCATE or PAGER_JOURNALMODE_PERSIST.  But anything
  ** past pPager->journalOff is off-limits to us.
  */
  szJ = pPager->journalOff;

  /* Begin by rolling back records from the main journal starting at
  ** PagerSavepoint.iOffset and continuing to the next journal header.
  ** There might be records in the main journal that have a page number
  ** greater than the current database size (pPager->dbSize) but those
  ** will be skipped automatically.  Pages are added to pDone as they
  ** are played back.
  */
  if( pSavepoint ){
    iHdrOff = pSavepoint->iHdrOffset ? pSavepoint->iHdrOffset : szJ;
    pPager->journalOff = pSavepoint->iOffset;
    while( rc==SQLITE_OK && pPager->journalOff<iHdrOff ){
      rc = pager_playback_one_page(pPager, 1, &pPager->journalOff, 1, pDone);

    }
    assert( rc!=SQLITE_DONE );
  }else{
    pPager->journalOff = 0;
  }

  /* Continue rolling back records out of the main journal starting at
  ** the first journal header seen and continuing until the effective end
  ** of the main journal file.  Continue to skip out-of-range pages and
  ** continue adding pages rolled back to pDone.
  */
  while( rc==SQLITE_OK && pPager->journalOff<szJ ){
    u32 ii;            /* Loop counter */
    u32 nJRec = 0;     /* Number of Journal Records */
    u32 dummy;
    rc = readJournalHdr(pPager, szJ, &nJRec, &dummy);
    assert( rc!=SQLITE_DONE );

    /*
    ** The "pPager->journalHdr+JOURNAL_HDR_SZ(pPager)==pPager->journalOff"
    ** test is related to ticket #2565.  See the discussion in the
    ** pager_playback() function for additional information.
    */
    assert( !(nJRec==0
         && pPager->journalHdr+JOURNAL_HDR_SZ(pPager)!=pPager->journalOff
         && ((szJ - pPager->journalOff) / JOURNAL_PG_SZ(pPager))>0
         && pagerNextJournalPageIsValid(pPager))
    );
    if( nJRec==0 
     && pPager->journalHdr+JOURNAL_HDR_SZ(pPager)==pPager->journalOff
    ){
      nJRec = (u32)((szJ - pPager->journalOff)/JOURNAL_PG_SZ(pPager));
    }
    for(ii=0; rc==SQLITE_OK && ii<nJRec && pPager->journalOff<szJ; ii++){
      rc = pager_playback_one_page(pPager, 1, &pPager->journalOff, 1, pDone);

    }
    assert( rc!=SQLITE_DONE );
  }
  assert( rc!=SQLITE_OK || pPager->journalOff==szJ );

  /* Finally,  rollback pages from the sub-journal.  Page that were
  ** previously rolled back out of the main journal (and are hence in pDone)
  ** will be skipped.  Out-of-range pages are also skipped.
  */
  if( pSavepoint ){
    u32 ii;            /* Loop counter */
    i64 offset = pSavepoint->iSubRec*(4+pPager->pageSize);
    for(ii=pSavepoint->iSubRec; rc==SQLITE_OK && ii<pPager->nSubRec; ii++){
      assert( offset==ii*(4+pPager->pageSize) );
      rc = pager_playback_one_page(pPager, 0, &offset, 1, pDone);

    }
    assert( rc!=SQLITE_DONE );
  }

  sqlite3BitvecDestroy(pDone);
  if( rc==SQLITE_OK ){
    pPager->journalOff = szJ;
  }
  return rc;

** testing and analysis only.  
*/
#ifdef SQLITE_TEST
SQLITE_API int sqlite3_opentemp_count = 0;
#endif

/*
** Open a temporary file.
**
** Write the file descriptor into *pFile. Return SQLITE_OK on success 
** or some other error code if we fail. The OS will automatically 
** delete the temporary file when it is closed.
**
** The flags passed to the VFS layer xOpen() call are those specified
** by parameter vfsFlags ORed with the following:
**
**     SQLITE_OPEN_READWRITE
**     SQLITE_OPEN_CREATE
**     SQLITE_OPEN_EXCLUSIVE
**     SQLITE_OPEN_DELETEONCLOSE
*/
static int pagerOpentemp(
  Pager *pPager,        /* The pager object */
  sqlite3_file *pFile,  /* Write the file descriptor here */
  int vfsFlags          /* Flags passed through to the VFS */
){
  int rc;               /* Return code */

#ifdef SQLITE_TEST
  sqlite3_opentemp_count++;  /* Used for testing and analysis only */
#endif

  vfsFlags |=  SQLITE_OPEN_READWRITE | SQLITE_OPEN_CREATE |
            SQLITE_OPEN_EXCLUSIVE | SQLITE_OPEN_DELETEONCLOSE;
  rc = sqlite3OsOpen(pPager->pVfs, 0, pFile, vfsFlags, 0);
  assert( rc!=SQLITE_OK || isOpen(pFile) );
  return rc;











































































































































































































































}

/*
** Set the busy handler function.
**
** The pager invokes the busy-handler if sqlite3OsLock() returns 
** SQLITE_BUSY when trying to upgrade from no-lock to a SHARED lock,
** or when trying to upgrade from a RESERVED lock to an EXCLUSIVE 
** lock. It does *not* invoke the busy handler when upgrading from
** SHARED to RESERVED, or when upgrading from SHARED to EXCLUSIVE
** (which occurs during hot-journal rollback). Summary:
**
**   Transition                        | Invokes xBusyHandler
**   --------------------------------------------------------
**   NO_LOCK       -> SHARED_LOCK      | Yes
**   SHARED_LOCK   -> RESERVED_LOCK    | No
**   SHARED_LOCK   -> EXCLUSIVE_LOCK   | No
**   RESERVED_LOCK -> EXCLUSIVE_LOCK   | Yes
**
** If the busy-handler callback returns non-zero, the lock is 
** retried. If it returns zero, then the SQLITE_BUSY error is
** returned to the caller of the pager API function.
*/
SQLITE_PRIVATE void sqlite3PagerSetBusyhandler(
  Pager *pPager,                       /* Pager object */
  int (*xBusyHandler)(void *),         /* Pointer to busy-handler function */
  void *pBusyHandlerArg                /* Argument to pass to xBusyHandler */
){  
  pPager->xBusyHandler = xBusyHandler;
  pPager->pBusyHandlerArg = pBusyHandlerArg;
}

/*
** Set the reinitializer for this pager. If not NULL, the reinitializer
** is called when the content of a page in cache is modified (restored)
** as part of a transaction or savepoint rollback. The callback gives 
** higher-level code an opportunity to restore the EXTRA section to 
** agree with the restored page data.
*/
SQLITE_PRIVATE void sqlite3PagerSetReiniter(Pager *pPager, void (*xReinit)(DbPage*)){
  pPager->xReiniter = xReinit;
}

/*
** Change the page size used by the Pager object. The new page size 
** is passed in *pPageSize.
**
** If the pager is in the error state when this function is called, it
** is a no-op. The value returned is the error state error code (i.e. 
** one of SQLITE_IOERR, SQLITE_CORRUPT or SQLITE_FULL).
**
** Otherwise, if all of the following are true:
**
**   * the new page size (value of *pPageSize) is valid (a power 
**     of two between 512 and SQLITE_MAX_PAGE_SIZE, inclusive), and
**
**   * there are no outstanding page references, and
**
**   * the database is either not an in-memory database or it is
**     an in-memory database that currently consists of zero pages.
**
** then the pager object page size is set to *pPageSize.
**
** If the page size is changed, then this function uses sqlite3PagerMalloc() 
** to obtain a new Pager.pTmpSpace buffer. If this allocation attempt 
** fails, SQLITE_NOMEM is returned and the page size remains unchanged. 
** In all other cases, SQLITE_OK is returned.
**
** If the page size is not changed, either because one of the enumerated
** conditions above is not true, the pager was in error state when this
** function was called, or because the memory allocation attempt failed, 
** then *pPageSize is set to the old, retained page size before returning.
*/
SQLITE_PRIVATE int sqlite3PagerSetPagesize(Pager *pPager, u16 *pPageSize){
  int rc = pPager->errCode;
  if( rc==SQLITE_OK ){
    u16 pageSize = *pPageSize;
    assert( pageSize==0 || (pageSize>=512 && pageSize<=SQLITE_MAX_PAGE_SIZE) );
    if( pageSize && pageSize!=pPager->pageSize 
     && (pPager->memDb==0 || pPager->dbSize==0)
     && sqlite3PcacheRefCount(pPager->pPCache)==0 
    ){
      char *pNew = (char *)sqlite3PageMalloc(pageSize);
      if( !pNew ){
        rc = SQLITE_NOMEM;
      }else{
        pager_reset(pPager);
        pPager->pageSize = pageSize;

        sqlite3PageFree(pPager->pTmpSpace);
        pPager->pTmpSpace = pNew;
        sqlite3PcacheSetPageSize(pPager->pPCache, pageSize);
      }
    }
    *pPageSize = (u16)pPager->pageSize;
  }

