[BACK]Return to sqlite3.h CVS log [TXT] [DIR] Up to [ELWIX - Embedded LightWeight unIX -] / elwix / files / sqlite / dist

Diff for /elwix/files/sqlite/dist/sqlite3.h between versions 1.1 and 1.2

version 1.1, 2012/04/08 16:31:42 version 1.2, 2012/10/12 08:22:47
Line 0 Line 1
   /*
   ** 2001 September 15
   **
   ** The author disclaims copyright to this source code.  In place of
   ** a legal notice, here is a blessing:
   **
   **    May you do good and not evil.
   **    May you find forgiveness for yourself and forgive others.
   **    May you share freely, never taking more than you give.
   **
   *************************************************************************
   ** This header file defines the interface that the SQLite library
   ** presents to client programs.  If a C-function, structure, datatype,
   ** or constant definition does not appear in this file, then it is
   ** not a published API of SQLite, is subject to change without
   ** notice, and should not be referenced by programs that use SQLite.
   **
   ** Some of the definitions that are in this file are marked as
   ** "experimental".  Experimental interfaces are normally new
   ** features recently added to SQLite.  We do not anticipate changes
   ** to experimental interfaces but reserve the right to make minor changes
   ** if experience from use "in the wild" suggest such changes are prudent.
   **
   ** The official C-language API documentation for SQLite is derived
   ** from comments in this file.  This file is the authoritative source
   ** 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.
   */
   #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++.
   */
   #ifdef __cplusplus
   extern "C" {
   #endif
   
   
   /*
   ** Add the ability to override 'extern'
   */
   #ifndef SQLITE_EXTERN
   # define SQLITE_EXTERN extern
   #endif
   
   #ifndef SQLITE_API
   # define SQLITE_API
   #endif
   
   
   /*
   ** These no-op macros are used in front of interfaces to mark those
   ** interfaces as either deprecated or experimental.  New applications
   ** should not use deprecated interfaces - they are support for backwards
   ** compatibility only.  Application writers should be aware that
   ** experimental interfaces are subject to change in point releases.
   **
   ** These macros used to resolve to various kinds of compiler magic that
   ** would generate warning messages when they were used.  But that
   ** compiler magic ended up generating such a flurry of bug reports
   ** that we have taken it all out and gone back to using simple
   ** noop macros.
   */
   #define SQLITE_DEPRECATED
   #define SQLITE_EXPERIMENTAL
   
   /*
   ** Ensure these symbols were not defined by some previous header file.
   */
   #ifdef SQLITE_VERSION
   # undef SQLITE_VERSION
   #endif
   #ifdef SQLITE_VERSION_NUMBER
   # undef SQLITE_VERSION_NUMBER
   #endif
   
   /*
   ** CAPI3REF: Compile-Time Library Version Numbers
   **
   ** ^(The [SQLITE_VERSION] C preprocessor macro in the sqlite3.h header
   ** evaluates to a string literal that is the SQLite version in the
   ** format "X.Y.Z" where X is the major version number (always 3 for
   ** SQLite3) and Y is the minor version number and Z is the release number.)^
   ** ^(The [SQLITE_VERSION_NUMBER] C preprocessor macro resolves to an integer
   ** with the value (X*1000000 + Y*1000 + Z) where X, Y, and Z are the same
   ** numbers used in [SQLITE_VERSION].)^
   ** The SQLITE_VERSION_NUMBER for any given release of SQLite will also
   ** be larger than the release from which it is derived.  Either Y will
   ** be held constant and Z will be incremented or else Y will be incremented
   ** and Z will be reset to zero.
   **
   ** Since version 3.6.18, SQLite source code has been stored in the
   ** <a href="http://www.fossil-scm.org/">Fossil configuration management
   ** system</a>.  ^The SQLITE_SOURCE_ID macro evaluates to
   ** a string which identifies a particular check-in of SQLite
   ** within its configuration management system.  ^The SQLITE_SOURCE_ID
   ** string contains the date and time of the check-in (UTC) and an SHA1
   ** hash of the entire source tree.
   **
   ** See also: [sqlite3_libversion()],
   ** [sqlite3_libversion_number()], [sqlite3_sourceid()],
   ** [sqlite_version()] and [sqlite_source_id()].
   */
   #define SQLITE_VERSION        "3.7.10"
   #define SQLITE_VERSION_NUMBER 3007010
   #define SQLITE_SOURCE_ID      "2012-01-16 13:28:40 ebd01a8deffb5024a5d7494eef800d2366d97204"
   
   /*
   ** CAPI3REF: Run-Time Library Version Numbers
   ** KEYWORDS: sqlite3_version, sqlite3_sourceid
   **
   ** These interfaces provide the same information as the [SQLITE_VERSION],
   ** [SQLITE_VERSION_NUMBER], and [SQLITE_SOURCE_ID] C preprocessor macros
   ** but are associated with the library instead of the header file.  ^(Cautious
   ** programmers might include assert() statements in their application to
   ** verify that values returned by these interfaces match the macros in
   ** the header, and thus insure that the application is
   ** compiled with matching library and header files.
   **
   ** <blockquote><pre>
   ** assert( sqlite3_libversion_number()==SQLITE_VERSION_NUMBER );
   ** assert( strcmp(sqlite3_sourceid(),SQLITE_SOURCE_ID)==0 );
   ** assert( strcmp(sqlite3_libversion(),SQLITE_VERSION)==0 );
   ** </pre></blockquote>)^
   **
   ** ^The sqlite3_version[] string constant contains the text of [SQLITE_VERSION]
   ** macro.  ^The sqlite3_libversion() function returns a pointer to the
   ** to the sqlite3_version[] string constant.  The sqlite3_libversion()
   ** function is provided for use in DLLs since DLL users usually do not have
   ** direct access to string constants within the DLL.  ^The
   ** sqlite3_libversion_number() function returns an integer equal to
   ** [SQLITE_VERSION_NUMBER].  ^The sqlite3_sourceid() function returns 
   ** a pointer to a string constant whose value is the same as the 
   ** [SQLITE_SOURCE_ID] C preprocessor macro.
   **
   ** See also: [sqlite_version()] and [sqlite_source_id()].
   */
   SQLITE_API SQLITE_EXTERN const char sqlite3_version[];
   SQLITE_API const char *sqlite3_libversion(void);
   SQLITE_API const char *sqlite3_sourceid(void);
   SQLITE_API int sqlite3_libversion_number(void);
   
   /*
   ** CAPI3REF: Run-Time Library Compilation Options Diagnostics
   **
   ** ^The sqlite3_compileoption_used() function returns 0 or 1 
   ** indicating whether the specified option was defined at 
   ** compile time.  ^The SQLITE_ prefix may be omitted from the 
   ** option name passed to sqlite3_compileoption_used().  
   **
   ** ^The sqlite3_compileoption_get() function allows iterating
   ** over the list of options that were defined at compile time by
   ** returning the N-th compile time option string.  ^If N is out of range,
   ** sqlite3_compileoption_get() returns a NULL pointer.  ^The SQLITE_ 
   ** prefix is omitted from any strings returned by 
   ** sqlite3_compileoption_get().
   **
   ** ^Support for the diagnostic functions sqlite3_compileoption_used()
   ** and sqlite3_compileoption_get() may be omitted by specifying the 
   ** [SQLITE_OMIT_COMPILEOPTION_DIAGS] option at compile time.
   **
   ** See also: SQL functions [sqlite_compileoption_used()] and
   ** [sqlite_compileoption_get()] and the [compile_options pragma].
   */
   #ifndef SQLITE_OMIT_COMPILEOPTION_DIAGS
   SQLITE_API int sqlite3_compileoption_used(const char *zOptName);
   SQLITE_API const char *sqlite3_compileoption_get(int N);
   #endif
   
   /*
   ** CAPI3REF: Test To See If The Library Is Threadsafe
   **
   ** ^The sqlite3_threadsafe() function returns zero if and only if
   ** SQLite was compiled with mutexing code omitted due to the
   ** [SQLITE_THREADSAFE] compile-time option being set to 0.
   **
   ** SQLite can be compiled with or without mutexes.  When
   ** the [SQLITE_THREADSAFE] C preprocessor macro is 1 or 2, mutexes
   ** are enabled and SQLite is threadsafe.  When the
   ** [SQLITE_THREADSAFE] macro is 0, 
   ** the mutexes are omitted.  Without the mutexes, it is not safe
   ** to use SQLite concurrently from more than one thread.
   **
   ** Enabling mutexes incurs a measurable performance penalty.
   ** So if speed is of utmost importance, it makes sense to disable
   ** the mutexes.  But for maximum safety, mutexes should be enabled.
   ** ^The default behavior is for mutexes to be enabled.
   **
   ** This interface can be used by an application to make sure that the
   ** version of SQLite that it is linking against was compiled with
   ** the desired setting of the [SQLITE_THREADSAFE] macro.
   **
   ** This interface only reports on the compile-time mutex setting
   ** of the [SQLITE_THREADSAFE] flag.  If SQLite is compiled with
   ** SQLITE_THREADSAFE=1 or =2 then mutexes are enabled by default but
   ** can be fully or partially disabled using a call to [sqlite3_config()]
   ** with the verbs [SQLITE_CONFIG_SINGLETHREAD], [SQLITE_CONFIG_MULTITHREAD],
   ** or [SQLITE_CONFIG_MUTEX].  ^(The return value of the
   ** sqlite3_threadsafe() function shows only the compile-time setting of
   ** thread safety, not any run-time changes to that setting made by
   ** sqlite3_config(). In other words, the return value from sqlite3_threadsafe()
   ** is unchanged by calls to sqlite3_config().)^
   **
   ** See the [threading mode] documentation for additional information.
   */
   SQLITE_API int sqlite3_threadsafe(void);
   
   /*
   ** CAPI3REF: Database Connection Handle
   ** KEYWORDS: {database connection} {database connections}
   **
   ** Each open SQLite database is represented by a pointer to an instance of
   ** the opaque structure named "sqlite3".  It is useful to think of an sqlite3
   ** pointer as an object.  The [sqlite3_open()], [sqlite3_open16()], and
   ** [sqlite3_open_v2()] interfaces are its constructors, and [sqlite3_close()]
   ** is its destructor.  There are many other interfaces (such as
   ** [sqlite3_prepare_v2()], [sqlite3_create_function()], and
   ** [sqlite3_busy_timeout()] to name but three) that are methods on an
   ** sqlite3 object.
   */
   typedef struct sqlite3 sqlite3;
   
   /*
   ** CAPI3REF: 64-Bit Integer Types
   ** KEYWORDS: sqlite_int64 sqlite_uint64
   **
   ** Because there is no cross-platform way to specify 64-bit integer types
   ** SQLite includes typedefs for 64-bit signed and unsigned integers.
   **
   ** The sqlite3_int64 and sqlite3_uint64 are the preferred type definitions.
   ** The sqlite_int64 and sqlite_uint64 types are supported for backwards
   ** compatibility only.
   **
   ** ^The sqlite3_int64 and sqlite_int64 types can store integer values
   ** between -9223372036854775808 and +9223372036854775807 inclusive.  ^The
   ** sqlite3_uint64 and sqlite_uint64 types can store integer values 
   ** between 0 and +18446744073709551615 inclusive.
   */
   #ifdef SQLITE_INT64_TYPE
     typedef SQLITE_INT64_TYPE sqlite_int64;
     typedef unsigned SQLITE_INT64_TYPE sqlite_uint64;
   #elif defined(_MSC_VER) || defined(__BORLANDC__)
     typedef __int64 sqlite_int64;
     typedef unsigned __int64 sqlite_uint64;
   #else
     typedef long long int sqlite_int64;
     typedef unsigned long long int sqlite_uint64;
   #endif
   typedef sqlite_int64 sqlite3_int64;
   typedef sqlite_uint64 sqlite3_uint64;
   
   /*
   ** If compiling for a processor that lacks floating point support,
   ** substitute integer for floating-point.
   */
   #ifdef SQLITE_OMIT_FLOATING_POINT
   # define double sqlite3_int64
   #endif
   
   /*
   ** CAPI3REF: Closing A Database Connection
   **
   ** ^The sqlite3_close() routine is the destructor for the [sqlite3] object.
   ** ^Calls to sqlite3_close() return SQLITE_OK if the [sqlite3] object is
   ** successfully destroyed and all associated resources are deallocated.
   **
   ** Applications must [sqlite3_finalize | finalize] all [prepared statements]
   ** and [sqlite3_blob_close | close] all [BLOB handles] associated with
   ** the [sqlite3] object prior to attempting to close the object.  ^If
   ** sqlite3_close() is called on a [database connection] that still has
   ** outstanding [prepared statements] or [BLOB handles], then it returns
   ** SQLITE_BUSY.
   **
   ** ^If [sqlite3_close()] is invoked while a transaction is open,
   ** the transaction is automatically rolled back.
   **
   ** The C parameter to [sqlite3_close(C)] must be either a NULL
   ** pointer or an [sqlite3] object pointer obtained
   ** from [sqlite3_open()], [sqlite3_open16()], or
   ** [sqlite3_open_v2()], and not previously closed.
   ** ^Calling sqlite3_close() with a NULL pointer argument is a 
   ** harmless no-op.
   */
   SQLITE_API int sqlite3_close(sqlite3 *);
   
   /*
   ** The type for a callback function.
   ** This is legacy and deprecated.  It is included for historical
   ** compatibility and is not documented.
   */
   typedef int (*sqlite3_callback)(void*,int,char**, char**);
   
   /*
   ** CAPI3REF: One-Step Query Execution Interface
   **
   ** The sqlite3_exec() interface is a convenience wrapper around
   ** [sqlite3_prepare_v2()], [sqlite3_step()], and [sqlite3_finalize()],
   ** that allows an application to run multiple statements of SQL
   ** without having to use a lot of C code. 
   **
   ** ^The sqlite3_exec() interface runs zero or more UTF-8 encoded,
   ** semicolon-separate SQL statements passed into its 2nd argument,
   ** in the context of the [database connection] passed in as its 1st
   ** argument.  ^If the callback function of the 3rd argument to
   ** sqlite3_exec() is not NULL, then it is invoked for each result row
   ** coming out of the evaluated SQL statements.  ^The 4th argument to
   ** sqlite3_exec() is relayed through to the 1st argument of each
   ** callback invocation.  ^If the callback pointer to sqlite3_exec()
   ** is NULL, then no callback is ever invoked and result rows are
   ** ignored.
   **
   ** ^If an error occurs while evaluating the SQL statements passed into
   ** sqlite3_exec(), then execution of the current statement stops and
   ** subsequent statements are skipped.  ^If the 5th parameter to sqlite3_exec()
   ** is not NULL then any error message is written into memory obtained
   ** from [sqlite3_malloc()] and passed back through the 5th parameter.
   ** To avoid memory leaks, the application should invoke [sqlite3_free()]
   ** on error message strings returned through the 5th parameter of
   ** of sqlite3_exec() after the error message string is no longer needed.
   ** ^If the 5th parameter to sqlite3_exec() is not NULL and no errors
   ** occur, then sqlite3_exec() sets the pointer in its 5th parameter to
   ** NULL before returning.
   **
   ** ^If an sqlite3_exec() callback returns non-zero, the sqlite3_exec()
   ** routine returns SQLITE_ABORT without invoking the callback again and
   ** without running any subsequent SQL statements.
   **
   ** ^The 2nd argument to the sqlite3_exec() callback function is the
   ** number of columns in the result.  ^The 3rd argument to the sqlite3_exec()
   ** callback is an array of pointers to strings obtained as if from
   ** [sqlite3_column_text()], one for each column.  ^If an element of a
   ** result row is NULL then the corresponding string pointer for the
   ** sqlite3_exec() callback is a NULL pointer.  ^The 4th argument to the
   ** sqlite3_exec() callback is an array of pointers to strings where each
   ** entry represents the name of corresponding result column as obtained
   ** from [sqlite3_column_name()].
   **
   ** ^If the 2nd parameter to sqlite3_exec() is a NULL pointer, a pointer
   ** to an empty string, or a pointer that contains only whitespace and/or 
   ** SQL comments, then no SQL statements are evaluated and the database
   ** is not changed.
   **
   ** Restrictions:
   **
   ** <ul>
   ** <li> The application must insure that the 1st parameter to sqlite3_exec()
   **      is a valid and open [database connection].
   ** <li> The application must not close [database connection] specified by
   **      the 1st parameter to sqlite3_exec() while sqlite3_exec() is running.
   ** <li> The application must not modify the SQL statement text passed into
   **      the 2nd parameter of sqlite3_exec() while sqlite3_exec() is running.
   ** </ul>
   */
   SQLITE_API int sqlite3_exec(
     sqlite3*,                                  /* An open database */
     const char *sql,                           /* SQL to be evaluated */
     int (*callback)(void*,int,char**,char**),  /* Callback function */
     void *,                                    /* 1st argument to callback */
     char **errmsg                              /* Error msg written here */
   );
   