# define enable_simulated_io_errors()
#endif

/*
** Read the first N bytes from the beginning of the file into memory
** that pDest points to. 
**
** If the pager was opened on a transient file (zFilename==""), or
** opened on a file less than N bytes in size, the output buffer is
** zeroed and SQLITE_OK returned. The rationale for this is that this 


** function is used to read database headers, and a new transient or
** zero sized database has a header than consists entirely of zeroes.
**
** If any IO error apart from SQLITE_IOERR_SHORT_READ is encountered,
** the error code is returned to the caller and the contents of the
** output buffer undefined.
*/
SQLITE_PRIVATE int sqlite3PagerReadFileheader(Pager *pPager, int N, unsigned char *pDest){
  int rc = SQLITE_OK;
  memset(pDest, 0, N);
  assert( isOpen(pPager->fd) || pPager->tempFile );
  if( isOpen(pPager->fd) ){
    IOTRACE(("DBHDR %p 0 %d\n", pPager, N))
    rc = sqlite3OsRead(pPager->fd, pDest, N, 0);
    if( rc==SQLITE_IOERR_SHORT_READ ){
      rc = SQLITE_OK;
    }
  }
  return rc;
}

/*
** Return the total number of pages in the database file associated 
** with pPager. Normally, this is calculated as (<db file size>/<page-size>).
** However, if the file is between 1 and <page-size> bytes in size, then 
** this is considered a 1 page file.
**
** If the pager is in error state when this function is called, then the
** error state error code is returned and *pnPage left unchanged. Or,
** if the file system has to be queried for the size of the file and

** the query attempt returns an IO error, the IO error code is returned
** and *pnPage is left unchanged.
**
** Otherwise, if everything is successful, then SQLITE_OK is returned
** and *pnPage is set to the number of pages in the database.
*/
SQLITE_PRIVATE int sqlite3PagerPagecount(Pager *pPager, int *pnPage){
  Pgno nPage;               /* Value to return via *pnPage */



  /* If the pager is already in the error state, return the error code. */
  if( pPager->errCode ){
    return pPager->errCode;

  }

  /* Determine the number of pages in the file. Store this in nPage. */
  if( pPager->dbSizeValid ){
    nPage = pPager->dbSize;
  }else{
    int rc;                 /* Error returned by OsFileSize() */
    i64 n = 0;              /* File size in bytes returned by OsFileSize() */

    assert( isOpen(pPager->fd) || pPager->tempFile );
    if( isOpen(pPager->fd) && (rc = sqlite3OsFileSize(pPager->fd, &n)) ){
      pager_error(pPager, rc);
      return rc;
    }
    if( n>0 && n<pPager->pageSize ){
      nPage = 1;
    }else{
      nPage = (Pgno)(n / pPager->pageSize);
    }
    if( pPager->state!=PAGER_UNLOCK ){
      pPager->dbSize = nPage;
      pPager->dbFileSize = nPage;
      pPager->dbSizeValid = 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;
  }

  /* Set the output variable and return SQLITE_OK */
  if( pnPage ){
    *pnPage = nPage;
  }
  return SQLITE_OK;
}






/*
** Try to obtain a lock of type locktype on the database file. If
** a similar or greater lock is already held, this function is a no-op
** (returning SQLITE_OK immediately).
**
** Otherwise, attempt to obtain the lock using sqlite3OsLock(). Invoke 
** the busy callback if the lock is currently not available. Repeat 
** until the busy callback returns false or until the attempt to 
** obtain the lock succeeds.
**
** Return SQLITE_OK on success and an error code if we cannot obtain
** the lock. If the lock is obtained successfully, set the Pager.state 
** variable to locktype before returning.
*/
static int pager_wait_on_lock(Pager *pPager, int locktype){
  int rc;                              /* Return code */

  /* The OS lock values must be the same as the Pager lock values */
  assert( PAGER_SHARED==SHARED_LOCK );
  assert( PAGER_RESERVED==RESERVED_LOCK );
  assert( PAGER_EXCLUSIVE==EXCLUSIVE_LOCK );

  /* If the file is currently unlocked then the size must be unknown */
  assert( pPager->state>=PAGER_SHARED || pPager->dbSizeValid==0 );

  /* Check that this is either a no-op (because the requested lock is 
  ** already held, or one of the transistions that the busy-handler
  ** may be invoked during, according to the comment above
  ** sqlite3PagerSetBusyhandler().
  */
  assert( (pPager->state>=locktype)
       || (pPager->state==PAGER_UNLOCK && locktype==PAGER_SHARED)
       || (pPager->state==PAGER_RESERVED && locktype==PAGER_EXCLUSIVE)
  );

  if( pPager->state>=locktype ){
    rc = SQLITE_OK;
  }else{
    do {
      rc = sqlite3OsLock(pPager->fd, locktype);
    }while( rc==SQLITE_BUSY && pPager->xBusyHandler(pPager->pBusyHandlerArg) );
    if( rc==SQLITE_OK ){
      pPager->state = (u8)locktype;
      IOTRACE(("LOCK %p %d\n", pPager, locktype))
    }
  }
  return rc;
}


/*
** Truncate the in-memory database file image to nPage pages. This 
** function does not actually modify the database file on disk. It 
** just sets the internal state of the pager object so that the 
** truncation will be done when the current transaction is committed.
*/
SQLITE_PRIVATE void sqlite3PagerTruncateImage(Pager *pPager, Pgno nPage){
  assert( pPager->dbSizeValid );
  assert( pPager->dbSize>=nPage );
  assert( pPager->state>=PAGER_RESERVED );
  pPager->dbSize = nPage;
}

















/*
** Shutdown the page cache.  Free all memory and close all files.
**
** If a transaction was in progress when this routine is called, that
** transaction is rolled back.  All outstanding pages are invalidated
** and their memory is freed.  Any attempt to use a page associated
** with this page cache after this function returns will likely
** result in a coredump.
**
** This function always succeeds. If a transaction is active an attempt
** is made to roll it back. If an error occurs during the rollback 
** a hot journal may be left in the filesystem but no error is returned
** to the caller.
*/
SQLITE_PRIVATE int sqlite3PagerClose(Pager *pPager){

  disable_simulated_io_errors();
  sqlite3BeginBenignMalloc();
  pPager->errCode = 0;
  pPager->exclusiveMode = 0;
  pager_reset(pPager);
  if( MEMDB ){
    pager_unlock(pPager);
  }else{
    /* Set Pager.journalHdr to -1 for the benefit of the pager_playback() 
    ** call which may be made from within pagerUnlockAndRollback(). If it
    ** is not -1, then the unsynced portion of an open journal file may
    ** be played back into the database. If a power failure occurs while
    ** this is happening, the database may become corrupt.
    */
    pPager->journalHdr = -1;
    pagerUnlockAndRollback(pPager);
  }

  sqlite3EndBenignMalloc();
  enable_simulated_io_errors();
  PAGERTRACE(("CLOSE %d\n", PAGERID(pPager)));
  IOTRACE(("CLOSE %p\n", pPager))

  sqlite3OsClose(pPager->fd);
  sqlite3PageFree(pPager->pTmpSpace);
  sqlite3PcacheClose(pPager->pPCache);



  assert( !pPager->aSavepoint && !pPager->pInJournal );



  assert( !isOpen(pPager->jfd) && !isOpen(pPager->sjfd) );





  sqlite3_free(pPager);
  return SQLITE_OK;
}

#if !defined(NDEBUG) || defined(SQLITE_TEST)
/*
** Return the page number for page pPg.
*/
SQLITE_PRIVATE Pgno sqlite3PagerPagenumber(DbPage *pPg){
  return pPg->pgno;
}
#endif

/*
** Increment the reference count for page pPg.

*/
SQLITE_PRIVATE void sqlite3PagerRef(DbPage *pPg){
  sqlite3PcacheRef(pPg);

}

/*
** Sync the journal. In other words, make sure all the pages that have
** been written to the journal have actually reached the surface of the
** disk and can be restored in the event of a hot-journal rollback.
**
** If the Pager.needSync flag is not set, then this function is a
** no-op. Otherwise, the actions required depend on the journal-mode
** and the device characteristics of the the file-system, as follows:
**
**   * If the journal file is an in-memory journal file, no action need


**     be taken.
**
**   * Otherwise, if the device does not support the SAFE_APPEND property,
**     then the nRec field of the most recently written journal header
**     is updated to contain the number of journal records that have
**     been written following it. If the pager is operating in full-sync
**     mode, then the journal file is synced before this field is updated.
**
**   * If the device does not support the SEQUENTIAL property, then 
**     journal file is synced.
**


** Or, in pseudo-code:
**

**   if( NOT <in-memory journal> ){
**     if( NOT SAFE_APPEND ){
**       if( <full-sync mode> ) xSync(<journal file>);
**       <update nRec field>
**     } 
**     if( NOT SEQUENTIAL ) xSync(<journal file>);
**   }
**
** The Pager.needSync flag is never be set for temporary files, or any
** file operating in no-sync mode (Pager.noSync set to non-zero).
**
** If successful, this routine clears the PGHDR_NEED_SYNC flag of every 
** page currently held in memory before returning SQLITE_OK. If an IO
** error is encountered, then the IO error code is returned to the caller.
*/
static int syncJournal(Pager *pPager){





  if( pPager->needSync ){
    assert( !pPager->tempFile );
    if( pPager->journalMode!=PAGER_JOURNALMODE_MEMORY ){
      int rc;                              /* Return code */
      const int iDc = sqlite3OsDeviceCharacteristics(pPager->fd);
      assert( isOpen(pPager->jfd) );

      if( 0==(iDc&SQLITE_IOCAP_SAFE_APPEND) ){
        /* Variable iNRecOffset is set to the offset in the journal file
        ** of the nRec field of the most recently written journal header.
        ** This field will be updated following the xSync() operation
        ** on the journal file. */
        i64 iNRecOffset = pPager->journalHdr + sizeof(aJournalMagic);

        /* This block deals with an obscure problem. If the last connection
        ** that wrote to this database was operating in persistent-journal
        ** mode, then the journal file may at this point actually be larger
        ** than Pager.journalOff bytes. If the next thing in the journal
        ** file happens to be a journal-header (written as part of the
        ** previous connections transaction), and a crash or power-failure 
        ** occurs after nRec is updated but before this connection writes 
        ** anything else to the journal file (or commits/rolls back its 
        ** transaction), then SQLite may become confused when doing the 
        ** hot-journal rollback following recovery. It may roll back all
        ** of this connections data, then proceed to rolling back the old,
        ** out-of-date data that follows it. Database corruption.
        **
        ** To work around this, if the journal file does appear to contain
        ** a valid header following Pager.journalOff, then write a 0x00
        ** byte to the start of it to prevent it from being recognized.
        **
        ** Variable iNextHdrOffset is set to the offset at which this
        ** problematic header will occur, if it exists. aMagic is used 
        ** as a temporary buffer to inspect the first couple of bytes of
        ** the potential journal header.
        */
        i64 iNextHdrOffset = journalHdrOffset(pPager);
        u8 aMagic[8];
        rc = sqlite3OsRead(pPager->jfd, aMagic, 8, iNextHdrOffset);
        if( rc==SQLITE_OK && 0==memcmp(aMagic, aJournalMagic, 8) ){
          static const u8 zerobyte = 0;
          rc = sqlite3OsWrite(pPager->jfd, &zerobyte, 1, iNextHdrOffset);
        }
        if( rc!=SQLITE_OK && rc!=SQLITE_IOERR_SHORT_READ ){
          return rc;
        }

        /* Write the nRec value into the journal file header. If in
        ** full-synchronous mode, sync the journal first. This ensures that
        ** all data has really hit the disk before nRec is updated to mark
        ** it as a candidate for rollback.
        **
        ** This is not required if the persistent media supports the
        ** SAFE_APPEND property. Because in this case it is not possible 
        ** for garbage data to be appended to the file, the nRec field
        ** is populated with 0xFFFFFFFF when the journal header is written
        ** and never needs to be updated.
        */
        if( pPager->fullSync && 0==(iDc&SQLITE_IOCAP_SEQUENTIAL) ){
          PAGERTRACE(("SYNC journal of %d\n", PAGERID(pPager)));
          IOTRACE(("JSYNC %p\n", pPager))
          rc = sqlite3OsSync(pPager->jfd, pPager->sync_flags);
          if( rc!=SQLITE_OK ) return rc;
        }


        IOTRACE(("JHDR %p %lld %d\n", pPager, iNRecOffset, 4));
        rc = write32bits(pPager->jfd, iNRecOffset, pPager->nRec);
        if( rc!=SQLITE_OK ) return rc;
      }
      if( 0==(iDc&SQLITE_IOCAP_SEQUENTIAL) ){
        PAGERTRACE(("SYNC journal of %d\n", PAGERID(pPager)));
        IOTRACE(("JSYNC %p\n", pPager))
        rc = sqlite3OsSync(pPager->jfd, pPager->sync_flags| 
          (pPager->sync_flags==SQLITE_SYNC_FULL?SQLITE_SYNC_DATAONLY:0)
        );
        if( rc!=SQLITE_OK ) return rc;
      }

    }


    /* The journal file was just successfully synced. Set Pager.needSync 
    ** to zero and clear the PGHDR_NEED_SYNC flag on all pagess.
    */
    pPager->needSync = 0;
    pPager->journalStarted = 1;
    sqlite3PcacheClearSyncFlags(pPager->pPCache);
  }

  return SQLITE_OK;
}

/*
** The argument is the first in a linked list of dirty pages connected
** by the PgHdr.pDirty pointer. This function writes each one of the
** in-memory pages in the list to the database file. The argument may
** be NULL, representing an empty list. In this case this function is
** a no-op.
**
** The pager must hold at least a RESERVED lock when this function
** is called. Before writing anything to the database file, this lock
** is upgraded to an EXCLUSIVE lock. If the lock cannot be obtained,
** SQLITE_BUSY is returned and no data is written to the database file.
** 
** If the pager is a temp-file pager and the actual file-system file
** is not yet open, it is created and opened before any data is 
** written out.
**
** Once the lock has been upgraded and, if necessary, the file opened,
** the pages are written out to the database file in list order. Writing
** a page is skipped if it meets either of the following criteria:
**
**   * The page number is greater than Pager.dbSize, or
**   * The PGHDR_DONT_WRITE flag is set on the page.
**
** If writing out a page causes the database file to grow, Pager.dbFileSize
** is updated accordingly. If page 1 is written out, then the value cached
** in Pager.dbFileVers[] is updated to match the new value stored in
** the database file.
**
** If everything is successful, SQLITE_OK is returned. If an IO error 
** occurs, an IO error code is returned. Or, if the EXCLUSIVE lock cannot
** be obtained, SQLITE_BUSY is returned.
*/
static int pager_write_pagelist(PgHdr *pList){
  Pager *pPager;                       /* Pager object */
  int rc;                              /* Return code */

  if( pList==0 ) return SQLITE_OK;
  pPager = pList->pPager;

  /* At this point there may be either a RESERVED or EXCLUSIVE lock on the
  ** database file. If there is already an EXCLUSIVE lock, the following
  ** call is a no-op.
  **
  ** Moving the lock from RESERVED to EXCLUSIVE actually involves going
  ** through an intermediate state PENDING.   A PENDING lock prevents new
  ** readers from attaching to the database but is unsufficient for us to
  ** write.  The idea of a PENDING lock is to prevent new readers from
  ** coming in while we wait for existing readers to clear.
  **
  ** While the pager is in the RESERVED state, the original database file
  ** is unchanged and we can rollback without having to playback the
  ** journal into the original database file.  Once we transition to
  ** EXCLUSIVE, it means the database file has been changed and any rollback
  ** will require a journal playback.
  */
  assert( pPager->state>=PAGER_RESERVED );
  rc = pager_wait_on_lock(pPager, EXCLUSIVE_LOCK);



  /* If the file is a temp-file has not yet been opened, open it now. It
  ** is not possible for rc to be other than SQLITE_OK if this branch
  ** is taken, as pager_wait_on_lock() is a no-op for temp-files.
  */

  if( !isOpen(pPager->fd) ){
    assert( pPager->tempFile && rc==SQLITE_OK );
    rc = pagerOpentemp(pPager, pPager->fd, pPager->vfsFlags);

  }

  while( rc==SQLITE_OK && pList ){
    Pgno pgno = pList->pgno;

    /* If there are dirty pages in the page cache with page numbers greater
    ** than Pager.dbSize, this means sqlite3PagerTruncateImage() was called to
    ** make the file smaller (presumably by auto-vacuum code). Do not write
    ** any such pages to the file.
    **
    ** Also, do not write out any page that has the PGHDR_DONT_WRITE flag
    ** set (set by sqlite3PagerDontWrite()).
    */
    if( pgno<=pPager->dbSize && 0==(pList->flags&PGHDR_DONT_WRITE) ){
      i64 offset = (pgno-1)*(i64)pPager->pageSize;         /* Offset to write */
      char *pData = CODEC2(pPager, pList->pData, pgno, 6); /* Data to write */

      /* Write out the page data. */


      rc = sqlite3OsWrite(pPager->fd, pData, pPager->pageSize, offset);

      /* If page 1 was just written, update Pager.dbFileVers to match
      ** the value now stored in the database file. If writing this 
      ** page caused the database file to grow, update dbFileSize. 
      */
      if( pgno==1 ){
        memcpy(&pPager->dbFileVers, &pData[24], sizeof(pPager->dbFileVers));
      }
      if( pgno>pPager->dbFileSize ){
        pPager->dbFileSize = pgno;
      }


      PAGERTRACE(("STORE %d page %d hash(%08x)\n",
                   PAGERID(pPager), pgno, pager_pagehash(pList)));
      IOTRACE(("PGOUT %p %d\n", pPager, pgno));
      PAGER_INCR(sqlite3_pager_writedb_count);
      PAGER_INCR(pPager->nWrite);
    }else{
      PAGERTRACE(("NOSTORE %d page %d\n", PAGERID(pPager), pgno));
    }