   /*
   ** CAPI3REF: Result Codes
   ** KEYWORDS: SQLITE_OK {error code} {error codes}
   ** KEYWORDS: {result code} {result codes}
   **
   ** Many SQLite functions return an integer result code from the set shown
   ** here in order to indicate success or failure.
   **
   ** New error codes may be added in future versions of SQLite.
   **
   ** See also: [SQLITE_IOERR_READ | extended result codes],
   ** [sqlite3_vtab_on_conflict()] [SQLITE_ROLLBACK | result codes].
   */
   #define SQLITE_OK           0   /* Successful result */
   /* beginning-of-error-codes */
   #define SQLITE_ERROR        1   /* SQL error or missing database */
   #define SQLITE_INTERNAL     2   /* Internal logic error in SQLite */
   #define SQLITE_PERM         3   /* Access permission denied */
   #define SQLITE_ABORT        4   /* Callback routine requested an abort */
   #define SQLITE_BUSY         5   /* The database file is locked */
   #define SQLITE_LOCKED       6   /* A table in the database is locked */
   #define SQLITE_NOMEM        7   /* A malloc() failed */
   #define SQLITE_READONLY     8   /* Attempt to write a readonly database */
   #define SQLITE_INTERRUPT    9   /* Operation terminated by sqlite3_interrupt()*/
   #define SQLITE_IOERR       10   /* Some kind of disk I/O error occurred */
   #define SQLITE_CORRUPT     11   /* The database disk image is malformed */
   #define SQLITE_NOTFOUND    12   /* Unknown opcode in sqlite3_file_control() */
   #define SQLITE_FULL        13   /* Insertion failed because database is full */
   #define SQLITE_CANTOPEN    14   /* Unable to open the database file */
   #define SQLITE_PROTOCOL    15   /* Database lock protocol error */
   #define SQLITE_EMPTY       16   /* Database is empty */
   #define SQLITE_SCHEMA      17   /* The database schema changed */
   #define SQLITE_TOOBIG      18   /* String or BLOB exceeds size limit */
   #define SQLITE_CONSTRAINT  19   /* Abort due to constraint violation */
   #define SQLITE_MISMATCH    20   /* Data type mismatch */
   #define SQLITE_MISUSE      21   /* Library used incorrectly */
   #define SQLITE_NOLFS       22   /* Uses OS features not supported on host */
   #define SQLITE_AUTH        23   /* Authorization denied */
   #define SQLITE_FORMAT      24   /* Auxiliary database format error */
   #define SQLITE_RANGE       25   /* 2nd parameter to sqlite3_bind out of range */
   #define SQLITE_NOTADB      26   /* File opened that is not a database file */
   #define SQLITE_ROW         100  /* sqlite3_step() has another row ready */
   #define SQLITE_DONE        101  /* sqlite3_step() has finished executing */
   /* end-of-error-codes */
   
   /*
   ** CAPI3REF: Extended Result Codes
   ** KEYWORDS: {extended error code} {extended error codes}
   ** KEYWORDS: {extended result code} {extended result codes}
   **
   ** In its default configuration, SQLite API routines return one of 26 integer
   ** [SQLITE_OK | result codes].  However, experience has shown that many of
   ** these result codes are too coarse-grained.  They do not provide as
   ** much information about problems as programmers might like.  In an effort to
   ** address this, newer versions of SQLite (version 3.3.8 and later) include
   ** support for additional result codes that provide more detailed information
   ** about errors. The extended result codes are enabled or disabled
   ** on a per database connection basis using the
   ** [sqlite3_extended_result_codes()] API.
   **
   ** Some of the available extended result codes are listed here.
   ** One may expect the number of extended result codes will be expand
   ** over time.  Software that uses extended result codes should expect
   ** to see new result codes in future releases of SQLite.
   **
   ** The SQLITE_OK result code will never be extended.  It will always
   ** be exactly zero.
   */
   #define SQLITE_IOERR_READ              (SQLITE_IOERR | (1<<8))
   #define SQLITE_IOERR_SHORT_READ        (SQLITE_IOERR | (2<<8))
   #define SQLITE_IOERR_WRITE             (SQLITE_IOERR | (3<<8))
   #define SQLITE_IOERR_FSYNC             (SQLITE_IOERR | (4<<8))
   #define SQLITE_IOERR_DIR_FSYNC         (SQLITE_IOERR | (5<<8))
   #define SQLITE_IOERR_TRUNCATE          (SQLITE_IOERR | (6<<8))
   #define SQLITE_IOERR_FSTAT             (SQLITE_IOERR | (7<<8))
   #define SQLITE_IOERR_UNLOCK            (SQLITE_IOERR | (8<<8))
   #define SQLITE_IOERR_RDLOCK            (SQLITE_IOERR | (9<<8))
   #define SQLITE_IOERR_DELETE            (SQLITE_IOERR | (10<<8))
   #define SQLITE_IOERR_BLOCKED           (SQLITE_IOERR | (11<<8))
   #define SQLITE_IOERR_NOMEM             (SQLITE_IOERR | (12<<8))
   #define SQLITE_IOERR_ACCESS            (SQLITE_IOERR | (13<<8))
   #define SQLITE_IOERR_CHECKRESERVEDLOCK (SQLITE_IOERR | (14<<8))
   #define SQLITE_IOERR_LOCK              (SQLITE_IOERR | (15<<8))
   #define SQLITE_IOERR_CLOSE             (SQLITE_IOERR | (16<<8))
   #define SQLITE_IOERR_DIR_CLOSE         (SQLITE_IOERR | (17<<8))
   #define SQLITE_IOERR_SHMOPEN           (SQLITE_IOERR | (18<<8))
   #define SQLITE_IOERR_SHMSIZE           (SQLITE_IOERR | (19<<8))
   #define SQLITE_IOERR_SHMLOCK           (SQLITE_IOERR | (20<<8))
   #define SQLITE_IOERR_SHMMAP            (SQLITE_IOERR | (21<<8))
   #define SQLITE_IOERR_SEEK              (SQLITE_IOERR | (22<<8))
   #define SQLITE_LOCKED_SHAREDCACHE      (SQLITE_LOCKED |  (1<<8))
   #define SQLITE_BUSY_RECOVERY           (SQLITE_BUSY   |  (1<<8))
   #define SQLITE_CANTOPEN_NOTEMPDIR      (SQLITE_CANTOPEN | (1<<8))
   #define SQLITE_CORRUPT_VTAB            (SQLITE_CORRUPT | (1<<8))
   #define SQLITE_READONLY_RECOVERY       (SQLITE_READONLY | (1<<8))
   #define SQLITE_READONLY_CANTLOCK       (SQLITE_READONLY | (2<<8))
   
   /*
   ** CAPI3REF: Flags For File Open Operations
   **
   ** These bit values are intended for use in the
   ** 3rd parameter to the [sqlite3_open_v2()] interface and
   ** in the 4th parameter to the [sqlite3_vfs.xOpen] method.
   */
   #define SQLITE_OPEN_READONLY         0x00000001  /* Ok for sqlite3_open_v2() */
   #define SQLITE_OPEN_READWRITE        0x00000002  /* Ok for sqlite3_open_v2() */
   #define SQLITE_OPEN_CREATE           0x00000004  /* Ok for sqlite3_open_v2() */
   #define SQLITE_OPEN_DELETEONCLOSE    0x00000008  /* VFS only */
   #define SQLITE_OPEN_EXCLUSIVE        0x00000010  /* VFS only */
   #define SQLITE_OPEN_AUTOPROXY        0x00000020  /* VFS only */
   #define SQLITE_OPEN_URI              0x00000040  /* Ok for sqlite3_open_v2() */
   #define SQLITE_OPEN_MAIN_DB          0x00000100  /* VFS only */
   #define SQLITE_OPEN_TEMP_DB          0x00000200  /* VFS only */
   #define SQLITE_OPEN_TRANSIENT_DB     0x00000400  /* VFS only */
   #define SQLITE_OPEN_MAIN_JOURNAL     0x00000800  /* VFS only */
   #define SQLITE_OPEN_TEMP_JOURNAL     0x00001000  /* VFS only */
   #define SQLITE_OPEN_SUBJOURNAL       0x00002000  /* VFS only */
   #define SQLITE_OPEN_MASTER_JOURNAL   0x00004000  /* VFS only */
   #define SQLITE_OPEN_NOMUTEX          0x00008000  /* Ok for sqlite3_open_v2() */
   #define SQLITE_OPEN_FULLMUTEX        0x00010000  /* Ok for sqlite3_open_v2() */
   #define SQLITE_OPEN_SHAREDCACHE      0x00020000  /* Ok for sqlite3_open_v2() */
   #define SQLITE_OPEN_PRIVATECACHE     0x00040000  /* Ok for sqlite3_open_v2() */
   #define SQLITE_OPEN_WAL              0x00080000  /* VFS only */
   
   /* Reserved:                         0x00F00000 */
   
   /*
   ** CAPI3REF: Device Characteristics
   **
   ** The xDeviceCharacteristics method of the [sqlite3_io_methods]
   ** object returns an integer which is a vector of the these
   ** bit values expressing I/O characteristics of the mass storage
   ** device that holds the file that the [sqlite3_io_methods]
   ** refers to.
   **
   ** The SQLITE_IOCAP_ATOMIC property means that all writes of
   ** any size are atomic.  The SQLITE_IOCAP_ATOMICnnn values
   ** mean that writes of blocks that are nnn bytes in size and
   ** are aligned to an address which is an integer multiple of
   ** nnn are atomic.  The SQLITE_IOCAP_SAFE_APPEND value means
   ** that when data is appended to a file, the data is appended
   ** first then the size of the file is extended, never the other
   ** way around.  The SQLITE_IOCAP_SEQUENTIAL property means that
   ** information is written to disk in the same order as calls
   ** to xWrite().  The SQLITE_IOCAP_POWERSAFE_OVERWRITE property means that
   ** after reboot following a crash or power loss, the only bytes in a
   ** file that were written at the application level might have changed
   ** and that adjacent bytes, even bytes within the same sector are
   ** guaranteed to be unchanged.
   */
   #define SQLITE_IOCAP_ATOMIC                 0x00000001
   #define SQLITE_IOCAP_ATOMIC512              0x00000002
   #define SQLITE_IOCAP_ATOMIC1K               0x00000004
   #define SQLITE_IOCAP_ATOMIC2K               0x00000008
   #define SQLITE_IOCAP_ATOMIC4K               0x00000010
   #define SQLITE_IOCAP_ATOMIC8K               0x00000020
   #define SQLITE_IOCAP_ATOMIC16K              0x00000040
   #define SQLITE_IOCAP_ATOMIC32K              0x00000080
   #define SQLITE_IOCAP_ATOMIC64K              0x00000100
   #define SQLITE_IOCAP_SAFE_APPEND            0x00000200
   #define SQLITE_IOCAP_SEQUENTIAL             0x00000400
   #define SQLITE_IOCAP_UNDELETABLE_WHEN_OPEN  0x00000800
   #define SQLITE_IOCAP_POWERSAFE_OVERWRITE    0x00001000
   
   /*
   ** CAPI3REF: File Locking Levels
   **
   ** SQLite uses one of these integer values as the second
   ** argument to calls it makes to the xLock() and xUnlock() methods
   ** of an [sqlite3_io_methods] object.
   */
   #define SQLITE_LOCK_NONE          0
   #define SQLITE_LOCK_SHARED        1
   #define SQLITE_LOCK_RESERVED      2
   #define SQLITE_LOCK_PENDING       3
   #define SQLITE_LOCK_EXCLUSIVE     4
   
   /*
   ** CAPI3REF: Synchronization Type Flags
   **
   ** When SQLite invokes the xSync() method of an
   ** [sqlite3_io_methods] object it uses a combination of
   ** these integer values as the second argument.
   **
   ** When the SQLITE_SYNC_DATAONLY flag is used, it means that the
   ** sync operation only needs to flush data to mass storage.  Inode
   ** information need not be flushed. If the lower four bits of the flag
   ** equal SQLITE_SYNC_NORMAL, that means to use normal fsync() semantics.
   ** If the lower four bits equal SQLITE_SYNC_FULL, that means
   ** to use Mac OS X style fullsync instead of fsync().
   **
   ** Do not confuse the SQLITE_SYNC_NORMAL and SQLITE_SYNC_FULL flags
   ** with the [PRAGMA synchronous]=NORMAL and [PRAGMA synchronous]=FULL
   ** settings.  The [synchronous pragma] determines when calls to the
   ** xSync VFS method occur and applies uniformly across all platforms.
   ** The SQLITE_SYNC_NORMAL and SQLITE_SYNC_FULL flags determine how
   ** energetic or rigorous or forceful the sync operations are and
   ** only make a difference on Mac OSX for the default SQLite code.
   ** (Third-party VFS implementations might also make the distinction
   ** between SQLITE_SYNC_NORMAL and SQLITE_SYNC_FULL, but among the
   ** operating systems natively supported by SQLite, only Mac OSX
   ** cares about the difference.)
   */
   #define SQLITE_SYNC_NORMAL        0x00002
   #define SQLITE_SYNC_FULL          0x00003
   #define SQLITE_SYNC_DATAONLY      0x00010
   
   /*
   ** CAPI3REF: OS Interface Open File Handle
   **
   ** An [sqlite3_file] object represents an open file in the 
   ** [sqlite3_vfs | OS interface layer].  Individual OS interface
   ** implementations will
   ** want to subclass this object by appending additional fields
   ** for their own use.  The pMethods entry is a pointer to an
   ** [sqlite3_io_methods] object that defines methods for performing
   ** I/O operations on the open file.
   */
   typedef struct sqlite3_file sqlite3_file;
   struct sqlite3_file {
     const struct sqlite3_io_methods *pMethods;  /* Methods for an open file */
   };
   