#ifdef SQLITE_CHECK_PAGES
    pList->pageHash = pager_pagehash(pList);
#endif
    pList = pList->pDirty;
  }

  return rc;
}

/*
** Append a record of the current state of page pPg to the sub-journal. 
** It is the callers responsibility to use subjRequiresPage() to check 
** that it is really required before calling this function.
**
** If successful, set the bit corresponding to pPg->pgno in the bitvecs
** for all open savepoints before returning.
**
** This function returns SQLITE_OK if everything is successful, an IO
** error code if the attempt to write to the sub-journal fails, or 
** SQLITE_NOMEM if a malloc fails while setting a bit in a savepoint
** bitvec.
*/
static int subjournalPage(PgHdr *pPg){
  int rc;
  void *pData = pPg->pData;
  Pager *pPager = pPg->pPager;
  i64 offset = pPager->nSubRec*(4+pPager->pageSize);
  char *pData2 = CODEC2(pPager, pData, pPg->pgno, 7);

  PAGERTRACE(("STMT-JOURNAL %d page %d\n", PAGERID(pPager), pPg->pgno));

  assert( pageInJournal(pPg) || pPg->pgno>pPager->dbOrigSize );
  rc = write32bits(pPager->sjfd, offset, pPg->pgno);
  if( rc==SQLITE_OK ){
    rc = sqlite3OsWrite(pPager->sjfd, pData2, pPager->pageSize, offset+4);
  }
  if( rc==SQLITE_OK ){
    pPager->nSubRec++;
    assert( pPager->nSavepoint>0 );
    rc = addToSavepointBitvecs(pPager, pPg->pgno);
    testcase( rc!=SQLITE_OK );
  }
  return rc;
}


/*
** This function is called by the pcache layer when it has reached some
** soft memory limit. The first argument is a pointer to a Pager object
** (cast as a void*). The pager is always 'purgeable' (not an in-memory
** database). The second argument is a reference to a page that is 
** currently dirty but has no outstanding references. The page
** is always associated with the Pager object passed as the first 
** argument.
**
** The job of this function is to make pPg clean by writing its contents
** out to the database file, if possible. This may involve syncing the
** journal file. 
**
** If successful, sqlite3PcacheMakeClean() is called on the page and
** SQLITE_OK returned. If an IO error occurs while trying to make the
** page clean, the IO error code is returned. If the page cannot be
** made clean for some other reason, but no error occurs, then SQLITE_OK
** is returned by sqlite3PcacheMakeClean() is not called.
*/
static int pagerStress(void *p, PgHdr *pPg){
  Pager *pPager = (Pager *)p;
  int rc = SQLITE_OK;




  assert( pPg->pPager==pPager );
  assert( pPg->flags&PGHDR_DIRTY );

  /* The doNotSync flag is set by the sqlite3PagerWrite() function while it
  ** is journalling a set of two or more database pages that are stored
  ** on the same disk sector. Syncing the journal is not allowed while
  ** this is happening as it is important that all members of such a
  ** set of pages are synced to disk together. So, if the page this function
  ** is trying to make clean will require a journal sync and the doNotSync
  ** flag is set, return without doing anything. The pcache layer will
  ** just have to go ahead and allocate a new page buffer instead of
  ** reusing pPg.
  **
  ** Similarly, if the pager has already entered the error state, do not
  ** try to write the contents of pPg to disk.
  */
  if( pPager->errCode || (pPager->doNotSync && pPg->flags&PGHDR_NEED_SYNC) ){
    return SQLITE_OK;
  }

  /* Sync the journal file if required. */
  if( pPg->flags&PGHDR_NEED_SYNC ){
    rc = syncJournal(pPager);
    if( rc==SQLITE_OK && pPager->fullSync && 
      !(pPager->journalMode==PAGER_JOURNALMODE_MEMORY) &&
      !(sqlite3OsDeviceCharacteristics(pPager->fd)&SQLITE_IOCAP_SAFE_APPEND)
    ){
      pPager->nRec = 0;
      rc = writeJournalHdr(pPager);
    }
  }

  /* If the page number of this page is larger than the current size of
  ** the database image, it may need to be written to the sub-journal.
  ** This is because the call to pager_write_pagelist() below will not
  ** actually write data to the file in this case.
  **
  ** Consider the following sequence of events:
  **
  **   BEGIN;
  **     <journal page X>
  **     <modify page X>
  **     SAVEPOINT sp;
  **       <shrink database file to Y pages>
  **       pagerStress(page X)
  **     ROLLBACK TO sp;
  **
  ** If (X>Y), then when pagerStress is called page X will not be written
  ** out to the database file, but will be dropped from the cache. Then,
  ** following the "ROLLBACK TO sp" statement, reading page X will read
  ** data from the database file. This will be the copy of page X as it
  ** was when the transaction started, not as it was when "SAVEPOINT sp"
  ** was executed.
  **
  ** The solution is to write the current data for page X into the 
  ** sub-journal file now (if it is not already there), so that it will
  ** be restored to its current value when the "ROLLBACK TO sp" is 
  ** executed.
  */
  if( rc==SQLITE_OK && pPg->pgno>pPager->dbSize && subjRequiresPage(pPg) ){
    rc = subjournalPage(pPg);
  }

  /* Write the contents of the page out to the database file. */
  if( rc==SQLITE_OK ){
    pPg->pDirty = 0;
    rc = pager_write_pagelist(pPg);
  }





  /* Mark the page as clean. */
  if( rc==SQLITE_OK ){
    PAGERTRACE(("STRESS %d page %d\n", PAGERID(pPager), pPg->pgno));
    sqlite3PcacheMakeClean(pPg);
  }

  return pager_error(pPager, rc);
}


/*
** Allocate and initialize a new Pager object and put a pointer to it
** in *ppPager. The pager should eventually be freed by passing it
** to sqlite3PagerClose().
**
** The zFilename argument is the path to the database file to open.
** If zFilename is NULL then a randomly-named temporary file is created
** and used as the file to be cached. Temporary files are be deleted
** automatically when they are closed. If zFilename is ":memory:" then 
** all information is held in cache. It is never written to disk. 
** This can be used to implement an in-memory database.
**
** The nExtra parameter specifies the number of bytes of space allocated
** along with each page reference. This space is available to the user
** via the sqlite3PagerGetExtra() API.
**
** The flags argument is used to specify properties that affect the
** operation of the pager. It should be passed some bitwise combination
** of the PAGER_OMIT_JOURNAL and PAGER_NO_READLOCK flags.
**
** The vfsFlags parameter is a bitmask to pass to the flags parameter
** of the xOpen() method of the supplied VFS when opening files. 
**
** If the pager object is allocated and the specified file opened 
** successfully, SQLITE_OK is returned and *ppPager set to point to
** the new pager object. If an error occurs, *ppPager is set to NULL
** and error code returned. This function may return SQLITE_NOMEM
** (sqlite3Malloc() is used to allocate memory), SQLITE_CANTOPEN or 
** various SQLITE_IO_XXX errors.
*/
SQLITE_PRIVATE int sqlite3PagerOpen(
  sqlite3_vfs *pVfs,       /* The virtual file system to use */
  Pager **ppPager,         /* OUT: Return the Pager structure here */
  const char *zFilename,   /* Name of the database file to open */
  int nExtra,              /* Extra bytes append to each in-memory page */
  int flags,               /* flags controlling this file */
  int vfsFlags             /* flags passed through to sqlite3_vfs.xOpen() */
){
  u8 *pPtr;
  Pager *pPager = 0;       /* Pager object to allocate and return */
  int rc = SQLITE_OK;      /* Return code */
  int tempFile = 0;        /* True for temp files (incl. in-memory files) */
  int memDb = 0;           /* True if this is an in-memory file */
  int readOnly = 0;        /* True if this is a read-only file */
  int journalFileSize;     /* Bytes to allocate for each journal fd */
  char *zPathname = 0;     /* Full path to database file */
  int nPathname = 0;       /* Number of bytes in zPathname */
  int useJournal = (flags & PAGER_OMIT_JOURNAL)==0; /* False to omit journal */
  int noReadlock = (flags & PAGER_NO_READLOCK)!=0;  /* True to omit read-lock */
  int pcacheSize = sqlite3PcacheSize();       /* Bytes to allocate for PCache */
  u16 szPageDflt = SQLITE_DEFAULT_PAGE_SIZE;  /* Default page size */

  /* Figure out how much space is required for each journal file-handle
  ** (there are two of them, the main journal and the sub-journal). This
  ** is the maximum space required for an in-memory journal file handle 
  ** and a regular journal file-handle. Note that a "regular journal-handle"
  ** may be a wrapper capable of caching the first portion of the journal
  ** file in memory to implement the atomic-write optimization (see 
  ** source file journal.c).
  */
  if( sqlite3JournalSize(pVfs)>sqlite3MemJournalSize() ){
    journalFileSize = sqlite3JournalSize(pVfs);
  }else{
    journalFileSize = sqlite3MemJournalSize();
  }

  /* Set the output variable to NULL in case an error occurs. */
  *ppPager = 0;

  /* Compute and store the full pathname in an allocated buffer pointed
  ** to by zPathname, length nPathname. Or, if this is a temporary file,
  ** leave both nPathname and zPathname set to 0.
  */
  if( zFilename && zFilename[0] ){
    nPathname = pVfs->mxPathname+1;
    zPathname = sqlite3Malloc(nPathname*2);
    if( zPathname==0 ){
      return SQLITE_NOMEM;
    }
#ifndef SQLITE_OMIT_MEMORYDB
    if( strcmp(zFilename,":memory:")==0 ){
      memDb = 1;
      zPathname[0] = 0;
    }else
#endif
    {
      rc = sqlite3OsFullPathname(pVfs, zFilename, nPathname, zPathname);
    }

    nPathname = sqlite3Strlen30(zPathname);
    if( rc==SQLITE_OK && nPathname+8>pVfs->mxPathname ){
      /* This branch is taken when the journal path required by
      ** the database being opened will be more than pVfs->mxPathname
      ** bytes in length. This means the database cannot be opened,
      ** as it will not be possible to open the journal file or even
      ** check for a hot-journal before reading.
      */
      rc = SQLITE_CANTOPEN;
    }
    if( rc!=SQLITE_OK ){
      sqlite3_free(zPathname);
      return rc;
    }
  }

  /* Allocate memory for the Pager structure, PCache object, the
  ** three file descriptors, the database file name and the journal 
  ** file name. The layout in memory is as follows:
  **
  **     Pager object                    (sizeof(Pager) bytes)
  **     PCache object                   (sqlite3PcacheSize() bytes)
  **     Database file handle            (pVfs->szOsFile bytes)
  **     Sub-journal file handle         (journalFileSize bytes)
  **     Main journal file handle        (journalFileSize bytes)
  **     Database file name              (nPathname+1 bytes)
  **     Journal file name               (nPathname+8+1 bytes)
  */
  pPtr = (u8 *)sqlite3MallocZero(
    sizeof(*pPager) +           /* Pager structure */
    pcacheSize      +           /* PCache object */
    pVfs->szOsFile  +           /* The main db file */
    journalFileSize * 2 +       /* The two journal files */ 
    nPathname + 1 +             /* zFilename */
    nPathname + 8 + 1           /* zJournal */
  );
  if( !pPtr ){
    sqlite3_free(zPathname);
    return SQLITE_NOMEM;
  }
  pPager =              (Pager*)(pPtr);
  pPager->pPCache =    (PCache*)(pPtr += sizeof(*pPager));
  pPager->fd =   (sqlite3_file*)(pPtr += pcacheSize);
  pPager->sjfd = (sqlite3_file*)(pPtr += pVfs->szOsFile);
  pPager->jfd =  (sqlite3_file*)(pPtr += journalFileSize);
  pPager->zFilename =    (char*)(pPtr += journalFileSize);

  /* Fill in the Pager.zFilename and Pager.zJournal buffers, if required. */
  if( zPathname ){
    pPager->zJournal =   (char*)(pPtr += nPathname + 1);
    memcpy(pPager->zFilename, zPathname, nPathname);
    memcpy(pPager->zJournal, zPathname, nPathname);
    memcpy(&pPager->zJournal[nPathname], "-journal", 8);
    sqlite3_free(zPathname);
  }
  pPager->pVfs = pVfs;
  pPager->vfsFlags = vfsFlags;

  /* Open the pager file.
  */
  if( zFilename && zFilename[0] && !memDb ){
    int fout = 0;                    /* VFS flags returned by xOpen() */
    rc = sqlite3OsOpen(pVfs, pPager->zFilename, pPager->fd, vfsFlags, &fout);
    readOnly = (fout&SQLITE_OPEN_READONLY);

    /* If the file was successfully opened for read/write access,
    ** choose a default page size in case we have to create the
    ** database file. The default page size is the maximum of:
    **
    **    + SQLITE_DEFAULT_PAGE_SIZE,
    **    + The value returned by sqlite3OsSectorSize()
    **    + The largest page size that can be written atomically.
    */
    if( rc==SQLITE_OK && !readOnly ){
      setSectorSize(pPager);
      assert(SQLITE_DEFAULT_PAGE_SIZE<=SQLITE_MAX_DEFAULT_PAGE_SIZE);
      if( szPageDflt<pPager->sectorSize ){
        if( pPager->sectorSize>SQLITE_MAX_DEFAULT_PAGE_SIZE ){
          szPageDflt = SQLITE_MAX_DEFAULT_PAGE_SIZE;
        }else{
          szPageDflt = (u16)pPager->sectorSize;
        }
      }
#ifdef SQLITE_ENABLE_ATOMIC_WRITE
      {
        int iDc = sqlite3OsDeviceCharacteristics(pPager->fd);
        int ii;
        assert(SQLITE_IOCAP_ATOMIC512==(512>>8));
        assert(SQLITE_IOCAP_ATOMIC64K==(65536>>8));
        assert(SQLITE_MAX_DEFAULT_PAGE_SIZE<=65536);
        for(ii=szPageDflt; ii<=SQLITE_MAX_DEFAULT_PAGE_SIZE; ii=ii*2){
          if( iDc&(SQLITE_IOCAP_ATOMIC|(ii>>8)) ){
            szPageDflt = ii;
          }
        }
      }
#endif
    }
  }else{
    /* If a temporary file is requested, it is not opened immediately.
    ** In this case we accept the default page size and delay actually
    ** opening the file until the first call to OsWrite().
    **
    ** This branch is also run for an in-memory database. An in-memory
    ** database is the same as a temp-file that is never written out to
    ** disk and uses an in-memory rollback journal.
    */ 
    tempFile = 1;
    pPager->state = PAGER_EXCLUSIVE;
  }

  /* The following call to PagerSetPagesize() serves to set the value of 
  ** Pager.pageSize and to allocate the Pager.pTmpSpace buffer.
  */
  if( rc==SQLITE_OK ){
    assert( pPager->memDb==0 );
    rc = sqlite3PagerSetPagesize(pPager, &szPageDflt);
    testcase( rc!=SQLITE_OK );
  }

  /* If an error occured in either of the blocks above, free the 
  ** Pager structure and close the file.
  */
  if( rc!=SQLITE_OK ){
    assert( !pPager->pTmpSpace );
    sqlite3OsClose(pPager->fd);
    sqlite3_free(pPager);
    return rc;
  }

  /* Initialize the PCache object. */
  nExtra = FORCE_ALIGNMENT(nExtra);
  sqlite3PcacheOpen(szPageDflt, nExtra, !memDb,
                    !memDb?pagerStress:0, (void *)pPager, pPager->pPCache);

  PAGERTRACE(("OPEN %d %s\n", FILEHANDLEID(pPager->fd), pPager->zFilename));
  IOTRACE(("OPEN %p %s\n", pPager, pPager->zFilename))

  pPager->useJournal = (u8)useJournal;
  pPager->noReadlock = (noReadlock && readOnly) ?1:0;
  /* pPager->stmtOpen = 0; */
  /* pPager->stmtInUse = 0; */
  /* pPager->nRef = 0; */
  pPager->dbSizeValid = (u8)memDb;
  /* pPager->stmtSize = 0; */
  /* pPager->stmtJSize = 0; */
  /* pPager->nPage = 0; */
  pPager->mxPgno = SQLITE_MAX_PAGE_COUNT;
  /* pPager->state = PAGER_UNLOCK; */
  assert( pPager->state == (tempFile ? PAGER_EXCLUSIVE : PAGER_UNLOCK) );
  /* pPager->errMask = 0; */
  pPager->tempFile = (u8)tempFile;
  assert( tempFile==PAGER_LOCKINGMODE_NORMAL 
          || tempFile==PAGER_LOCKINGMODE_EXCLUSIVE );
  assert( PAGER_LOCKINGMODE_EXCLUSIVE==1 );
  pPager->exclusiveMode = (u8)tempFile; 
  pPager->changeCountDone = pPager->tempFile;
  pPager->memDb = (u8)memDb;
  pPager->readOnly = (u8)readOnly;
  /* pPager->needSync = 0; */
  pPager->noSync = (pPager->tempFile || !useJournal) ?1:0;
  pPager->fullSync = pPager->noSync ?0:1;
  pPager->sync_flags = SQLITE_SYNC_NORMAL;
  /* pPager->pFirst = 0; */
  /* pPager->pFirstSynced = 0; */
  /* pPager->pLast = 0; */
  pPager->nExtra = nExtra;
  pPager->journalSizeLimit = SQLITE_DEFAULT_JOURNAL_SIZE_LIMIT;
  assert( isOpen(pPager->fd) || tempFile );
  setSectorSize(pPager);
  if( memDb ){
    pPager->journalMode = PAGER_JOURNALMODE_MEMORY;
  }
  /* pPager->xBusyHandler = 0; */
  /* pPager->pBusyHandlerArg = 0; */
  /* memset(pPager->aHash, 0, sizeof(pPager->aHash)); */
  *ppPager = pPager;
  return SQLITE_OK;
}



/*
** This function is called after transitioning from PAGER_UNLOCK to
** PAGER_SHARED state. It tests if there is a hot journal present in
** the file-system for the given pager. A hot journal is one that 
** needs to be played back. According to this function, a hot-journal
** file exists if the following three criteria are met:
**
**   * The journal file exists in the file system, and
**   * No process holds a RESERVED or greater lock on the database file, and
**   * The database file itself is greater than 0 bytes in size.
**
** If the current size of the database file is 0 but a journal file
** exists, that is probably an old journal left over from a prior
** database with the same name. In this case the journal file is
** just deleted using OsDelete, *pExists is set to 0 and SQLITE_OK
** is returned.