   /*
   ** CAPI3REF: OS Interface File Virtual Methods Object
   **
   ** Every file opened by the [sqlite3_vfs.xOpen] method populates an
   ** [sqlite3_file] object (or, more commonly, a subclass of the
   ** [sqlite3_file] object) with a pointer to an instance of this object.
   ** This object defines the methods used to perform various operations
   ** against the open file represented by the [sqlite3_file] object.
   **
   ** If the [sqlite3_vfs.xOpen] method sets the sqlite3_file.pMethods element 
   ** to a non-NULL pointer, then the sqlite3_io_methods.xClose method
   ** may be invoked even if the [sqlite3_vfs.xOpen] reported that it failed.  The
   ** only way to prevent a call to xClose following a failed [sqlite3_vfs.xOpen]
   ** is for the [sqlite3_vfs.xOpen] to set the sqlite3_file.pMethods element
   ** to NULL.
   **
   ** The flags argument to xSync may be one of [SQLITE_SYNC_NORMAL] or
   ** [SQLITE_SYNC_FULL].  The first choice is the normal fsync().
   ** The second choice is a Mac OS X style fullsync.  The [SQLITE_SYNC_DATAONLY]
   ** flag may be ORed in to indicate that only the data of the file
   ** and not its inode needs to be synced.
   **
   ** The integer values to xLock() and xUnlock() are one of
   ** <ul>
   ** <li> [SQLITE_LOCK_NONE],
   ** <li> [SQLITE_LOCK_SHARED],
   ** <li> [SQLITE_LOCK_RESERVED],
   ** <li> [SQLITE_LOCK_PENDING], or
   ** <li> [SQLITE_LOCK_EXCLUSIVE].
   ** </ul>
   ** xLock() increases the lock. xUnlock() decreases the lock.
   ** The xCheckReservedLock() method checks whether any database connection,
   ** either in this process or in some other process, is holding a RESERVED,
   ** PENDING, or EXCLUSIVE lock on the file.  It returns true
   ** if such a lock exists and false otherwise.
   **
   ** The xFileControl() method is a generic interface that allows custom
   ** VFS implementations to directly control an open file using the
   ** [sqlite3_file_control()] interface.  The second "op" argument is an
   ** integer opcode.  The third argument is a generic pointer intended to
   ** point to a structure that may contain arguments or space in which to
   ** write return values.  Potential uses for xFileControl() might be
   ** functions to enable blocking locks with timeouts, to change the
   ** locking strategy (for example to use dot-file locks), to inquire
   ** about the status of a lock, or to break stale locks.  The SQLite
   ** core reserves all opcodes less than 100 for its own use.
   ** A [SQLITE_FCNTL_LOCKSTATE | list of opcodes] less than 100 is available.
   ** Applications that define a custom xFileControl method should use opcodes
   ** greater than 100 to avoid conflicts.  VFS implementations should
   ** return [SQLITE_NOTFOUND] for file control opcodes that they do not
   ** recognize.
   **
   ** The xSectorSize() method returns the sector size of the
   ** device that underlies the file.  The sector size is the
   ** minimum write that can be performed without disturbing
   ** other bytes in the file.  The xDeviceCharacteristics()
   ** method returns a bit vector describing behaviors of the
   ** underlying device:
   **
   ** <ul>
   ** <li> [SQLITE_IOCAP_ATOMIC]
   ** <li> [SQLITE_IOCAP_ATOMIC512]
   ** <li> [SQLITE_IOCAP_ATOMIC1K]
   ** <li> [SQLITE_IOCAP_ATOMIC2K]
   ** <li> [SQLITE_IOCAP_ATOMIC4K]
   ** <li> [SQLITE_IOCAP_ATOMIC8K]
   ** <li> [SQLITE_IOCAP_ATOMIC16K]
   ** <li> [SQLITE_IOCAP_ATOMIC32K]
   ** <li> [SQLITE_IOCAP_ATOMIC64K]
   ** <li> [SQLITE_IOCAP_SAFE_APPEND]
   ** <li> [SQLITE_IOCAP_SEQUENTIAL]
   ** </ul>
   **
   ** The SQLITE_IOCAP_ATOMIC property means that all writes of
   ** any size are atomic.  The SQLITE_IOCAP_ATOMICnnn values
   ** mean that writes of blocks that are nnn bytes in size and
   ** are aligned to an address which is an integer multiple of
   ** nnn are atomic.  The SQLITE_IOCAP_SAFE_APPEND value means
   ** that when data is appended to a file, the data is appended
   ** first then the size of the file is extended, never the other
   ** way around.  The SQLITE_IOCAP_SEQUENTIAL property means that
   ** information is written to disk in the same order as calls
   ** to xWrite().
   **
   ** If xRead() returns SQLITE_IOERR_SHORT_READ it must also fill
   ** in the unread portions of the buffer with zeros.  A VFS that
   ** fails to zero-fill short reads might seem to work.  However,
   ** failure to zero-fill short reads will eventually lead to
   ** database corruption.
   */
   typedef struct sqlite3_io_methods sqlite3_io_methods;
   struct sqlite3_io_methods {
     int iVersion;
     int (*xClose)(sqlite3_file*);
     int (*xRead)(sqlite3_file*, void*, int iAmt, sqlite3_int64 iOfst);
     int (*xWrite)(sqlite3_file*, const void*, int iAmt, sqlite3_int64 iOfst);
     int (*xTruncate)(sqlite3_file*, sqlite3_int64 size);
     int (*xSync)(sqlite3_file*, int flags);
     int (*xFileSize)(sqlite3_file*, sqlite3_int64 *pSize);
     int (*xLock)(sqlite3_file*, int);
     int (*xUnlock)(sqlite3_file*, int);
     int (*xCheckReservedLock)(sqlite3_file*, int *pResOut);
     int (*xFileControl)(sqlite3_file*, int op, void *pArg);
     int (*xSectorSize)(sqlite3_file*);
     int (*xDeviceCharacteristics)(sqlite3_file*);
     /* Methods above are valid for version 1 */
     int (*xShmMap)(sqlite3_file*, int iPg, int pgsz, int, void volatile**);
     int (*xShmLock)(sqlite3_file*, int offset, int n, int flags);
     void (*xShmBarrier)(sqlite3_file*);
     int (*xShmUnmap)(sqlite3_file*, int deleteFlag);
     /* Methods above are valid for version 2 */
     /* Additional methods may be added in future releases */
   };
   
   /*
   ** CAPI3REF: Standard File Control Opcodes
   **
   ** These integer constants are opcodes for the xFileControl method
   ** of the [sqlite3_io_methods] object and for the [sqlite3_file_control()]
   ** interface.
   **
   ** The [SQLITE_FCNTL_LOCKSTATE] opcode is used for debugging.  This
   ** opcode causes the xFileControl method to write the current state of
   ** the lock (one of [SQLITE_LOCK_NONE], [SQLITE_LOCK_SHARED],
   ** [SQLITE_LOCK_RESERVED], [SQLITE_LOCK_PENDING], or [SQLITE_LOCK_EXCLUSIVE])
   ** into an integer that the pArg argument points to. This capability
   ** is used during testing and only needs to be supported when SQLITE_TEST
   ** is defined.
   **
   ** The [SQLITE_FCNTL_SIZE_HINT] opcode is used by SQLite to give the VFS
   ** layer a hint of how large the database file will grow to be during the
   ** current transaction.  This hint is not guaranteed to be accurate but it
   ** is often close.  The underlying VFS might choose to preallocate database
   ** file space based on this hint in order to help writes to the database
   ** file run faster.
   **
   ** The [SQLITE_FCNTL_CHUNK_SIZE] opcode is used to request that the VFS
   ** extends and truncates the database file in chunks of a size specified
   ** by the user. The fourth argument to [sqlite3_file_control()] should 
   ** point to an integer (type int) containing the new chunk-size to use
   ** for the nominated database. Allocating database file space in large
   ** chunks (say 1MB at a time), may reduce file-system fragmentation and
   ** improve performance on some systems.
   **
   ** The [SQLITE_FCNTL_FILE_POINTER] opcode is used to obtain a pointer
   ** to the [sqlite3_file] object associated with a particular database
   ** connection.  See the [sqlite3_file_control()] documentation for
   ** additional information.
   **
   ** ^(The [SQLITE_FCNTL_SYNC_OMITTED] opcode is generated internally by
   ** SQLite and sent to all VFSes in place of a call to the xSync method
   ** when the database connection has [PRAGMA synchronous] set to OFF.)^
   ** Some specialized VFSes need this signal in order to operate correctly
   ** when [PRAGMA synchronous | PRAGMA synchronous=OFF] is set, but most 
   ** VFSes do not need this signal and should silently ignore this opcode.
   ** Applications should not call [sqlite3_file_control()] with this
   ** opcode as doing so may disrupt the operation of the specialized VFSes
   ** that do require it.  
   **
   ** ^The [SQLITE_FCNTL_WIN32_AV_RETRY] opcode is used to configure automatic
   ** retry counts and intervals for certain disk I/O operations for the
   ** windows [VFS] in order to provide robustness in the presence of
   ** anti-virus programs.  By default, the windows VFS will retry file read,
   ** file write, and file delete operations up to 10 times, with a delay
   ** of 25 milliseconds before the first retry and with the delay increasing
   ** by an additional 25 milliseconds with each subsequent retry.  This
   ** opcode allows these two values (10 retries and 25 milliseconds of delay)
   ** to be adjusted.  The values are changed for all database connections
   ** within the same process.  The argument is a pointer to an array of two
   ** integers where the first integer i the new retry count and the second
   ** integer is the delay.  If either integer is negative, then the setting
   ** is not changed but instead the prior value of that setting is written
   ** into the array entry, allowing the current retry settings to be
   ** interrogated.  The zDbName parameter is ignored.
   **
   ** ^The [SQLITE_FCNTL_PERSIST_WAL] opcode is used to set or query the
   ** persistent [WAL | Write AHead Log] setting.  By default, the auxiliary
   ** write ahead log and shared memory files used for transaction control
   ** are automatically deleted when the latest connection to the database
   ** closes.  Setting persistent WAL mode causes those files to persist after
   ** close.  Persisting the files is useful when other processes that do not
   ** have write permission on the directory containing the database file want
   ** to read the database file, as the WAL and shared memory files must exist
   ** in order for the database to be readable.  The fourth parameter to
   ** [sqlite3_file_control()] for this opcode should be a pointer to an integer.
   ** That integer is 0 to disable persistent WAL mode or 1 to enable persistent
   ** WAL mode.  If the integer is -1, then it is overwritten with the current
   ** WAL persistence setting.
   **
   ** ^The [SQLITE_FCNTL_POWERSAFE_OVERWRITE] opcode is used to set or query the
   ** persistent "powersafe-overwrite" or "PSOW" setting.  The PSOW setting
   ** determines the [SQLITE_IOCAP_POWERSAFE_OVERWRITE] bit of the
   ** xDeviceCharacteristics methods. The fourth parameter to
   ** [sqlite3_file_control()] for this opcode should be a pointer to an integer.
   ** That integer is 0 to disable zero-damage mode or 1 to enable zero-damage
   ** mode.  If the integer is -1, then it is overwritten with the current
   ** zero-damage mode setting.
   **
   ** ^The [SQLITE_FCNTL_OVERWRITE] opcode is invoked by SQLite after opening
   ** a write transaction to indicate that, unless it is rolled back for some
   ** reason, the entire database file will be overwritten by the current 
   ** transaction. This is used by VACUUM operations.
   **
   ** ^The [SQLITE_FCNTL_VFSNAME] opcode can be used to obtain the names of
   ** all [VFSes] in the VFS stack.  The names are of all VFS shims and the
   ** final bottom-level VFS are written into memory obtained from 
   ** [sqlite3_malloc()] and the result is stored in the char* variable
   ** that the fourth parameter of [sqlite3_file_control()] points to.
   ** The caller is responsible for freeing the memory when done.  As with
   ** all file-control actions, there is no guarantee that this will actually
   ** do anything.  Callers should initialize the char* variable to a NULL
   ** pointer in case this file-control is not implemented.  This file-control
   ** is intended for diagnostic use only.
   */
   #define SQLITE_FCNTL_LOCKSTATE               1
   #define SQLITE_GET_LOCKPROXYFILE             2
   #define SQLITE_SET_LOCKPROXYFILE             3
   #define SQLITE_LAST_ERRNO                    4
   #define SQLITE_FCNTL_SIZE_HINT               5
   #define SQLITE_FCNTL_CHUNK_SIZE              6
   #define SQLITE_FCNTL_FILE_POINTER            7
   #define SQLITE_FCNTL_SYNC_OMITTED            8
   #define SQLITE_FCNTL_WIN32_AV_RETRY          9
   #define SQLITE_FCNTL_PERSIST_WAL            10
   #define SQLITE_FCNTL_OVERWRITE              11
   #define SQLITE_FCNTL_VFSNAME                12
   #define SQLITE_FCNTL_POWERSAFE_OVERWRITE    13
   
   /*
   ** CAPI3REF: Mutex Handle
   **
   ** The mutex module within SQLite defines [sqlite3_mutex] to be an
   ** abstract type for a mutex object.  The SQLite core never looks
   ** at the internal representation of an [sqlite3_mutex].  It only
   ** deals with pointers to the [sqlite3_mutex] object.
   **
   ** Mutexes are created using [sqlite3_mutex_alloc()].
   */
   typedef struct sqlite3_mutex sqlite3_mutex;
   