**
** This routine does not open the journal file to examine its
** content.  Hence, the journal might contain the name of a master
** journal file that has been deleted, and hence not be hot.  Or
** the header of the journal might be zeroed out.  This routine
** does not discover these cases of a non-hot journal - if the
** journal file exists and is not empty this routine assumes it
** is hot.  The pager_playback() routine will discover that the
** journal file is not really hot and will no-op.
**
** If a hot-journal file is found to exist, *pExists is set to 1 and 
** SQLITE_OK returned. If no hot-journal file is present, *pExists is
** set to 0 and SQLITE_OK returned. If an IO error occurs while trying
** to determine whether or not a hot-journal file exists, the IO error
** code is returned and the value of *pExists is undefined.
*/
static int hasHotJournal(Pager *pPager, int *pExists){
  sqlite3_vfs * const pVfs = pPager->pVfs;
  int rc;                       /* Return code */
  int exists = 0;               /* True if a journal file is present */
  int locked = 0;               /* True if some process holds a RESERVED lock */

  assert( pPager!=0 );
  assert( pPager->useJournal );
  assert( isOpen(pPager->fd) );

  *pExists = 0;
  rc = sqlite3OsAccess(pVfs, pPager->zJournal, SQLITE_ACCESS_EXISTS, &exists);
  if( rc==SQLITE_OK && exists ){
    rc = sqlite3OsCheckReservedLock(pPager->fd, &locked);

    if( rc==SQLITE_OK && !locked ){
      int nPage;
      rc = sqlite3PagerPagecount(pPager, &nPage);
      if( rc==SQLITE_OK ){
       if( nPage==0 ){
          sqlite3OsDelete(pVfs, pPager->zJournal, 0);
        }else{
          *pExists = 1;
        }
      }
    }
  }
  return rc;
}

/*
** Read the content for page pPg out of the database file and into 
** pPg->pData. A shared lock or greater must be held on the database
** file before this function is called.
**
** If page 1 is read, then the value of Pager.dbFileVers[] is set to
** the value read from the database file.
**
** If an IO error occurs, then the IO error is returned to the caller.
** Otherwise, SQLITE_OK is returned.
*/
static int readDbPage(PgHdr *pPg){
  Pager *pPager = pPg->pPager; /* Pager object associated with page pPg */
  Pgno pgno = pPg->pgno;       /* Page number to read */
  int rc;                      /* Return code */
  i64 iOffset;                 /* Byte offset of file to read from */

  assert( pPager->state>=PAGER_SHARED && !MEMDB );

  if( !isOpen(pPager->fd) ){
    assert( pPager->tempFile );
    return SQLITE_IOERR_SHORT_READ;
  }
  iOffset = (pgno-1)*(i64)pPager->pageSize;
  rc = sqlite3OsRead(pPager->fd, pPg->pData, pPager->pageSize, iOffset);
  if( pgno==1 ){
    u8 *dbFileVers = &((u8*)pPg->pData)[24];
    memcpy(&pPager->dbFileVers, dbFileVers, sizeof(pPager->dbFileVers));
  }
  CODEC1(pPager, pPg->pData, pgno, 3);

  PAGER_INCR(sqlite3_pager_readdb_count);
  PAGER_INCR(pPager->nRead);
  IOTRACE(("PGIN %p %d\n", pPager, pgno));





  PAGERTRACE(("FETCH %d page %d hash(%08x)\n",
               PAGERID(pPager), pgno, pager_pagehash(pPg)));

  return rc;
}


/*
** This function is called whenever the upper layer requests a database
** page is requested, before the cache is checked for a suitable page
** or any data is read from the database. It performs the following
** two functions:
**
**   1) If the pager is currently in PAGER_UNLOCK state (no lock held
**      on the database file), then an attempt is made to obtain a
**      SHARED lock on the database file. Immediately after obtaining
**      the SHARED lock, the file-system is checked for a hot-journal,
**      which is played back if present. Following any hot-journal 
**      rollback, the contents of the cache are validated by checking
**      the 'change-counter' field of the database file header and
**      discarded if they are found to be invalid.
**
**   2) If the pager is running in exclusive-mode, and there are currently
**      no outstanding references to any pages, and is in the error state,
**      then an attempt is made to clear the error state by discarding
**      the contents of the page cache and rolling back any open journal
**      file.
**
** If the operation described by (2) above is not attempted, and if the
** pager is in an error state other than SQLITE_FULL when this is called,
** the error state error code is returned. It is permitted to read the
** database when in SQLITE_FULL error state.
**
** Otherwise, if everything is successful, SQLITE_OK is returned. If an
** IO error occurs while locking the database, checking for a hot-journal
** file or rolling back a journal file, the IO error code is returned.
*/
static int pagerSharedLock(Pager *pPager){
  int rc = SQLITE_OK;                /* Return code */
  int isErrorReset = 0;              /* True if recovering from error state */

  /* If this database is opened for exclusive access, has no outstanding 
  ** page references and is in an error-state, this is a chance to clear
  ** the error. Discard the contents of the pager-cache and treat any
  ** open journal file as a hot-journal.
  */
  if( !MEMDB && pPager->exclusiveMode 
   && sqlite3PcacheRefCount(pPager->pPCache)==0 && pPager->errCode 
  ){
    if( isOpen(pPager->jfd) ){
      isErrorReset = 1;
    }
    pPager->errCode = SQLITE_OK;
    pager_reset(pPager);
  }

  /* If the pager is still in an error state, do not proceed. The error 
  ** state will be cleared at some point in the future when all page 
  ** references are dropped and the cache can be discarded.
  */
  if( pPager->errCode && pPager->errCode!=SQLITE_FULL ){
    return pPager->errCode;
  }

  if( pPager->state==PAGER_UNLOCK || isErrorReset ){
    sqlite3_vfs * const pVfs = pPager->pVfs;
    int isHotJournal = 0;
    assert( !MEMDB );
    assert( sqlite3PcacheRefCount(pPager->pPCache)==0 );
    if( !pPager->noReadlock ){
      rc = pager_wait_on_lock(pPager, SHARED_LOCK);
      if( rc!=SQLITE_OK ){
        assert( pPager->state==PAGER_UNLOCK );

      }
    }
    if( isErrorReset || isHotJournal ){
      /* Get an EXCLUSIVE lock on the database file. At this point it is
      ** important that a RESERVED lock is not obtained on the way to the
      ** EXCLUSIVE lock. If it were, another process might open the
      ** database file, detect the RESERVED lock, and conclude that the
      ** database is safe to read while this process is still rolling the 
      ** hot-journal back.
      ** 
      ** Because the intermediate RESERVED lock is not requested, any
      ** other process attempting to access the database file will get to 
      ** this point in the code and fail to obtain its own EXCLUSIVE lock 
      ** on the database file.
      */
      if( pPager->state<EXCLUSIVE_LOCK ){
        rc = sqlite3OsLock(pPager->fd, EXCLUSIVE_LOCK);
        if( rc!=SQLITE_OK ){
          rc = pager_error(pPager, rc);
          goto failed;
        }
        pPager->state = PAGER_EXCLUSIVE;
      }
 
      /* Open the journal for read/write access. This is because in 
      ** exclusive-access mode the file descriptor will be kept open and
      ** possibly used for a transaction later on. On some systems, the
      ** OsTruncate() call used in exclusive-access mode also requires
      ** a read/write file handle.
      */
      if( !isOpen(pPager->jfd) ){
        int res;
        rc = sqlite3OsAccess(pVfs,pPager->zJournal,SQLITE_ACCESS_EXISTS,&res);
        if( rc==SQLITE_OK ){
          if( res ){
            int fout = 0;
            int f = SQLITE_OPEN_READWRITE|SQLITE_OPEN_MAIN_JOURNAL;
            assert( !pPager->tempFile );
            rc = sqlite3OsOpen(pVfs, pPager->zJournal, pPager->jfd, f, &fout);
            assert( rc!=SQLITE_OK || isOpen(pPager->jfd) );
            if( rc==SQLITE_OK && fout&SQLITE_OPEN_READONLY ){
              rc = SQLITE_CANTOPEN;
              sqlite3OsClose(pPager->jfd);
            }
          }else{
            /* If the journal does not exist, that means some other process
            ** has already rolled it back */
            rc = SQLITE_BUSY;
          }
        }
      }
      if( rc!=SQLITE_OK ){
        goto failed;
      }

      /* TODO: Why are these cleared here? Is it necessary? */
      pPager->journalStarted = 0;
      pPager->journalOff = 0;
      pPager->setMaster = 0;
      pPager->journalHdr = 0;
 