   /*
   ** CAPI3REF: OS Interface Object
   **
   ** An instance of the sqlite3_vfs object defines the interface between
   ** the SQLite core and the underlying operating system.  The "vfs"
   ** in the name of the object stands for "virtual file system".  See
   ** the [VFS | VFS documentation] for further information.
   **
   ** The value of the iVersion field is initially 1 but may be larger in
   ** future versions of SQLite.  Additional fields may be appended to this
   ** object when the iVersion value is increased.  Note that the structure
   ** of the sqlite3_vfs object changes in the transaction between
   ** SQLite version 3.5.9 and 3.6.0 and yet the iVersion field was not
   ** modified.
   **
   ** The szOsFile field is the size of the subclassed [sqlite3_file]
   ** structure used by this VFS.  mxPathname is the maximum length of
   ** a pathname in this VFS.
   **
   ** Registered sqlite3_vfs objects are kept on a linked list formed by
   ** the pNext pointer.  The [sqlite3_vfs_register()]
   ** and [sqlite3_vfs_unregister()] interfaces manage this list
   ** in a thread-safe way.  The [sqlite3_vfs_find()] interface
   ** searches the list.  Neither the application code nor the VFS
   ** implementation should use the pNext pointer.
   **
   ** The pNext field is the only field in the sqlite3_vfs
   ** structure that SQLite will ever modify.  SQLite will only access
   ** or modify this field while holding a particular static mutex.
   ** The application should never modify anything within the sqlite3_vfs
   ** object once the object has been registered.
   **
   ** The zName field holds the name of the VFS module.  The name must
   ** be unique across all VFS modules.
   **
   ** [[sqlite3_vfs.xOpen]]
   ** ^SQLite guarantees that the zFilename parameter to xOpen
   ** is either a NULL pointer or string obtained
   ** from xFullPathname() with an optional suffix added.
   ** ^If a suffix is added to the zFilename parameter, it will
   ** consist of a single "-" character followed by no more than
   ** 11 alphanumeric and/or "-" characters.
   ** ^SQLite further guarantees that
   ** the string will be valid and unchanged until xClose() is
   ** called. Because of the previous sentence,
   ** the [sqlite3_file] can safely store a pointer to the
   ** filename if it needs to remember the filename for some reason.
   ** If the zFilename parameter to xOpen is a NULL pointer then xOpen
   ** must invent its own temporary name for the file.  ^Whenever the 
   ** xFilename parameter is NULL it will also be the case that the
   ** flags parameter will include [SQLITE_OPEN_DELETEONCLOSE].
   **
   ** The flags argument to xOpen() includes all bits set in
   ** the flags argument to [sqlite3_open_v2()].  Or if [sqlite3_open()]
   ** or [sqlite3_open16()] is used, then flags includes at least
   ** [SQLITE_OPEN_READWRITE] | [SQLITE_OPEN_CREATE]. 
   ** If xOpen() opens a file read-only then it sets *pOutFlags to
   ** include [SQLITE_OPEN_READONLY].  Other bits in *pOutFlags may be set.
   **
   ** ^(SQLite will also add one of the following flags to the xOpen()
   ** call, depending on the object being opened:
   **
   ** <ul>
   ** <li>  [SQLITE_OPEN_MAIN_DB]
   ** <li>  [SQLITE_OPEN_MAIN_JOURNAL]
   ** <li>  [SQLITE_OPEN_TEMP_DB]
   ** <li>  [SQLITE_OPEN_TEMP_JOURNAL]
   ** <li>  [SQLITE_OPEN_TRANSIENT_DB]
   ** <li>  [SQLITE_OPEN_SUBJOURNAL]
   ** <li>  [SQLITE_OPEN_MASTER_JOURNAL]
   ** <li>  [SQLITE_OPEN_WAL]
   ** </ul>)^
   **
   ** The file I/O implementation can use the object type flags to
   ** change the way it deals with files.  For example, an application
   ** that does not care about crash recovery or rollback might make
   ** the open of a journal file a no-op.  Writes to this journal would
   ** also be no-ops, and any attempt to read the journal would return
   ** SQLITE_IOERR.  Or the implementation might recognize that a database
   ** file will be doing page-aligned sector reads and writes in a random
   ** order and set up its I/O subsystem accordingly.
   **
   ** SQLite might also add one of the following flags to the xOpen method:
   **
   ** <ul>
   ** <li> [SQLITE_OPEN_DELETEONCLOSE]
   ** <li> [SQLITE_OPEN_EXCLUSIVE]
   ** </ul>
   **
   ** The [SQLITE_OPEN_DELETEONCLOSE] flag means the file should be
   ** deleted when it is closed.  ^The [SQLITE_OPEN_DELETEONCLOSE]
   ** will be set for TEMP databases and their journals, transient
   ** databases, and subjournals.
   **
   ** ^The [SQLITE_OPEN_EXCLUSIVE] flag is always used in conjunction
   ** with the [SQLITE_OPEN_CREATE] flag, which are both directly
   ** analogous to the O_EXCL and O_CREAT flags of the POSIX open()
   ** API.  The SQLITE_OPEN_EXCLUSIVE flag, when paired with the 
   ** SQLITE_OPEN_CREATE, is used to indicate that file should always
   ** be created, and that it is an error if it already exists.
   ** It is <i>not</i> used to indicate the file should be opened 
   ** for exclusive access.
   **
   ** ^At least szOsFile bytes of memory are allocated by SQLite
   ** to hold the  [sqlite3_file] structure passed as the third
   ** argument to xOpen.  The xOpen method does not have to
   ** allocate the structure; it should just fill it in.  Note that
   ** the xOpen method must set the sqlite3_file.pMethods to either
   ** a valid [sqlite3_io_methods] object or to NULL.  xOpen must do
   ** this even if the open fails.  SQLite expects that the sqlite3_file.pMethods
   ** element will be valid after xOpen returns regardless of the success
   ** or failure of the xOpen call.
   **
   ** [[sqlite3_vfs.xAccess]]
   ** ^The flags argument to xAccess() may be [SQLITE_ACCESS_EXISTS]
   ** to test for the existence of a file, or [SQLITE_ACCESS_READWRITE] to
   ** test whether a file is readable and writable, or [SQLITE_ACCESS_READ]
   ** to test whether a file is at least readable.   The file can be a
   ** directory.
   **
   ** ^SQLite will always allocate at least mxPathname+1 bytes for the
   ** output buffer xFullPathname.  The exact size of the output buffer
   ** is also passed as a parameter to both  methods. If the output buffer
   ** is not large enough, [SQLITE_CANTOPEN] should be returned. Since this is
   ** handled as a fatal error by SQLite, vfs implementations should endeavor
   ** to prevent this by setting mxPathname to a sufficiently large value.
   **
   ** The xRandomness(), xSleep(), xCurrentTime(), and xCurrentTimeInt64()
   ** interfaces are not strictly a part of the filesystem, but they are
   ** included in the VFS structure for completeness.
   ** The xRandomness() function attempts to return nBytes bytes
   ** of good-quality randomness into zOut.  The return value is
   ** the actual number of bytes of randomness obtained.
   ** The xSleep() method causes the calling thread to sleep for at
   ** least the number of microseconds given.  ^The xCurrentTime()
   ** method returns a Julian Day Number for the current date and time as
   ** a floating point value.
   ** ^The xCurrentTimeInt64() method returns, as an integer, the Julian
   ** Day Number multiplied by 86400000 (the number of milliseconds in 
   ** a 24-hour day).  
   ** ^SQLite will use the xCurrentTimeInt64() method to get the current
   ** date and time if that method is available (if iVersion is 2 or 
   ** greater and the function pointer is not NULL) and will fall back
   ** to xCurrentTime() if xCurrentTimeInt64() is unavailable.
   **
   ** ^The xSetSystemCall(), xGetSystemCall(), and xNestSystemCall() interfaces
   ** are not used by the SQLite core.  These optional interfaces are provided
   ** by some VFSes to facilitate testing of the VFS code. By overriding 
   ** system calls with functions under its control, a test program can
   ** simulate faults and error conditions that would otherwise be difficult
   ** or impossible to induce.  The set of system calls that can be overridden
   ** varies from one VFS to another, and from one version of the same VFS to the
   ** next.  Applications that use these interfaces must be prepared for any
   ** or all of these interfaces to be NULL or for their behavior to change
   ** from one release to the next.  Applications must not attempt to access
   ** any of these methods if the iVersion of the VFS is less than 3.
   */
   typedef struct sqlite3_vfs sqlite3_vfs;
   typedef void (*sqlite3_syscall_ptr)(void);
   struct sqlite3_vfs {
     int iVersion;            /* Structure version number (currently 3) */
     int szOsFile;            /* Size of subclassed sqlite3_file */
     int mxPathname;          /* Maximum file pathname length */
     sqlite3_vfs *pNext;      /* Next registered VFS */
     const char *zName;       /* Name of this virtual file system */
     void *pAppData;          /* Pointer to application-specific data */
     int (*xOpen)(sqlite3_vfs*, const char *zName, sqlite3_file*,
                  int flags, int *pOutFlags);
     int (*xDelete)(sqlite3_vfs*, const char *zName, int syncDir);
     int (*xAccess)(sqlite3_vfs*, const char *zName, int flags, int *pResOut);
     int (*xFullPathname)(sqlite3_vfs*, const char *zName, int nOut, char *zOut);
     void *(*xDlOpen)(sqlite3_vfs*, const char *zFilename);
     void (*xDlError)(sqlite3_vfs*, int nByte, char *zErrMsg);
     void (*(*xDlSym)(sqlite3_vfs*,void*, const char *zSymbol))(void);
     void (*xDlClose)(sqlite3_vfs*, void*);
     int (*xRandomness)(sqlite3_vfs*, int nByte, char *zOut);
     int (*xSleep)(sqlite3_vfs*, int microseconds);
     int (*xCurrentTime)(sqlite3_vfs*, double*);
     int (*xGetLastError)(sqlite3_vfs*, int, char *);
     /*
     ** The methods above are in version 1 of the sqlite_vfs object
     ** definition.  Those that follow are added in version 2 or later
     */
     int (*xCurrentTimeInt64)(sqlite3_vfs*, sqlite3_int64*);
     /*
     ** The methods above are in versions 1 and 2 of the sqlite_vfs object.
     ** Those below are for version 3 and greater.
     */
     int (*xSetSystemCall)(sqlite3_vfs*, const char *zName, sqlite3_syscall_ptr);
     sqlite3_syscall_ptr (*xGetSystemCall)(sqlite3_vfs*, const char *zName);
     const char *(*xNextSystemCall)(sqlite3_vfs*, const char *zName);
     /*
     ** The methods above are in versions 1 through 3 of the sqlite_vfs object.
     ** New fields may be appended in figure versions.  The iVersion
     ** value will increment whenever this happens. 
     */
   };
   
   /*
   ** CAPI3REF: Flags for the xAccess VFS method
   **
   ** These integer constants can be used as the third parameter to
   ** the xAccess method of an [sqlite3_vfs] object.  They determine
   ** what kind of permissions the xAccess method is looking for.
   ** With SQLITE_ACCESS_EXISTS, the xAccess method
   ** simply checks whether the file exists.
   ** With SQLITE_ACCESS_READWRITE, the xAccess method
   ** checks whether the named directory is both readable and writable
   ** (in other words, if files can be added, removed, and renamed within
   ** the directory).
   ** The SQLITE_ACCESS_READWRITE constant is currently used only by the
   ** [temp_store_directory pragma], though this could change in a future
   ** release of SQLite.
   ** With SQLITE_ACCESS_READ, the xAccess method
   ** checks whether the file is readable.  The SQLITE_ACCESS_READ constant is
   ** currently unused, though it might be used in a future release of
   ** SQLite.
   */
   #define SQLITE_ACCESS_EXISTS    0
   #define SQLITE_ACCESS_READWRITE 1   /* Used by PRAGMA temp_store_directory */
   #define SQLITE_ACCESS_READ      2   /* Unused */
   
   /*
   ** CAPI3REF: Flags for the xShmLock VFS method
   **
   ** These integer constants define the various locking operations
   ** allowed by the xShmLock method of [sqlite3_io_methods].  The
   ** following are the only legal combinations of flags to the
   ** xShmLock method:
   **
   ** <ul>
   ** <li>  SQLITE_SHM_LOCK | SQLITE_SHM_SHARED
   ** <li>  SQLITE_SHM_LOCK | SQLITE_SHM_EXCLUSIVE
   ** <li>  SQLITE_SHM_UNLOCK | SQLITE_SHM_SHARED
   ** <li>  SQLITE_SHM_UNLOCK | SQLITE_SHM_EXCLUSIVE
   ** </ul>
   **
   ** When unlocking, the same SHARED or EXCLUSIVE flag must be supplied as
   ** was given no the corresponding lock.  
   **
   ** The xShmLock method can transition between unlocked and SHARED or
   ** between unlocked and EXCLUSIVE.  It cannot transition between SHARED
   ** and EXCLUSIVE.
   */
   #define SQLITE_SHM_UNLOCK       1
   #define SQLITE_SHM_LOCK         2
   #define SQLITE_SHM_SHARED       4
   #define SQLITE_SHM_EXCLUSIVE    8
   
   /*
   ** CAPI3REF: Maximum xShmLock index
   **
   ** The xShmLock method on [sqlite3_io_methods] may use values
   ** between 0 and this upper bound as its "offset" argument.
   ** The SQLite core will never attempt to acquire or release a
   ** lock outside of this range
   */
   #define SQLITE_SHM_NLOCK        8
   
   
   /*
   ** CAPI3REF: Initialize The SQLite Library
   **
   ** ^The sqlite3_initialize() routine initializes the
   ** SQLite library.  ^The sqlite3_shutdown() routine
   ** deallocates any resources that were allocated by sqlite3_initialize().
   ** These routines are designed to aid in process initialization and
   ** shutdown on embedded systems.  Workstation applications using
   ** SQLite normally do not need to invoke either of these routines.
   **
   ** A call to sqlite3_initialize() is an "effective" call if it is
   ** the first time sqlite3_initialize() is invoked during the lifetime of
   ** the process, or if it is the first time sqlite3_initialize() is invoked
   ** following a call to sqlite3_shutdown().  ^(Only an effective call
   ** of sqlite3_initialize() does any initialization.  All other calls
   ** are harmless no-ops.)^
   **
   ** A call to sqlite3_shutdown() is an "effective" call if it is the first
   ** call to sqlite3_shutdown() since the last sqlite3_initialize().  ^(Only
   ** an effective call to sqlite3_shutdown() does any deinitialization.
   ** All other valid calls to sqlite3_shutdown() are harmless no-ops.)^
   **
   ** The sqlite3_initialize() interface is threadsafe, but sqlite3_shutdown()
   ** is not.  The sqlite3_shutdown() interface must only be called from a
   ** single thread.  All open [database connections] must be closed and all
   ** other SQLite resources must be deallocated prior to invoking
   ** sqlite3_shutdown().
   **
   ** Among other things, ^sqlite3_initialize() will invoke
   ** sqlite3_os_init().  Similarly, ^sqlite3_shutdown()
   ** will invoke sqlite3_os_end().
   **
   ** ^The sqlite3_initialize() routine returns [SQLITE_OK] on success.
   ** ^If for some reason, sqlite3_initialize() is unable to initialize
   ** the library (perhaps it is unable to allocate a needed resource such
   ** as a mutex) it returns an [error code] other than [SQLITE_OK].
   **
   ** ^The sqlite3_initialize() routine is called internally by many other
   ** SQLite interfaces so that an application usually does not need to
   ** invoke sqlite3_initialize() directly.  For example, [sqlite3_open()]
   ** calls sqlite3_initialize() so the SQLite library will be automatically
   ** initialized when [sqlite3_open()] is called if it has not be initialized
   ** already.  ^However, if SQLite is compiled with the [SQLITE_OMIT_AUTOINIT]
   ** compile-time option, then the automatic calls to sqlite3_initialize()
   ** are omitted and the application must call sqlite3_initialize() directly
   ** prior to using any other SQLite interface.  For maximum portability,
   ** it is recommended that applications always invoke sqlite3_initialize()
   ** directly prior to using any other SQLite interface.  Future releases
   ** of SQLite may require this.  In other words, the behavior exhibited
   ** when SQLite is compiled with [SQLITE_OMIT_AUTOINIT] might become the
   ** default behavior in some future release of SQLite.
   **
   ** The sqlite3_os_init() routine does operating-system specific
   ** initialization of the SQLite library.  The sqlite3_os_end()
   ** routine undoes the effect of sqlite3_os_init().  Typical tasks
   ** performed by these routines include allocation or deallocation
   ** of static resources, initialization of global variables,
   ** setting up a default [sqlite3_vfs] module, or setting up
   ** a default configuration using [sqlite3_config()].
   **
   ** The application should never invoke either sqlite3_os_init()
   ** or sqlite3_os_end() directly.  The application should only invoke
   ** sqlite3_initialize() and sqlite3_shutdown().  The sqlite3_os_init()
   ** interface is called automatically by sqlite3_initialize() and
   ** sqlite3_os_end() is called by sqlite3_shutdown().  Appropriate
   ** implementations for sqlite3_os_init() and sqlite3_os_end()
   ** are built into SQLite when it is compiled for Unix, Windows, or OS/2.
   ** When [custom builds | built for other platforms]
   ** (using the [SQLITE_OS_OTHER=1] compile-time
   ** option) the application must supply a suitable implementation for
   ** sqlite3_os_init() and sqlite3_os_end().  An application-supplied
   ** implementation of sqlite3_os_init() or sqlite3_os_end()
   ** must return [SQLITE_OK] on success and some other [error code] upon
   ** failure.
   */
   SQLITE_API int sqlite3_initialize(void);
   SQLITE_API int sqlite3_shutdown(void);
   SQLITE_API int sqlite3_os_init(void);
   SQLITE_API int sqlite3_os_end(void);
   
   /*
   ** CAPI3REF: Configuring The SQLite Library
   **
   ** The sqlite3_config() interface is used to make global configuration
   ** changes to SQLite in order to tune SQLite to the specific needs of
   ** the application.  The default configuration is recommended for most
   ** applications and so this routine is usually not necessary.  It is
   ** provided to support rare applications with unusual needs.
   **
   ** The sqlite3_config() interface is not threadsafe.  The application
   ** must insure that no other SQLite interfaces are invoked by other
   ** threads while sqlite3_config() is running.  Furthermore, sqlite3_config()
   ** may only be invoked prior to library initialization using
   ** [sqlite3_initialize()] or after shutdown by [sqlite3_shutdown()].
   ** ^If sqlite3_config() is called after [sqlite3_initialize()] and before
   ** [sqlite3_shutdown()] then it will return SQLITE_MISUSE.
   ** Note, however, that ^sqlite3_config() can be called as part of the
   ** implementation of an application-defined [sqlite3_os_init()].
   **
   ** The first argument to sqlite3_config() is an integer
   ** [configuration option] that determines
   ** what property of SQLite is to be configured.  Subsequent arguments
   ** vary depending on the [configuration option]
   ** in the first argument.
   **
   ** ^When a configuration option is set, sqlite3_config() returns [SQLITE_OK].
   ** ^If the option is unknown or SQLite is unable to set the option
   ** then this routine returns a non-zero [error code].
   */
   SQLITE_API int sqlite3_config(int, ...);
   
   /*
   ** CAPI3REF: Configure database connections
   **
   ** The sqlite3_db_config() interface is used to make configuration
   ** changes to a [database connection].  The interface is similar to
   ** [sqlite3_config()] except that the changes apply to a single
   ** [database connection] (specified in the first argument).
   **
   ** The second argument to sqlite3_db_config(D,V,...)  is the
   ** [SQLITE_DBCONFIG_LOOKASIDE | configuration verb] - an integer code 
   ** that indicates what aspect of the [database connection] is being configured.
   ** Subsequent arguments vary depending on the configuration verb.
   **
   ** ^Calls to sqlite3_db_config() return SQLITE_OK if and only if
   ** the call is considered successful.
   */
   SQLITE_API int sqlite3_db_config(sqlite3*, int op, ...);
   
   /*
   ** CAPI3REF: Memory Allocation Routines
   **
   ** An instance of this object defines the interface between SQLite
   ** and low-level memory allocation routines.
   **
   ** This object is used in only one place in the SQLite interface.
   ** A pointer to an instance of this object is the argument to
   ** [sqlite3_config()] when the configuration option is
   ** [SQLITE_CONFIG_MALLOC] or [SQLITE_CONFIG_GETMALLOC].  
   ** By creating an instance of this object
   ** and passing it to [sqlite3_config]([SQLITE_CONFIG_MALLOC])
   ** during configuration, an application can specify an alternative
   ** memory allocation subsystem for SQLite to use for all of its
   ** dynamic memory needs.
   **
   ** Note that SQLite comes with several [built-in memory allocators]
   ** that are perfectly adequate for the overwhelming majority of applications
   ** and that this object is only useful to a tiny minority of applications
   ** with specialized memory allocation requirements.  This object is
   ** also used during testing of SQLite in order to specify an alternative
   ** memory allocator that simulates memory out-of-memory conditions in
   ** order to verify that SQLite recovers gracefully from such
   ** conditions.
   **
   ** The xMalloc, xRealloc, and xFree methods must work like the
   ** malloc(), realloc() and free() functions from the standard C library.
   ** ^SQLite guarantees that the second argument to
   ** xRealloc is always a value returned by a prior call to xRoundup.
   **
   ** xSize should return the allocated size of a memory allocation
   ** previously obtained from xMalloc or xRealloc.  The allocated size
   ** is always at least as big as the requested size but may be larger.
   **
   ** The xRoundup method returns what would be the allocated size of
   ** a memory allocation given a particular requested size.  Most memory
   ** allocators round up memory allocations at least to the next multiple
   ** of 8.  Some allocators round up to a larger multiple or to a power of 2.
   ** Every memory allocation request coming in through [sqlite3_malloc()]
   ** or [sqlite3_realloc()] first calls xRoundup.  If xRoundup returns 0, 
   ** that causes the corresponding memory allocation to fail.
   **
   ** The xInit method initializes the memory allocator.  (For example,
   ** it might allocate any require mutexes or initialize internal data
   ** structures.  The xShutdown method is invoked (indirectly) by
   ** [sqlite3_shutdown()] and should deallocate any resources acquired
   ** by xInit.  The pAppData pointer is used as the only parameter to
   ** xInit and xShutdown.
   **
   ** SQLite holds the [SQLITE_MUTEX_STATIC_MASTER] mutex when it invokes
   ** the xInit method, so the xInit method need not be threadsafe.  The
   ** xShutdown method is only called from [sqlite3_shutdown()] so it does
   ** not need to be threadsafe either.  For all other methods, SQLite
   ** holds the [SQLITE_MUTEX_STATIC_MEM] mutex as long as the
   ** [SQLITE_CONFIG_MEMSTATUS] configuration option is turned on (which
   ** it is by default) and so the methods are automatically serialized.
   ** However, if [SQLITE_CONFIG_MEMSTATUS] is disabled, then the other
   ** methods must be threadsafe or else make their own arrangements for
   ** serialization.
   **
   ** SQLite will never invoke xInit() more than once without an intervening
   ** call to xShutdown().
   */
   typedef struct sqlite3_mem_methods sqlite3_mem_methods;
   struct sqlite3_mem_methods {
     void *(*xMalloc)(int);         /* Memory allocation function */
     void (*xFree)(void*);          /* Free a prior allocation */
     void *(*xRealloc)(void*,int);  /* Resize an allocation */
     int (*xSize)(void*);           /* Return the size of an allocation */
     int (*xRoundup)(int);          /* Round up request size to allocation size */
     int (*xInit)(void*);           /* Initialize the memory allocator */
     void (*xShutdown)(void*);      /* Deinitialize the memory allocator */
     void *pAppData;                /* Argument to xInit() and xShutdown() */
   };
   
   /*
   ** CAPI3REF: Configuration Options
   ** KEYWORDS: {configuration option}
   **
   ** These constants are the available integer configuration options that
   ** can be passed as the first argument to the [sqlite3_config()] interface.
   **
   ** New configuration options may be added in future releases of SQLite.
   ** Existing configuration options might be discontinued.  Applications
   ** should check the return code from [sqlite3_config()] to make sure that
   ** the call worked.  The [sqlite3_config()] interface will return a
   ** non-zero [error code] if a discontinued or unsupported configuration option
   ** is invoked.
   **
   ** <dl>
   ** [[SQLITE_CONFIG_SINGLETHREAD]] <dt>SQLITE_CONFIG_SINGLETHREAD</dt>
   ** <dd>There are no arguments to this option.  ^This option sets the
   ** [threading mode] to Single-thread.  In other words, it disables
   ** all mutexing and puts SQLite into a mode where it can only be used
   ** by a single thread.   ^If SQLite is compiled with
   ** the [SQLITE_THREADSAFE | SQLITE_THREADSAFE=0] compile-time option then
   ** it is not possible to change the [threading mode] from its default
   ** value of Single-thread and so [sqlite3_config()] will return 
   ** [SQLITE_ERROR] if called with the SQLITE_CONFIG_SINGLETHREAD
   ** configuration option.</dd>
   **
   ** [[SQLITE_CONFIG_MULTITHREAD]] <dt>SQLITE_CONFIG_MULTITHREAD</dt>
   ** <dd>There are no arguments to this option.  ^This option sets the
   ** [threading mode] to Multi-thread.  In other words, it disables
   ** mutexing on [database connection] and [prepared statement] objects.
   ** The application is responsible for serializing access to
   ** [database connections] and [prepared statements].  But other mutexes
   ** are enabled so that SQLite will be safe to use in a multi-threaded
   ** environment as long as no two threads attempt to use the same
   ** [database connection] at the same time.  ^If SQLite is compiled with
   ** the [SQLITE_THREADSAFE | SQLITE_THREADSAFE=0] compile-time option then
   ** it is not possible to set the Multi-thread [threading mode] and
   ** [sqlite3_config()] will return [SQLITE_ERROR] if called with the
   ** SQLITE_CONFIG_MULTITHREAD configuration option.</dd>
   **
   ** [[SQLITE_CONFIG_SERIALIZED]] <dt>SQLITE_CONFIG_SERIALIZED</dt>
   ** <dd>There are no arguments to this option.  ^This option sets the
   ** [threading mode] to Serialized. In other words, this option enables
   ** all mutexes including the recursive
   ** mutexes on [database connection] and [prepared statement] objects.
   ** In this mode (which is the default when SQLite is compiled with
   ** [SQLITE_THREADSAFE=1]) the SQLite library will itself serialize access
   ** to [database connections] and [prepared statements] so that the
   ** application is free to use the same [database connection] or the
   ** same [prepared statement] in different threads at the same time.
   ** ^If SQLite is compiled with
   ** the [SQLITE_THREADSAFE | SQLITE_THREADSAFE=0] compile-time option then
   ** it is not possible to set the Serialized [threading mode] and
   ** [sqlite3_config()] will return [SQLITE_ERROR] if called with the
   ** SQLITE_CONFIG_SERIALIZED configuration option.</dd>
   **
   ** [[SQLITE_CONFIG_MALLOC]] <dt>SQLITE_CONFIG_MALLOC</dt>
   ** <dd> ^(This option takes a single argument which is a pointer to an
   ** instance of the [sqlite3_mem_methods] structure.  The argument specifies
   ** alternative low-level memory allocation routines to be used in place of
   ** the memory allocation routines built into SQLite.)^ ^SQLite makes
   ** its own private copy of the content of the [sqlite3_mem_methods] structure
   ** before the [sqlite3_config()] call returns.</dd>
   **
   ** [[SQLITE_CONFIG_GETMALLOC]] <dt>SQLITE_CONFIG_GETMALLOC</dt>
   ** <dd> ^(This option takes a single argument which is a pointer to an
   ** instance of the [sqlite3_mem_methods] structure.  The [sqlite3_mem_methods]
   ** structure is filled with the currently defined memory allocation routines.)^
   ** This option can be used to overload the default memory allocation
   ** routines with a wrapper that simulations memory allocation failure or
   ** tracks memory usage, for example. </dd>
   **
   ** [[SQLITE_CONFIG_MEMSTATUS]] <dt>SQLITE_CONFIG_MEMSTATUS</dt>
   ** <dd> ^This option takes single argument of type int, interpreted as a 
   ** boolean, which enables or disables the collection of memory allocation 
   ** statistics. ^(When memory allocation statistics are disabled, the 
   ** following SQLite interfaces become non-operational:
   **   <ul>
   **   <li> [sqlite3_memory_used()]
   **   <li> [sqlite3_memory_highwater()]
   **   <li> [sqlite3_soft_heap_limit64()]
   **   <li> [sqlite3_status()]
   **   </ul>)^
   ** ^Memory allocation statistics are enabled by default unless SQLite is
   ** compiled with [SQLITE_DEFAULT_MEMSTATUS]=0 in which case memory
   ** allocation statistics are disabled by default.
   ** </dd>
   **
   ** [[SQLITE_CONFIG_SCRATCH]] <dt>SQLITE_CONFIG_SCRATCH</dt>
   ** <dd> ^This option specifies a static memory buffer that SQLite can use for
   ** scratch memory.  There are three arguments:  A pointer an 8-byte
   ** aligned memory buffer from which the scratch allocations will be
   ** drawn, the size of each scratch allocation (sz),
   ** and the maximum number of scratch allocations (N).  The sz
   ** argument must be a multiple of 16.
   ** The first argument must be a pointer to an 8-byte aligned buffer
   ** of at least sz*N bytes of memory.
   ** ^SQLite will use no more than two scratch buffers per thread.  So
   ** N should be set to twice the expected maximum number of threads.
   ** ^SQLite will never require a scratch buffer that is more than 6
   ** times the database page size. ^If SQLite needs needs additional
   ** scratch memory beyond what is provided by this configuration option, then 
   ** [sqlite3_malloc()] will be used to obtain the memory needed.</dd>
   **
   ** [[SQLITE_CONFIG_PAGECACHE]] <dt>SQLITE_CONFIG_PAGECACHE</dt>
   ** <dd> ^This option specifies a static memory buffer that SQLite can use for
   ** the database page cache with the default page cache implementation.  
   ** This configuration should not be used if an application-define page
   ** cache implementation is loaded using the SQLITE_CONFIG_PCACHE2 option.
   ** There are three arguments to this option: A pointer to 8-byte aligned
   ** memory, the size of each page buffer (sz), and the number of pages (N).
   ** The sz argument should be the size of the largest database page
   ** (a power of two between 512 and 32768) plus a little extra for each
   ** page header.  ^The page header size is 20 to 40 bytes depending on
   ** the host architecture.  ^It is harmless, apart from the wasted memory,
   ** to make sz a little too large.  The first
   ** argument should point to an allocation of at least sz*N bytes of memory.
   ** ^SQLite will use the memory provided by the first argument to satisfy its
   ** memory needs for the first N pages that it adds to cache.  ^If additional
   ** page cache memory is needed beyond what is provided by this option, then
   ** SQLite goes to [sqlite3_malloc()] for the additional storage space.
   ** The pointer in the first argument must
   ** be aligned to an 8-byte boundary or subsequent behavior of SQLite
   ** will be undefined.</dd>
   **
   ** [[SQLITE_CONFIG_HEAP]] <dt>SQLITE_CONFIG_HEAP</dt>
   ** <dd> ^This option specifies a static memory buffer that SQLite will use
   ** for all of its dynamic memory allocation needs beyond those provided
   ** for by [SQLITE_CONFIG_SCRATCH] and [SQLITE_CONFIG_PAGECACHE].
   ** There are three arguments: An 8-byte aligned pointer to the memory,
   ** the number of bytes in the memory buffer, and the minimum allocation size.
   ** ^If the first pointer (the memory pointer) is NULL, then SQLite reverts
   ** to using its default memory allocator (the system malloc() implementation),
   ** undoing any prior invocation of [SQLITE_CONFIG_MALLOC].  ^If the
   ** memory pointer is not NULL and either [SQLITE_ENABLE_MEMSYS3] or
   ** [SQLITE_ENABLE_MEMSYS5] are defined, then the alternative memory
   ** allocator is engaged to handle all of SQLites memory allocation needs.
   ** The first pointer (the memory pointer) must be aligned to an 8-byte
   ** boundary or subsequent behavior of SQLite will be undefined.
   ** The minimum allocation size is capped at 2**12. Reasonable values
   ** for the minimum allocation size are 2**5 through 2**8.</dd>
   **
   ** [[SQLITE_CONFIG_MUTEX]] <dt>SQLITE_CONFIG_MUTEX</dt>
   ** <dd> ^(This option takes a single argument which is a pointer to an
   ** instance of the [sqlite3_mutex_methods] structure.  The argument specifies
   ** alternative low-level mutex routines to be used in place
   ** the mutex routines built into SQLite.)^  ^SQLite makes a copy of the
   ** content of the [sqlite3_mutex_methods] structure before the call to
   ** [sqlite3_config()] returns. ^If SQLite is compiled with
   ** the [SQLITE_THREADSAFE | SQLITE_THREADSAFE=0] compile-time option then
   ** the entire mutexing subsystem is omitted from the build and hence calls to
   ** [sqlite3_config()] with the SQLITE_CONFIG_MUTEX configuration option will
   ** return [SQLITE_ERROR].</dd>
   **
   ** [[SQLITE_CONFIG_GETMUTEX]] <dt>SQLITE_CONFIG_GETMUTEX</dt>
   ** <dd> ^(This option takes a single argument which is a pointer to an
   ** instance of the [sqlite3_mutex_methods] structure.  The
   ** [sqlite3_mutex_methods]
   ** structure is filled with the currently defined mutex routines.)^
   ** This option can be used to overload the default mutex allocation
   ** routines with a wrapper used to track mutex usage for performance
   ** profiling or testing, for example.   ^If SQLite is compiled with
   ** the [SQLITE_THREADSAFE | SQLITE_THREADSAFE=0] compile-time option then
   ** the entire mutexing subsystem is omitted from the build and hence calls to
   ** [sqlite3_config()] with the SQLITE_CONFIG_GETMUTEX configuration option will
   ** return [SQLITE_ERROR].</dd>
   **
   ** [[SQLITE_CONFIG_LOOKASIDE]] <dt>SQLITE_CONFIG_LOOKASIDE</dt>
   ** <dd> ^(This option takes two arguments that determine the default
   ** memory allocation for the lookaside memory allocator on each
   ** [database connection].  The first argument is the
   ** size of each lookaside buffer slot and the second is the number of
   ** slots allocated to each database connection.)^  ^(This option sets the
   ** <i>default</i> lookaside size. The [SQLITE_DBCONFIG_LOOKASIDE]
   ** verb to [sqlite3_db_config()] can be used to change the lookaside
   ** configuration on individual connections.)^ </dd>
   **
   ** [[SQLITE_CONFIG_PCACHE2]] <dt>SQLITE_CONFIG_PCACHE2</dt>
   ** <dd> ^(This option takes a single argument which is a pointer to
   ** an [sqlite3_pcache_methods2] object.  This object specifies the interface
   ** to a custom page cache implementation.)^  ^SQLite makes a copy of the
   ** object and uses it for page cache memory allocations.</dd>
   **
   ** [[SQLITE_CONFIG_GETPCACHE2]] <dt>SQLITE_CONFIG_GETPCACHE2</dt>
   ** <dd> ^(This option takes a single argument which is a pointer to an
   ** [sqlite3_pcache_methods2] object.  SQLite copies of the current
   ** page cache implementation into that object.)^ </dd>
   **
   ** [[SQLITE_CONFIG_LOG]] <dt>SQLITE_CONFIG_LOG</dt>
   ** <dd> ^The SQLITE_CONFIG_LOG option takes two arguments: a pointer to a
   ** function with a call signature of void(*)(void*,int,const char*), 
   ** and a pointer to void. ^If the function pointer is not NULL, it is
   ** invoked by [sqlite3_log()] to process each logging event.  ^If the
   ** function pointer is NULL, the [sqlite3_log()] interface becomes a no-op.
   ** ^The void pointer that is the second argument to SQLITE_CONFIG_LOG is
   ** passed through as the first parameter to the application-defined logger
   ** function whenever that function is invoked.  ^The second parameter to
   ** the logger function is a copy of the first parameter to the corresponding
   ** [sqlite3_log()] call and is intended to be a [result code] or an
   ** [extended result code].  ^The third parameter passed to the logger is
   ** log message after formatting via [sqlite3_snprintf()].
   ** The SQLite logging interface is not reentrant; the logger function
   ** supplied by the application must not invoke any SQLite interface.
   ** In a multi-threaded application, the application-defined logger
   ** function must be threadsafe. </dd>
   **
   ** [[SQLITE_CONFIG_URI]] <dt>SQLITE_CONFIG_URI
   ** <dd> This option takes a single argument of type int. If non-zero, then
   ** URI handling is globally enabled. If the parameter is zero, then URI handling
   ** is globally disabled. If URI handling is globally enabled, all filenames
   ** passed to [sqlite3_open()], [sqlite3_open_v2()], [sqlite3_open16()] or
   ** specified as part of [ATTACH] commands are interpreted as URIs, regardless
   ** of whether or not the [SQLITE_OPEN_URI] flag is set when the database
   ** connection is opened. If it is globally disabled, filenames are
   ** only interpreted as URIs if the SQLITE_OPEN_URI flag is set when the
   ** database connection is opened. By default, URI handling is globally
   ** disabled. The default value may be changed by compiling with the
   ** [SQLITE_USE_URI] symbol defined.
   **
   ** [[SQLITE_CONFIG_PCACHE]] [[SQLITE_CONFIG_GETPCACHE]]
   ** <dt>SQLITE_CONFIG_PCACHE and SQLITE_CONFNIG_GETPCACHE
   ** <dd> These options are obsolete and should not be used by new code.
   ** They are retained for backwards compatibility but are now no-ops.
   ** </dl>
   */
   #define SQLITE_CONFIG_SINGLETHREAD  1  /* nil */
   #define SQLITE_CONFIG_MULTITHREAD   2  /* nil */
   #define SQLITE_CONFIG_SERIALIZED    3  /* nil */
   #define SQLITE_CONFIG_MALLOC        4  /* sqlite3_mem_methods* */
   #define SQLITE_CONFIG_GETMALLOC     5  /* sqlite3_mem_methods* */
   #define SQLITE_CONFIG_SCRATCH       6  /* void*, int sz, int N */
   #define SQLITE_CONFIG_PAGECACHE     7  /* void*, int sz, int N */
   #define SQLITE_CONFIG_HEAP          8  /* void*, int nByte, int min */
   #define SQLITE_CONFIG_MEMSTATUS     9  /* boolean */
   #define SQLITE_CONFIG_MUTEX        10  /* sqlite3_mutex_methods* */
   #define SQLITE_CONFIG_GETMUTEX     11  /* sqlite3_mutex_methods* */
   /* previously SQLITE_CONFIG_CHUNKALLOC 12 which is now unused. */ 
   #define SQLITE_CONFIG_LOOKASIDE    13  /* int int */
   #define SQLITE_CONFIG_PCACHE       14  /* no-op */
   #define SQLITE_CONFIG_GETPCACHE    15  /* no-op */
   #define SQLITE_CONFIG_LOG          16  /* xFunc, void* */
   #define SQLITE_CONFIG_URI          17  /* int */
   #define SQLITE_CONFIG_PCACHE2      18  /* sqlite3_pcache_methods2* */
   #define SQLITE_CONFIG_GETPCACHE2   19  /* sqlite3_pcache_methods2* */
   