      /* Playback and delete the journal.  Drop the database write
      ** lock and reacquire the read lock. Purge the cache before
      ** playing back the hot-journal so that we don't end up with
      ** an inconsistent cache.
      */
      sqlite3PcacheClear(pPager->pPCache);
      rc = pager_playback(pPager, 1);
      if( rc!=SQLITE_OK ){
        rc = pager_error(pPager, rc);
        goto failed;
      }
      assert( (pPager->state==PAGER_SHARED)
           || (pPager->exclusiveMode && pPager->state>PAGER_SHARED)
      );
    }

    if( sqlite3PcachePagecount(pPager->pPCache)>0 ){
      /* The shared-lock has just been acquired on the database file
      ** and there are already pages in the cache (from a previous
      ** read or write transaction).  Check to see if the database

    /* pager_unlock() is a no-op for exclusive mode and in-memory databases. */
    pager_unlock(pPager);
  }
  return rc;
}

/*



















** If the reference count has reached zero, rollback any active
** transaction and unlock the pager.
*/ 
static void pagerUnlockIfUnused(Pager *pPager){
  if( sqlite3PcacheRefCount(pPager->pPCache)==0 ){


    pagerUnlockAndRollback(pPager);
  }
}

/*
** Drop a page from the cache using sqlite3PcacheDrop().
**
** If this means there are now no pages with references to them, a rollback
** occurs and the lock on the database is removed.
*/
static void pagerDropPage(DbPage *pPg){
  Pager *pPager = pPg->pPager;
  sqlite3PcacheDrop(pPg);
  pagerUnlockIfUnused(pPager);
}

/*
** Acquire a reference to page number pgno in pager pPager (a page
** reference has type DbPage*). If the requested reference is 
** successfully obtained, it is copied to *ppPage and SQLITE_OK returned.
**

** This function calls pagerSharedLock() to obtain a SHARED lock on
** the database file if such a lock or greater is not already held.
** This may cause hot-journal rollback or a cache purge. See comments
** above function pagerSharedLock() for details.
**

** If the requested page is already in the cache, it is returned. 
** Otherwise, a new page object is allocated and populated with data
** read from the database file. In some cases, the pcache module may
** choose not to allocate a new page object and may reuse an existing
** object with no outstanding references.
**
** The extra data appended to a page is always initialized to zeros the 
** first time a page is loaded into memory. If the page requested is 
** already in the cache when this function is called, then the extra
** data is left as it was when the page object was last used.
**
** If the database image is smaller than the requested page or if a 
** non-zero value is passed as the noContent parameter and the 
** requested page is not already stored in the cache, then no 
** actual disk read occurs. In this case the memory image of the 
** page is initialized to all zeros. 
**
** If noContent is true, it means that we do not care about the contents
** of the page. This occurs in two seperate scenarios:
**
**   a) When reading a free-list leaf page from the database, and
**
**   b) When a savepoint is being rolled back and we need to load
**      a new page into the cache to populate with the data read
**      from the savepoint journal.
**
** If noContent is true, then the data returned is zeroed instead of
** being read from the database. Additionally, the bits corresponding
** to pgno in Pager.pInJournal (bitvec of pages already written to the
** journal file) and the PagerSavepoint.pInSavepoint bitvecs of any open
** savepoints are set. This means if the page is made writable at any
** point in the future, using a call to sqlite3PagerWrite(), its contents
** will not be journaled. This saves IO.
**
** The acquisition might fail for several reasons.  In all cases,
** an appropriate error code is returned and *ppPage is set to NULL.
**
** See also sqlite3PagerLookup().  Both this routine and Lookup() attempt
** to find a page in the in-memory cache first.  If the page is not already
** in memory, this routine goes to disk to read it in whereas Lookup()
** just returns 0.  This routine acquires a read-lock the first time it
** has to go to disk, and could also playback an old journal if necessary.
** Since Lookup() never goes to disk, it never has to deal with locks
** or journal files.









*/
SQLITE_PRIVATE int sqlite3PagerAcquire(
  Pager *pPager,      /* The pager open on the database file */
  Pgno pgno,          /* Page number to fetch */
  DbPage **ppPage,    /* Write a pointer to the page here */
  int noContent       /* Do not bother reading content from disk if true */
){
  PgHdr *pPg = 0;
  int rc;

  assert( assert_pager_state(pPager) );
  assert( pPager->state==PAGER_UNLOCK 
       || sqlite3PcacheRefCount(pPager->pPCache)>0 
       || pgno==1
  );

  /* The maximum page number is 2^31. Return SQLITE_CORRUPT if a page
  ** number greater than this, or zero, is requested.

  }
  assert( pPager->state!=PAGER_UNLOCK );

  rc = sqlite3PcacheFetch(pPager->pPCache, pgno, 1, &pPg);
  if( rc!=SQLITE_OK ){
    return rc;
  }
  assert( pPg->pgno==pgno );
  assert( pPg->pPager==pPager || pPg->pPager==0 );
  if( pPg->pPager==0 ){
    /* The pager cache has created a new page. Its content needs to 
    ** be initialized.
    */
    int nMax;
    PAGER_INCR(pPager->nMiss);
    pPg->pPager = pPager;


    rc = sqlite3PagerPagecount(pPager, &nMax);
    if( rc!=SQLITE_OK ){
      sqlite3PagerUnref(pPg);
      return rc;
    }

    if( nMax<(int)pgno || MEMDB || noContent ){
      if( pgno>pPager->mxPgno ){
        sqlite3PagerUnref(pPg);
        return SQLITE_FULL;
      }

      if( noContent ){
        /* Failure to set the bits in the InJournal bit-vectors is benign.
        ** It merely means that we might do some extra work to journal a 
        ** page that does not need to be journaled.  Nevertheless, be sure 
        ** to test the case where a malloc error occurs while trying to set 
        ** a bit in a bit vector.
        */
        sqlite3BeginBenignMalloc();
        if( pgno<=pPager->dbOrigSize ){
          TESTONLY( rc = ) sqlite3BitvecSet(pPager->pInJournal, pgno);
          testcase( rc==SQLITE_NOMEM );
        }
        TESTONLY( rc = ) addToSavepointBitvecs(pPager, pgno);
        testcase( rc==SQLITE_NOMEM );
        sqlite3EndBenignMalloc();
      }else{
        memset(pPg->pData, 0, pPager->pageSize);
      }
      IOTRACE(("ZERO %p %d\n", pPager, pgno));
    }else{
      assert( pPg->pPager==pPager );
      rc = readDbPage(pPg);
      if( rc!=SQLITE_OK && rc!=SQLITE_IOERR_SHORT_READ ){

        pagerDropPage(pPg);
        return rc;
      }
    }
#ifdef SQLITE_CHECK_PAGES
    pPg->pageHash = pager_pagehash(pPg);
#endif
  }else{
    /* The requested page is in the page cache. */

    PAGER_INCR(pPager->nHit);







  }

  *ppPage = pPg;
  return SQLITE_OK;
}

/*
** Acquire a page if it is already in the in-memory cache.  Do
** not read the page from disk.  Return a pointer to the page,
** or 0 if the page is not in cache. Also, return 0 if the 
** pager is in PAGER_UNLOCK state when this function is called,
** or if the pager is in an error state other than SQLITE_FULL.
**
** See also sqlite3PagerGet().  The difference between this routine
** and sqlite3PagerGet() is that _get() will go to the disk and read
** in the page if the page is not already in cache.  This routine
** returns NULL if the page is not in cache or if a disk I/O error 
** has ever happened.
*/

    sqlite3PcacheFetch(pPager->pPCache, pgno, 0, &pPg);
  }

  return pPg;
}

/*
** Release a page reference.
**
** If the number of references to the page drop to zero, then the
** page is added to the LRU list.  When all references to all pages
** are released, a rollback occurs and the lock on the database is
** removed.
*/
SQLITE_PRIVATE void sqlite3PagerUnref(DbPage *pPg){
  if( pPg ){
    Pager *pPager = pPg->pPager;
    sqlite3PcacheRelease(pPg);
    pagerUnlockIfUnused(pPager);
  }

}

/*
** If the main journal file has already been opened, ensure that the
** sub-journal file is open too. If the main journal is not open,
** this function is a no-op.
**
** SQLITE_OK is returned if everything goes according to plan. 
** An SQLITE_IOERR_XXX error code is returned if a call to 
** sqlite3OsOpen() fails.
*/
static int openSubJournal(Pager *pPager){
  int rc = SQLITE_OK;
  if( isOpen(pPager->jfd) && !isOpen(pPager->sjfd) ){
    if( pPager->journalMode==PAGER_JOURNALMODE_MEMORY ){
      sqlite3MemJournalOpen(pPager->sjfd);
    }else{
      rc = pagerOpentemp(pPager, pPager->sjfd, SQLITE_OPEN_SUBJOURNAL);
    }
  }
  return rc;
}