   /*
   ** CAPI3REF: Database Connection Configuration Options
   **
   ** These constants are the available integer configuration options that
   ** can be passed as the second argument to the [sqlite3_db_config()] interface.
   **
   ** New configuration options may be added in future releases of SQLite.
   ** Existing configuration options might be discontinued.  Applications
   ** should check the return code from [sqlite3_db_config()] to make sure that
   ** the call worked.  ^The [sqlite3_db_config()] interface will return a
   ** non-zero [error code] if a discontinued or unsupported configuration option
   ** is invoked.
   **
   ** <dl>
   ** <dt>SQLITE_DBCONFIG_LOOKASIDE</dt>
   ** <dd> ^This option takes three additional arguments that determine the 
   ** [lookaside memory allocator] configuration for the [database connection].
   ** ^The first argument (the third parameter to [sqlite3_db_config()] is a
   ** pointer to a memory buffer to use for lookaside memory.
   ** ^The first argument after the SQLITE_DBCONFIG_LOOKASIDE verb
   ** may be NULL in which case SQLite will allocate the
   ** lookaside buffer itself using [sqlite3_malloc()]. ^The second argument is the
   ** size of each lookaside buffer slot.  ^The third argument is the number of
   ** slots.  The size of the buffer in the first argument must be greater than
   ** or equal to the product of the second and third arguments.  The buffer
   ** must be aligned to an 8-byte boundary.  ^If the second argument to
   ** SQLITE_DBCONFIG_LOOKASIDE is not a multiple of 8, it is internally
   ** rounded down to the next smaller multiple of 8.  ^(The lookaside memory
   ** configuration for a database connection can only be changed when that
   ** connection is not currently using lookaside memory, or in other words
   ** when the "current value" returned by
   ** [sqlite3_db_status](D,[SQLITE_CONFIG_LOOKASIDE],...) is zero.
   ** Any attempt to change the lookaside memory configuration when lookaside
   ** memory is in use leaves the configuration unchanged and returns 
   ** [SQLITE_BUSY].)^</dd>
   **
   ** <dt>SQLITE_DBCONFIG_ENABLE_FKEY</dt>
   ** <dd> ^This option is used to enable or disable the enforcement of
   ** [foreign key constraints].  There should be two additional arguments.
   ** The first argument is an integer which is 0 to disable FK enforcement,
   ** positive to enable FK enforcement or negative to leave FK enforcement
   ** unchanged.  The second parameter is a pointer to an integer into which
   ** is written 0 or 1 to indicate whether FK enforcement is off or on
   ** following this call.  The second parameter may be a NULL pointer, in
   ** which case the FK enforcement setting is not reported back. </dd>
   **
   ** <dt>SQLITE_DBCONFIG_ENABLE_TRIGGER</dt>
   ** <dd> ^This option is used to enable or disable [CREATE TRIGGER | triggers].
   ** There should be two additional arguments.
   ** The first argument is an integer which is 0 to disable triggers,
   ** positive to enable triggers or negative to leave the setting unchanged.
   ** The second parameter is a pointer to an integer into which
   ** is written 0 or 1 to indicate whether triggers are disabled or enabled
   ** following this call.  The second parameter may be a NULL pointer, in
   ** which case the trigger setting is not reported back. </dd>
   **
   ** </dl>
   */
   #define SQLITE_DBCONFIG_LOOKASIDE       1001  /* void* int int */
   #define SQLITE_DBCONFIG_ENABLE_FKEY     1002  /* int int* */
   #define SQLITE_DBCONFIG_ENABLE_TRIGGER  1003  /* int int* */
   
   
   /*
   ** CAPI3REF: Enable Or Disable Extended Result Codes
   **
   ** ^The sqlite3_extended_result_codes() routine enables or disables the
   ** [extended result codes] feature of SQLite. ^The extended result
   ** codes are disabled by default for historical compatibility.
   */
   SQLITE_API int sqlite3_extended_result_codes(sqlite3*, int onoff);
   
   /*
   ** CAPI3REF: Last Insert Rowid
   **
   ** ^Each entry in an SQLite table has a unique 64-bit signed
   ** integer key called the [ROWID | "rowid"]. ^The rowid is always available
   ** as an undeclared column named ROWID, OID, or _ROWID_ as long as those
   ** names are not also used by explicitly declared columns. ^If
   ** the table has a column of type [INTEGER PRIMARY KEY] then that column
   ** is another alias for the rowid.
   **
   ** ^This routine returns the [rowid] of the most recent
   ** successful [INSERT] into the database from the [database connection]
   ** in the first argument.  ^As of SQLite version 3.7.7, this routines
   ** records the last insert rowid of both ordinary tables and [virtual tables].
   ** ^If no successful [INSERT]s
   ** have ever occurred on that database connection, zero is returned.
   **
   ** ^(If an [INSERT] occurs within a trigger or within a [virtual table]
   ** method, then this routine will return the [rowid] of the inserted
   ** row as long as the trigger or virtual table method is running.
   ** But once the trigger or virtual table method ends, the value returned 
   ** by this routine reverts to what it was before the trigger or virtual
   ** table method began.)^
   **
   ** ^An [INSERT] that fails due to a constraint violation is not a
   ** successful [INSERT] and does not change the value returned by this
   ** routine.  ^Thus INSERT OR FAIL, INSERT OR IGNORE, INSERT OR ROLLBACK,
   ** and INSERT OR ABORT make no changes to the return value of this
   ** routine when their insertion fails.  ^(When INSERT OR REPLACE
   ** encounters a constraint violation, it does not fail.  The
   ** INSERT continues to completion after deleting rows that caused
   ** the constraint problem so INSERT OR REPLACE will always change
   ** the return value of this interface.)^
   **
   ** ^For the purposes of this routine, an [INSERT] is considered to
   ** be successful even if it is subsequently rolled back.
   **
   ** This function is accessible to SQL statements via the
   ** [last_insert_rowid() SQL function].
   **
   ** If a separate thread performs a new [INSERT] on the same
   ** database connection while the [sqlite3_last_insert_rowid()]
   ** function is running and thus changes the last insert [rowid],
   ** then the value returned by [sqlite3_last_insert_rowid()] is
   ** unpredictable and might not equal either the old or the new
   ** last insert [rowid].
   */
   SQLITE_API sqlite3_int64 sqlite3_last_insert_rowid(sqlite3*);
   
   /*
   ** CAPI3REF: Count The Number Of Rows Modified
   **
   ** ^This function returns the number of database rows that were changed
   ** or inserted or deleted by the most recently completed SQL statement
   ** on the [database connection] specified by the first parameter.
   ** ^(Only changes that are directly specified by the [INSERT], [UPDATE],
   ** or [DELETE] statement are counted.  Auxiliary changes caused by
   ** triggers or [foreign key actions] are not counted.)^ Use the
   ** [sqlite3_total_changes()] function to find the total number of changes
   ** including changes caused by triggers and foreign key actions.
   **
   ** ^Changes to a view that are simulated by an [INSTEAD OF trigger]
   ** are not counted.  Only real table changes are counted.
   **
   ** ^(A "row change" is a change to a single row of a single table
   ** caused by an INSERT, DELETE, or UPDATE statement.  Rows that
   ** are changed as side effects of [REPLACE] constraint resolution,
   ** rollback, ABORT processing, [DROP TABLE], or by any other
   ** mechanisms do not count as direct row changes.)^
   **
   ** A "trigger context" is a scope of execution that begins and
   ** ends with the script of a [CREATE TRIGGER | trigger]. 
   ** Most SQL statements are
   ** evaluated outside of any trigger.  This is the "top level"
   ** trigger context.  If a trigger fires from the top level, a
   ** new trigger context is entered for the duration of that one
   ** trigger.  Subtriggers create subcontexts for their duration.
   **
   ** ^Calling [sqlite3_exec()] or [sqlite3_step()] recursively does
   ** not create a new trigger context.
   **
   ** ^This function returns the number of direct row changes in the
   ** most recent INSERT, UPDATE, or DELETE statement within the same
   ** trigger context.
   **
   ** ^Thus, when called from the top level, this function returns the
   ** number of changes in the most recent INSERT, UPDATE, or DELETE
   ** that also occurred at the top level.  ^(Within the body of a trigger,
   ** the sqlite3_changes() interface can be called to find the number of
   ** changes in the most recently completed INSERT, UPDATE, or DELETE
   ** statement within the body of the same trigger.
   ** However, the number returned does not include changes
   ** caused by subtriggers since those have their own context.)^
   **
   ** See also the [sqlite3_total_changes()] interface, the
   ** [count_changes pragma], and the [changes() SQL function].
   **
   ** If a separate thread makes changes on the same database connection
   ** while [sqlite3_changes()] is running then the value returned
   ** is unpredictable and not meaningful.
   */
   SQLITE_API int sqlite3_changes(sqlite3*);
   
   /*
   ** CAPI3REF: Total Number Of Rows Modified
   **
   ** ^This function returns the number of row changes caused by [INSERT],
   ** [UPDATE] or [DELETE] statements since the [database connection] was opened.
   ** ^(The count returned by sqlite3_total_changes() includes all changes
   ** from all [CREATE TRIGGER | trigger] contexts and changes made by
   ** [foreign key actions]. However,
   ** the count does not include changes used to implement [REPLACE] constraints,
   ** do rollbacks or ABORT processing, or [DROP TABLE] processing.  The
   ** count does not include rows of views that fire an [INSTEAD OF trigger],
   ** though if the INSTEAD OF trigger makes changes of its own, those changes 
   ** are counted.)^
   ** ^The sqlite3_total_changes() function counts the changes as soon as
   ** the statement that makes them is completed (when the statement handle
   ** is passed to [sqlite3_reset()] or [sqlite3_finalize()]).
   **
   ** See also the [sqlite3_changes()] interface, the
   ** [count_changes pragma], and the [total_changes() SQL function].
   **
   ** If a separate thread makes changes on the same database connection
   ** while [sqlite3_total_changes()] is running then the value
   ** returned is unpredictable and not meaningful.
   */
   SQLITE_API int sqlite3_total_changes(sqlite3*);
   