/*
** This function is called at the start of every write transaction.
** There must already be a RESERVED or EXCLUSIVE lock on the database 
** file when this routine is called.
**
** Open the journal file for pager pPager and write a journal header
** to the start of it. If there are active savepoints, open the sub-journal
** as well. This function is only used when the journal file is being 
** opened to write a rollback log for a transaction. It is not used 
** when opening a hot journal file to roll it back.
**
** If the journal file is already open (as it may be in exclusive mode),
** then this function just writes a journal header to the start of the
** already open file. 
**
** Whether or not the journal file is opened by this function, the
** Pager.pInJournal bitvec structure is allocated.
**
** Return SQLITE_OK if everything is successful. Otherwise, return 

** SQLITE_NOMEM if the attempt to allocate Pager.pInJournal fails, or 
** an IO error code if opening or writing the journal file fails.
*/
static int pager_open_journal(Pager *pPager){
  int rc = SQLITE_OK;                        /* Return code */
  sqlite3_vfs * const pVfs = pPager->pVfs;   /* Local cache of vfs pointer */



  assert( pPager->state>=PAGER_RESERVED );
  assert( pPager->useJournal );
  assert( pPager->pInJournal==0 );
  
  /* If already in the error state, this function is a no-op. */
  if( pPager->errCode ){
    return pPager->errCode;
  }

  /* TODO: Is it really possible to get here with dbSizeValid==0? If not,
  ** the call to PagerPagecount() can be removed.
  */
  testcase( pPager->dbSizeValid==0 );
  sqlite3PagerPagecount(pPager, 0);

  pPager->pInJournal = sqlite3BitvecCreate(pPager->dbSize);
  if( pPager->pInJournal==0 ){
    return SQLITE_NOMEM;

  }

  /* Open the journal file if it is not already open. */
  if( !isOpen(pPager->jfd) ){




    if( pPager->journalMode==PAGER_JOURNALMODE_MEMORY ){
      sqlite3MemJournalOpen(pPager->jfd);

    }else{
      const int flags =                   /* VFS flags to open journal file */
        SQLITE_OPEN_READWRITE|SQLITE_OPEN_EXCLUSIVE|SQLITE_OPEN_CREATE|
        (pPager->tempFile ? 
          (SQLITE_OPEN_DELETEONCLOSE|SQLITE_OPEN_TEMP_JOURNAL):
          (SQLITE_OPEN_MAIN_JOURNAL)
        );
#ifdef SQLITE_ENABLE_ATOMIC_WRITE
      rc = sqlite3JournalOpen(
          pVfs, pPager->zJournal, pPager->jfd, flags, jrnlBufferSize(pPager)
      );
#else
      rc = sqlite3OsOpen(pVfs, pPager->zJournal, pPager->jfd, flags, 0);
#endif
    }
    assert( rc!=SQLITE_OK || isOpen(pPager->jfd) );






  }



  /* Write the first journal header to the journal file and open 
  ** the sub-journal if necessary.
  */
  if( rc==SQLITE_OK ){
    /* TODO: Check if all of these are really required. */
    pPager->dbOrigSize = pPager->dbSize;
    pPager->journalStarted = 0;
    pPager->needSync = 0;
    pPager->nRec = 0;
    pPager->journalOff = 0;
    pPager->setMaster = 0;



    pPager->journalHdr = 0;
    rc = writeJournalHdr(pPager);
  }
  if( rc==SQLITE_OK && pPager->nSavepoint ){
    rc = openSubJournal(pPager);
  }







  if( rc!=SQLITE_OK ){

    sqlite3BitvecDestroy(pPager->pInJournal);
    pPager->pInJournal = 0;
  }
  return rc;
}

/*
** Begin a write-transaction on the specified pager object. If a 
** 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.
*/
SQLITE_PRIVATE int sqlite3PagerBegin(Pager *pPager, int exFlag){

  int rc = SQLITE_OK;

  assert( pPager->state!=PAGER_UNLOCK );
  if( pPager->state==PAGER_SHARED ){
    assert( pPager->pInJournal==0 );
    assert( !MEMDB && !pPager->tempFile );

    /* Obtain a RESERVED lock on the database file. If the exFlag parameter
    ** is true, then immediately upgrade this to an EXCLUSIVE lock. The
    ** busy-handler callback can be used when upgrading to the EXCLUSIVE
    ** lock, but not when obtaining the RESERVED lock.
    */
    rc = sqlite3OsLock(pPager->fd, RESERVED_LOCK);
    if( rc==SQLITE_OK ){
      pPager->state = PAGER_RESERVED;
      if( exFlag ){
        rc = pager_wait_on_lock(pPager, EXCLUSIVE_LOCK);
      }
    }





    /* If the required locks were successfully obtained, open the journal
    ** file and write the first journal-header to it.
    */
    if( rc==SQLITE_OK && pPager->useJournal
     && pPager->journalMode!=PAGER_JOURNALMODE_OFF 
    ){
      rc = pager_open_journal(pPager);
    }
  }else if( isOpen(pPager->jfd) && pPager->journalOff==0 ){
    /* This happens when the pager was in exclusive-access mode the last
    ** time a (read or write) transaction was successfully concluded
    ** by this connection. Instead of deleting the journal file it was 
    ** kept open and either was truncated to 0 bytes or its header was
    ** overwritten with zeros.
    */
    assert( pPager->nRec==0 );
    assert( pPager->dbOrigSize==0 );
    assert( pPager->pInJournal==0 );






    rc = pager_open_journal(pPager);
  }

  PAGERTRACE(("TRANSACTION %d\n", PAGERID(pPager)));
  assert( !isOpen(pPager->jfd) || pPager->journalOff>0 || rc!=SQLITE_OK );
  return rc;
}

/*
** Mark a single data page as writeable. The page is written into the 



** main journal or sub-journal as required. If the page is written into
** one of the journals, the corresponding bit is set in the 




** Pager.pInJournal bitvec and the PagerSavepoint.pInSavepoint bitvecs



** of any open savepoints as appropriate.
*/
static int pager_write(PgHdr *pPg){
  void *pData = pPg->pData;
  Pager *pPager = pPg->pPager;
  int rc = SQLITE_OK;

  /* Check for errors
  */
  if( pPager->errCode ){ 
    return pPager->errCode;
  }
  if( pPager->readOnly ){
    return SQLITE_PERM;
  }

  assert( !pPager->setMaster );

  CHECK_PAGE(pPg);














  /* Mark the page as dirty.  If the page has already been written
  ** to the journal then we can return right away.
  */
  sqlite3PcacheMakeDirty(pPg);
  if( pageInJournal(pPg) && !subjRequiresPage(pPg) ){

    pPager->dbModified = 1;
  }else{

    /* If we get this far, it means that the page needs to be
    ** written to the transaction journal or the ckeckpoint journal
    ** or both.
    **
    ** First check to see that the transaction journal exists and
    ** create it if it does not.
    */
    assert( pPager->state!=PAGER_UNLOCK );
    rc = sqlite3PagerBegin(pPager, 0);
    if( rc!=SQLITE_OK ){
      return rc;
    }
    assert( pPager->state>=PAGER_RESERVED );
    if( !isOpen(pPager->jfd) && pPager->useJournal
          && pPager->journalMode!=PAGER_JOURNALMODE_OFF ){
      rc = pager_open_journal(pPager);
      if( rc!=SQLITE_OK ) return rc;
    }

    pPager->dbModified = 1;
  
    /* The transaction journal now exists and we have a RESERVED or an
    ** EXCLUSIVE lock on the main database file.  Write the current page to
    ** the transaction journal if it is not there already.
    */
    if( !pageInJournal(pPg) && isOpen(pPager->jfd) ){
      if( pPg->pgno<=pPager->dbOrigSize ){
        u32 cksum;
        char *pData2;

        /* We should never write to the journal file the page that
        ** contains the database locks.  The following assert verifies
        ** that we do not. */

      pPager->dbSize++;
    }
  }
  return rc;
}

/*
** Mark a data page as writeable. This routine must be called before 
** making changes to a page. The caller must check the return value 
** of this function and be careful not to change any page data unless 
** this routine returns SQLITE_OK.
**
** The difference between this function and pager_write() is that this
** function also deals with the special case where 2 or more pages
** fit on a single disk sector. In this case all co-resident pages
** must have been written to the journal file before returning.
**
** If an error occurs, SQLITE_NOMEM or an IO error code is returned
** as appropriate. Otherwise, SQLITE_OK.
*/
SQLITE_PRIVATE int sqlite3PagerWrite(DbPage *pDbPage){
  int rc = SQLITE_OK;

  PgHdr *pPg = pDbPage;
  Pager *pPager = pPg->pPager;
  Pgno nPagePerSector = (pPager->sectorSize/pPager->pageSize);

  if( nPagePerSector>1 ){
    Pgno nPageCount;          /* Total number of pages in database file */
    Pgno pg1;                 /* First page of the sector pPg is located on. */
    int nPage;                /* Number of pages starting at pg1 to journal */
    int ii;                   /* Loop counter */
    int needSync = 0;         /* True if any page has PGHDR_NEED_SYNC */

    /* Set the doNotSync flag to 1. This is because we cannot allow a journal
    ** header to be written between the pages journaled by this function.
    */
    assert( !MEMDB );
    assert( pPager->doNotSync==0 );
    pPager->doNotSync = 1;