   /*
   ** CAPI3REF: Interrupt A Long-Running Query
   **
   ** ^This function causes any pending database operation to abort and
   ** return at its earliest opportunity. This routine is typically
   ** called in response to a user action such as pressing "Cancel"
   ** or Ctrl-C where the user wants a long query operation to halt
   ** immediately.
   **
   ** ^It is safe to call this routine from a thread different from the
   ** thread that is currently running the database operation.  But it
   ** is not safe to call this routine with a [database connection] that
   ** is closed or might close before sqlite3_interrupt() returns.
   **
   ** ^If an SQL operation is very nearly finished at the time when
   ** sqlite3_interrupt() is called, then it might not have an opportunity
   ** to be interrupted and might continue to completion.
   **
   ** ^An SQL operation that is interrupted will return [SQLITE_INTERRUPT].
   ** ^If the interrupted SQL operation is an INSERT, UPDATE, or DELETE
   ** that is inside an explicit transaction, then the entire transaction
   ** will be rolled back automatically.
   **
   ** ^The sqlite3_interrupt(D) call is in effect until all currently running
   ** SQL statements on [database connection] D complete.  ^Any new SQL statements
   ** that are started after the sqlite3_interrupt() call and before the 
   ** running statements reaches zero are interrupted as if they had been
   ** running prior to the sqlite3_interrupt() call.  ^New SQL statements
   ** that are started after the running statement count reaches zero are
   ** not effected by the sqlite3_interrupt().
   ** ^A call to sqlite3_interrupt(D) that occurs when there are no running
   ** SQL statements is a no-op and has no effect on SQL statements
   ** that are started after the sqlite3_interrupt() call returns.
   **
   ** If the database connection closes while [sqlite3_interrupt()]
   ** is running then bad things will likely happen.
   */
   SQLITE_API void sqlite3_interrupt(sqlite3*);
   
   /*
   ** CAPI3REF: Determine If An SQL Statement Is Complete
   **
   ** These routines are useful during command-line input to determine if the
   ** currently entered text seems to form a complete SQL statement or
   ** if additional input is needed before sending the text into
   ** SQLite for parsing.  ^These routines return 1 if the input string
   ** appears to be a complete SQL statement.  ^A statement is judged to be
   ** complete if it ends with a semicolon token and is not a prefix of a
   ** well-formed CREATE TRIGGER statement.  ^Semicolons that are embedded within
   ** string literals or quoted identifier names or comments are not
   ** independent tokens (they are part of the token in which they are
   ** embedded) and thus do not count as a statement terminator.  ^Whitespace
   ** and comments that follow the final semicolon are ignored.
   **
   ** ^These routines return 0 if the statement is incomplete.  ^If a
   ** memory allocation fails, then SQLITE_NOMEM is returned.
   **
   ** ^These routines do not parse the SQL statements thus
   ** will not detect syntactically incorrect SQL.
   **
   ** ^(If SQLite has not been initialized using [sqlite3_initialize()] prior 
   ** to invoking sqlite3_complete16() then sqlite3_initialize() is invoked
   ** automatically by sqlite3_complete16().  If that initialization fails,
   ** then the return value from sqlite3_complete16() will be non-zero
   ** regardless of whether or not the input SQL is complete.)^
   **
   ** The input to [sqlite3_complete()] must be a zero-terminated
   ** UTF-8 string.
   **
   ** The input to [sqlite3_complete16()] must be a zero-terminated
   ** UTF-16 string in native byte order.
   */
   SQLITE_API int sqlite3_complete(const char *sql);
   SQLITE_API int sqlite3_complete16(const void *sql);
   
   /*
   ** CAPI3REF: Register A Callback To Handle SQLITE_BUSY Errors
   **
   ** ^This routine sets a callback function that might be invoked whenever
   ** an attempt is made to open a database table that another thread
   ** or process has locked.
   **
   ** ^If the busy callback is NULL, then [SQLITE_BUSY] or [SQLITE_IOERR_BLOCKED]
   ** is returned immediately upon encountering the lock.  ^If the busy callback
   ** is not NULL, then the callback might be invoked with two arguments.
   **
   ** ^The first argument to the busy handler is a copy of the void* pointer which
   ** is the third argument to sqlite3_busy_handler().  ^The second argument to
   ** the busy handler callback is the number of times that the busy handler has
   ** been invoked for this locking event.  ^If the
   ** busy callback returns 0, then no additional attempts are made to
   ** access the database and [SQLITE_BUSY] or [SQLITE_IOERR_BLOCKED] is returned.
   ** ^If the callback returns non-zero, then another attempt
   ** is made to open the database for reading and the cycle repeats.
   **
   ** The presence of a busy handler does not guarantee that it will be invoked
   ** when there is lock contention. ^If SQLite determines that invoking the busy
   ** handler could result in a deadlock, it will go ahead and return [SQLITE_BUSY]
   ** or [SQLITE_IOERR_BLOCKED] instead of invoking the busy handler.
   ** Consider a scenario where one process is holding a read lock that
   ** it is trying to promote to a reserved lock and
   ** a second process is holding a reserved lock that it is trying
   ** to promote to an exclusive lock.  The first process cannot proceed
   ** because it is blocked by the second and the second process cannot
   ** proceed because it is blocked by the first.  If both processes
   ** invoke the busy handlers, neither will make any progress.  Therefore,
   ** SQLite returns [SQLITE_BUSY] for the first process, hoping that this
   ** will induce the first process to release its read lock and allow
   ** the second process to proceed.
   **
   ** ^The default busy callback is NULL.
   **
   ** ^The [SQLITE_BUSY] error is converted to [SQLITE_IOERR_BLOCKED]
   ** when SQLite is in the middle of a large transaction where all the
   ** changes will not fit into the in-memory cache.  SQLite will
   ** already hold a RESERVED lock on the database file, but it needs
   ** to promote this lock to EXCLUSIVE so that it can spill cache
   ** pages into the database file without harm to concurrent
   ** readers.  ^If it is unable to promote the lock, then the in-memory
   ** cache will be left in an inconsistent state and so the error
   ** code is promoted from the relatively benign [SQLITE_BUSY] to
   ** the more severe [SQLITE_IOERR_BLOCKED].  ^This error code promotion
   ** forces an automatic rollback of the changes.  See the
   ** <a href="/cvstrac/wiki?p=CorruptionFollowingBusyError">
   ** CorruptionFollowingBusyError</a> wiki page for a discussion of why
   ** this is important.
   **
   ** ^(There can only be a single busy handler defined for each
   ** [database connection].  Setting a new busy handler clears any
   ** previously set handler.)^  ^Note that calling [sqlite3_busy_timeout()]
   ** will also set or clear the busy handler.
   **
   ** The busy callback should not take any actions which modify the
   ** database connection that invoked the busy handler.  Any such actions
   ** result in undefined behavior.
   ** 
   ** A busy handler must not close the database connection
   ** or [prepared statement] that invoked the busy handler.
   */
   SQLITE_API int sqlite3_busy_handler(sqlite3*, int(*)(void*,int), void*);
   
   /*
   ** CAPI3REF: Set A Busy Timeout
   **
   ** ^This routine sets a [sqlite3_busy_handler | busy handler] that sleeps
   ** for a specified amount of time when a table is locked.  ^The handler
   ** will sleep multiple times until at least "ms" milliseconds of sleeping
   ** have accumulated.  ^After at least "ms" milliseconds of sleeping,
   ** the handler returns 0 which causes [sqlite3_step()] to return
   ** [SQLITE_BUSY] or [SQLITE_IOERR_BLOCKED].
   **
   ** ^Calling this routine with an argument less than or equal to zero
   ** turns off all busy handlers.
   **
   ** ^(There can only be a single busy handler for a particular
   ** [database connection] any any given moment.  If another busy handler
   ** was defined  (using [sqlite3_busy_handler()]) prior to calling
   ** this routine, that other busy handler is cleared.)^
   */
   SQLITE_API int sqlite3_busy_timeout(sqlite3*, int ms);
   
   /*
   ** CAPI3REF: Convenience Routines For Running Queries
   **
   ** This is a legacy interface that is preserved for backwards compatibility.
   ** Use of this interface is not recommended.
   **
   ** Definition: A <b>result table</b> is memory data structure created by the
   ** [sqlite3_get_table()] interface.  A result table records the
   ** complete query results from one or more queries.
   **
   ** The table conceptually has a number of rows and columns.  But
   ** these numbers are not part of the result table itself.  These
   ** numbers are obtained separately.  Let N be the number of rows
   ** and M be the number of columns.
   **
   ** A result table is an array of pointers to zero-terminated UTF-8 strings.
   ** There are (N+1)*M elements in the array.  The first M pointers point
   ** to zero-terminated strings that  contain the names of the columns.
   ** The remaining entries all point to query results.  NULL values result
   ** in NULL pointers.  All other values are in their UTF-8 zero-terminated
   ** string representation as returned by [sqlite3_column_text()].
   **
   ** A result table might consist of one or more memory allocations.
   ** It is not safe to pass a result table directly to [sqlite3_free()].
   ** A result table should be deallocated using [sqlite3_free_table()].
   **
   ** ^(As an example of the result table format, suppose a query result
   ** is as follows:
   **
   ** <blockquote><pre>
   **        Name        | Age
   **        -----------------------
   **        Alice       | 43
   **        Bob         | 28
   **        Cindy       | 21
   ** </pre></blockquote>
   **
   ** There are two column (M==2) and three rows (N==3).  Thus the
   ** result table has 8 entries.  Suppose the result table is stored
   ** in an array names azResult.  Then azResult holds this content:
   **
   ** <blockquote><pre>
   **        azResult&#91;0] = "Name";
   **        azResult&#91;1] = "Age";
   **        azResult&#91;2] = "Alice";
   **        azResult&#91;3] = "43";
   **        azResult&#91;4] = "Bob";
   **        azResult&#91;5] = "28";
   **        azResult&#91;6] = "Cindy";
   **        azResult&#91;7] = "21";
   ** </pre></blockquote>)^
   **
   ** ^The sqlite3_get_table() function evaluates one or more
   ** semicolon-separated SQL statements in the zero-terminated UTF-8
   ** string of its 2nd parameter and returns a result table to the
   ** pointer given in its 3rd parameter.
   **
   ** After the application has finished with the result from sqlite3_get_table(),
   ** it must pass the result table pointer to sqlite3_free_table() in order to
   ** release the memory that was malloced.  Because of the way the
   ** [sqlite3_malloc()] happens within sqlite3_get_table(), the calling
   ** function must not try to call [sqlite3_free()] directly.  Only
   ** [sqlite3_free_table()] is able to release the memory properly and safely.
   **
   ** The sqlite3_get_table() interface is implemented as a wrapper around
   ** [sqlite3_exec()].  The sqlite3_get_table() routine does not have access
   ** to any internal data structures of SQLite.  It uses only the public
   ** interface defined here.  As a consequence, errors that occur in the
   ** wrapper layer outside of the internal [sqlite3_exec()] call are not
   ** reflected in subsequent calls to [sqlite3_errcode()] or
   ** [sqlite3_errmsg()].
   */
   SQLITE_API int sqlite3_get_table(
     sqlite3 *db,          /* An open database */
     const char *zSql,     /* SQL to be evaluated */
     char ***pazResult,    /* Results of the query */
     int *pnRow,           /* Number of result rows written here */
     int *pnColumn,        /* Number of result columns written here */
     char **pzErrmsg       /* Error msg written here */
   );
   SQLITE_API void sqlite3_free_table(char **result);
   
   /*
   ** CAPI3REF: Formatted String Printing Functions
   **
   ** These routines are work-alikes of the "printf()" family of functions
   ** from the standard C library.
   **
   ** ^The sqlite3_mprintf() and sqlite3_vmprintf() routines write their
   ** results into memory obtained from [sqlite3_malloc()].
   ** The strings returned by these two routines should be
   ** released by [sqlite3_free()].  ^Both routines return a
   ** NULL pointer if [sqlite3_malloc()] is unable to allocate enough
   ** memory to hold the resulting string.
   **
   ** ^(The sqlite3_snprintf() routine is similar to "snprintf()" from
   ** the standard C library.  The result is written into the
   ** buffer supplied as the second parameter whose size is given by
   ** the first parameter. Note that the order of the
   ** first two parameters is reversed from snprintf().)^  This is an
   ** historical accident that cannot be fixed without breaking
   ** backwards compatibility.  ^(Note also that sqlite3_snprintf()
   ** returns a pointer to its buffer instead of the number of
   ** characters actually written into the buffer.)^  We admit that
   ** the number of characters written would be a more useful return
   ** value but we cannot change the implementation of sqlite3_snprintf()
   ** now without breaking compatibility.
   **
   ** ^As long as the buffer size is greater than zero, sqlite3_snprintf()
   ** guarantees that the buffer is always zero-terminated.  ^The first
   ** parameter "n" is the total size of the buffer, including space for
   ** the zero terminator.  So the longest string that can be completely
   ** written will be n-1 characters.
   **
   ** ^The sqlite3_vsnprintf() routine is a varargs version of sqlite3_snprintf().
   **
   ** These routines all implement some additional formatting
   ** options that are useful for constructing SQL statements.
   ** All of the usual printf() formatting options apply.  In addition, there
   ** is are "%q", "%Q", and "%z" options.
   **
   ** ^(The %q option works like %s in that it substitutes a nul-terminated
   ** string from the argument list.  But %q also doubles every '\'' character.
   ** %q is designed for use inside a string literal.)^  By doubling each '\''
   ** character it escapes that character and allows it to be inserted into
   ** the string.
   **
   ** For example, assume the string variable zText contains text as follows:
   **
   ** <blockquote><pre>
   **  char *zText = "It's a happy day!";
   ** </pre></blockquote>
   **
   ** One can use this text in an SQL statement as follows:
   **
   ** <blockquote><pre>
   **  char *zSQL = sqlite3_mprintf("INSERT INTO table VALUES('%q')", zText);
   **  sqlite3_exec(db, zSQL, 0, 0, 0);
   **  sqlite3_free(zSQL);
   ** </pre></blockquote>
   **
   ** Because the %q format string is used, the '\'' character in zText
   ** is escaped and the SQL generated is as follows:
   **
   ** <blockquote><pre>
   **  INSERT INTO table1 VALUES('It''s a happy day!')
   ** </pre></blockquote>
   **
   ** This is correct.  Had we used %s instead of %q, the generated SQL
   ** would have looked like this:
   **
   ** <blockquote><pre>
   **  INSERT INTO table1 VALUES('It's a happy day!');
   ** </pre></blockquote>
   **
   ** This second example is an SQL syntax error.  As a general rule you should
   ** always use %q instead of %s when inserting text into a string literal.
   **
   ** ^(The %Q option works like %q except it also adds single quotes around
   ** the outside of the total string.  Additionally, if the parameter in the
   ** argument list is a NULL pointer, %Q substitutes the text "NULL" (without
   ** single quotes).)^  So, for example, one could say:
   **
   ** <blockquote><pre>
   **  char *zSQL = sqlite3_mprintf("INSERT INTO table VALUES(%Q)", zText);
   **  sqlite3_exec(db, zSQL, 0, 0, 0);
   **  sqlite3_free(zSQL);
   ** </pre></blockquote>
   **
   ** The code above will render a correct SQL statement in the zSQL
   ** variable even if the zText variable is a NULL pointer.
   **
   ** ^(The "%z" formatting option works like "%s" but with the
   ** addition that after the string has been read and copied into
   ** the result, [sqlite3_free()] is called on the input string.)^
   */
   SQLITE_API char *sqlite3_mprintf(const char*,...);
   SQLITE_API char *sqlite3_vmprintf(const char*, va_list);
   SQLITE_API char *sqlite3_snprintf(int,char*,const char*, ...);
   SQLITE_API char *sqlite3_vsnprintf(int,char*,const char*, va_list);
   
   /*
   ** CAPI3REF: Memory Allocation Subsystem
   **
   ** The SQLite core uses these three routines for all of its own
   ** internal memory allocation needs. "Core" in the previous sentence
   ** does not include operating-system specific VFS implementation.  The
   ** Windows VFS uses native malloc() and free() for some operations.
   **
   ** ^The sqlite3_malloc() routine returns a pointer to a block
   ** of memory at least N bytes in length, where N is the parameter.
   ** ^If sqlite3_malloc() is unable to obtain sufficient free
   ** memory, it returns a NULL pointer.  ^If the parameter N to
   ** sqlite3_malloc() is zero or negative then sqlite3_malloc() returns
   ** a NULL pointer.
   **
   ** ^Calling sqlite3_free() with a pointer previously returned
   ** by sqlite3_malloc() or sqlite3_realloc() releases that memory so
   ** that it might be reused.  ^The sqlite3_free() routine is
   ** a no-op if is called with a NULL pointer.  Passing a NULL pointer
   ** to sqlite3_free() is harmless.  After being freed, memory
   ** should neither be read nor written.  Even reading previously freed
   ** memory might result in a segmentation fault or other severe error.
   ** Memory corruption, a segmentation fault, or other severe error
   ** might result if sqlite3_free() is called with a non-NULL pointer that
   ** was not obtained from sqlite3_malloc() or sqlite3_realloc().
   **
   ** ^(The sqlite3_realloc() interface attempts to resize a
   ** prior memory allocation to be at least N bytes, where N is the
   ** second parameter.  The memory allocation to be resized is the first
   ** parameter.)^ ^ If the first parameter to sqlite3_realloc()
   ** is a NULL pointer then its behavior is identical to calling
   ** sqlite3_malloc(N) where N is the second parameter to sqlite3_realloc().
   ** ^If the second parameter to sqlite3_realloc() is zero or
   ** negative then the behavior is exactly the same as calling
   ** sqlite3_free(P) where P is the first parameter to sqlite3_realloc().
   ** ^sqlite3_realloc() returns a pointer to a memory allocation
   ** of at least N bytes in size or NULL if sufficient memory is unavailable.
   ** ^If M is the size of the prior allocation, then min(N,M) bytes
   ** of the prior allocation are copied into the beginning of buffer returned
   ** by sqlite3_realloc() and the prior allocation is freed.
   ** ^If sqlite3_realloc() returns NULL, then the prior allocation
   ** is not freed.
   **
   ** ^The memory returned by sqlite3_malloc() and sqlite3_realloc()
   ** is always aligned to at least an 8 byte boundary, or to a
   ** 4 byte boundary if the [SQLITE_4_BYTE_ALIGNED_MALLOC] compile-time
   ** option is used.
   **
   ** In SQLite version 3.5.0 and 3.5.1, it was possible to define
   ** the SQLITE_OMIT_MEMORY_ALLOCATION which would cause the built-in
   ** implementation of these routines to be omitted.  That capability
   ** is no longer provided.  Only built-in memory allocators can be used.
   **
   ** The Windows OS interface layer calls
   ** the system malloc() and free() directly when converting
   ** filenames between the UTF-8 encoding used by SQLite
   ** and whatever filename encoding is used by the particular Windows
   ** installation.  Memory allocation errors are detected, but
   ** they are reported back as [SQLITE_CANTOPEN] or
   ** [SQLITE_IOERR] rather than [SQLITE_NOMEM].
   **
   ** The pointer arguments to [sqlite3_free()] and [sqlite3_realloc()]
   ** must be either NULL or else pointers obtained from a prior
   ** invocation of [sqlite3_malloc()] or [sqlite3_realloc()] that have
   ** not yet been released.
   **
   ** The application must not read or write any part of
   ** a block of memory after it has been released using
   ** [sqlite3_free()] or [sqlite3_realloc()].
   */
   SQLITE_API void *sqlite3_malloc(int);
   SQLITE_API void *sqlite3_realloc(void*, int);
   SQLITE_API void sqlite3_free(void*);
   
   /*
   ** CAPI3REF: Memory Allocator Statistics
   **
   ** SQLite provides these two interfaces for reporting on the status
   ** of the [sqlite3_malloc()], [sqlite3_free()], and [sqlite3_realloc()]
   ** routines, which form the built-in memory allocation subsystem.
   **
   ** ^The [sqlite3_memory_used()] routine returns the number of bytes
   ** of memory currently outstanding (malloced but not freed).
   ** ^The [sqlite3_memory_highwater()] routine returns the maximum
   ** value of [sqlite3_memory_used()] since the high-water mark
   ** was last reset.  ^The values returned by [sqlite3_memory_used()] and
   ** [sqlite3_memory_highwater()] include any overhead
   ** added by SQLite in its implementation of [sqlite3_malloc()],
   ** but not overhead added by the any underlying system library
   ** routines that [sqlite3_malloc()] may call.
   **
   ** ^The memory high-water mark is reset to the current value of
   ** [sqlite3_memory_used()] if and only if the parameter to
   ** [sqlite3_memory_highwater()] is true.  ^The value returned
   ** by [sqlite3_memory_highwater(1)] is the high-water mark
   ** prior to the reset.
   */
   SQLITE_API sqlite3_int64 sqlite3_memory_used(void);
   SQLITE_API sqlite3_int64 sqlite3_memory_highwater(int resetFlag);
   
   /*
   ** CAPI3REF: Pseudo-Random Number Generator
   **
   ** SQLite contains a high-quality pseudo-random number generator (PRNG) used to
   ** select random [ROWID | ROWIDs] when inserting new records into a table that
   ** already uses the largest possible [ROWID].  The PRNG is also used for
   ** the build-in random() and randomblob() SQL functions.  This interface allows
   ** applications to access the same PRNG for other purposes.
   **
   ** ^A call to this routine stores N bytes of randomness into buffer P.
   **
   ** ^The first time this routine is invoked (either internally or by
   ** the application) the PRNG is seeded using randomness obtained
   ** from the xRandomness method of the default [sqlite3_vfs] object.
   ** ^On all subsequent invocations, the pseudo-randomness is generated
   ** internally and without recourse to the [sqlite3_vfs] xRandomness
   ** method.
   */
   SQLITE_API void sqlite3_randomness(int N, void *P);
   
   /*
   ** CAPI3REF: Compile-Time Authorization Callbacks
   **
   ** ^This routine registers an authorizer callback with a particular
   ** [database connection], supplied in the first argument.
   ** ^The authorizer callback is invoked as SQL statements are being compiled
   ** by [sqlite3_prepare()] or its variants [sqlite3_prepare_v2()],
   ** [sqlite3_prepare16()] and [sqlite3_prepare16_v2()].  ^At various
   ** points during the compilation process, as logic is being created
   ** to perform various actions, the authorizer callback is invoked to
   ** see if those actions are allowed.  ^The authorizer callback should
   ** return [SQLITE_OK] to allow the action, [SQLITE_IGNORE] to disallow the
   ** specific action but allow the SQL statement to continue to be
   ** compiled, or [SQLITE_DENY] to cause the entire SQL statement to be
   ** rejected with an error.  ^If the authorizer callback returns
   ** any value other than [SQLITE_IGNORE], [SQLITE_OK], or [SQLITE_DENY]
   ** then the [sqlite3_prepare_v2()] or equivalent call that triggered
   ** the authorizer will fail with an error message.
   **
   ** When the callback returns [SQLITE_OK], that means the operation
   ** requested is ok.  ^When the callback returns [SQLITE_DENY], the
   ** [sqlite3_prepare_v2()] or equivalent call that triggered the
   ** authorizer will fail with an error message explaining that
   ** access is denied. 
   **
   ** ^The first parameter to the authorizer callback is a copy of the third
   ** parameter to the sqlite3_set_authorizer() interface. ^The second parameter
   ** to the callback is an integer [SQLITE_COPY | action code] that specifies
   ** the particular action to be authorized. ^The third through sixth parameters
   ** to the callback are zero-terminated strings that contain additional
   ** details about the action to be authorized.
   **
   ** ^If the action code is [SQLITE_READ]
   ** and the callback returns [SQLITE_IGNORE] then the
   ** [prepared statement] statement is constructed to substitute
   ** a NULL value in place of the table column that would have
   ** been read if [SQLITE_OK] had been returned.  The [SQLITE_IGNORE]
   ** return can be used to deny an untrusted user access to individual
   ** columns of a table.
   ** ^If the action code is [SQLITE_DELETE] and the callback returns
   ** [SQLITE_IGNORE] then the [DELETE] operation proceeds but the
   ** [truncate optimization] is disabled and all rows are deleted individually.
   **
   ** An authorizer is used when [sqlite3_prepare | preparing]
   ** SQL statements from an untrusted source, to ensure that the SQL statements
   ** do not try to access data they are not allowed to see, or that they do not
   ** try to execute malicious statements that damage the database.  For
   ** example, an application may allow a user to enter arbitrary
   ** SQL queries for evaluation by a database.  But the application does
   ** not want the user to be able to make arbitrary changes to the
   ** database.  An authorizer could then be put in place while the
   ** user-entered SQL is being [sqlite3_prepare | prepared] that
   ** disallows everything except [SELECT] statements.
   **
   ** Applications that need to process SQL from untrusted sources
   ** might also consider lowering resource limits using [sqlite3_limit()]
   ** and limiting database size using the [max_page_count] [PRAGMA]
   ** in addition to using an authorizer.
   **
   ** ^(Only a single authorizer can be in place on a database connection
   ** at a time.  Each call to sqlite3_set_authorizer overrides the
   ** previous call.)^  ^Disable the authorizer by installing a NULL callback.
   ** The authorizer is disabled by default.
   **
   ** The authorizer callback must not do anything that will modify
   ** the database connection that invoked the authorizer callback.
   ** Note that [sqlite3_prepare_v2()] and [sqlite3_step()] both modify their
   ** database connections for the meaning of "modify" in this paragraph.
   **
   ** ^When [sqlite3_prepare_v2()] is used to prepare a statement, the
   ** statement might be re-prepared during [sqlite3_step()] due to a 
   ** schema change.  Hence, the application should ensure that the
   ** correct authorizer callback remains in place during the [sqlite3_step()].
   **
   ** ^Note that the authorizer callback is invoked only during
   ** [sqlite3_prepare()] or its variants.  Authorization is not
   ** performed during statement evaluation in [sqlite3_step()], unless
   ** as stated in the previous paragraph, sqlite3_step() invokes
   ** sqlite3_prepare_v2() to reprepare a statement after a schema change.
   */
   SQLITE_API int sqlite3_set_authorizer(
     sqlite3*,
     int (*xAuth)(void*,int,const char*,const char*,const char*,const char*),
     void *pUserData
   );
   
   /*
   ** CAPI3REF: Authorizer Return Codes
   **
   ** The [sqlite3_set_authorizer | authorizer callback function] must
   ** return either [SQLITE_OK] or one of these two constants in order
   ** to signal SQLite whether or not the action is permitted.  See the
   ** [sqlite3_set_authorizer | authorizer documentation] for additional
   ** information.
   **
   ** Note that SQLITE_IGNORE is also used as a [SQLITE_ROLLBACK | return code]
   ** from the [sqlite3_vtab_on_conflict()] interface.
   */
   #define SQLITE_DENY   1   /* Abort the SQL statement with an error */
   #define SQLITE_IGNORE 2   /* Don't allow access, but don't generate an error */
   
   /*
   ** CAPI3REF: Authorizer Action Codes
   **
   ** The [sqlite3_set_authorizer()] interface registers a callback function
   ** that is invoked to authorize certain SQL statement actions.  The
   ** second parameter to the callback is an integer code that specifies
   ** what action is being authorized.  These are the integer action codes that
   ** the authorizer callback may be passed.
   **
   ** These action code values signify what kind of operation is to be
   ** authorized.  The 3rd and 4th parameters to the authorization
   ** callback function will be parameters or NULL depending on which of these
   ** codes is used as the second parameter.  ^(The 5th parameter to the
   ** authorizer callback is the name of the database ("main", "temp",
   ** etc.) if applicable.)^  ^The 6th parameter to the authorizer callback
   ** is the name of the inner-most trigger or view that is responsible for
   ** the access attempt or NULL if this access attempt is directly from
   ** top-level SQL code.
   */
   /******************************************* 3rd ************ 4th ***********/
   #define SQLITE_CREATE_INDEX          1   /* Index Name      Table Name      */
   #define SQLITE_CREATE_TABLE          2   /* Table Name      NULL            */
   #define SQLITE_CREATE_TEMP_INDEX     3   /* Index Name      Table Name      */
   #define SQLITE_CREATE_TEMP_TABLE     4   /* Table Name      NULL            */
   #define SQLITE_CREATE_TEMP_TRIGGER   5   /* Trigger Name    Table Name      */
   #define SQLITE_CREATE_TEMP_VIEW      6   /* View Name       NULL            */
   #define SQLITE_CREATE_TRIGGER        7   /* Trigger Name    Table Name      */
   #define SQLITE_CREATE_VIEW           8   /* View Name       NULL            */
   #define SQLITE_DELETE                9   /* Table Name      NULL            */
   #define SQLITE_DROP_INDEX           10   /* Index Name      Table Name      */
   #define SQLITE_DROP_TABLE           11   /* Table Name      NULL            */
   #define SQLITE_DROP_TEMP_INDEX      12   /* Index Name      Table Name      */
   #define SQLITE_DROP_TEMP_TABLE      13   /* Table Name      NULL            */
   #define SQLITE_DROP_TEMP_TRIGGER    14   /* Trigger Name    Table Name      */
   #define SQLITE_DROP_TEMP_VIEW       15   /* View Name       NULL            */
   #define SQLITE_DROP_TRIGGER         16   /* Trigger Name    Table Name      */
   #define SQLITE_DROP_VIEW            17   /* View Name       NULL            */
   #define SQLITE_INSERT               18   /* Table Name      NULL            */
   #define SQLITE_PRAGMA               19   /* Pragma Name     1st arg or NULL */
   #define SQLITE_READ                 20   /* Table Name      Column Name     */
   #define SQLITE_SELECT               21   /* NULL            NULL            */
   #define SQLITE_TRANSACTION          22   /* Operation       NULL            */
   #define SQLITE_UPDATE               23   /* Table Name      Column Name     */
   #define SQLITE_ATTACH               24   /* Filename        NULL            */
   #define SQLITE_DETACH               25   /* Database Name   NULL            */
   #define SQLITE_ALTER_TABLE          26   /* Database Name   Table Name      */
   #define SQLITE_REINDEX              27   /* Index Name      NULL            */
   #define SQLITE_ANALYZE              28   /* Table Name      NULL            */
   #define SQLITE_CREATE_VTABLE        29   /* Table Name      Module Name     */
   #define SQLITE_DROP_VTABLE          30   /* Table Name      Module Name     */
   #define SQLITE_FUNCTION             31   /* NULL            Function Name   */
   #define SQLITE_SAVEPOINT            32   /* Operation       Savepoint Name  */
   #define SQLITE_COPY                  0   /* No longer used */
   
   /*
   ** CAPI3REF: Tracing And Profiling Functions
   **
   ** These routines register callback functions that can be used for
   ** tracing and profiling the execution of SQL statements.
   **
   ** ^The callback function registered by sqlite3_trace() is invoked at
   ** various times when an SQL statement is being run by [sqlite3_step()].
   ** ^The sqlite3_trace() callback is invoked with a UTF-8 rendering of the
   ** SQL statement text as the statement first begins executing.
   ** ^(Additional sqlite3_trace() callbacks might occur
   ** as each triggered subprogram is entered.  The callbacks for triggers
   ** contain a UTF-8 SQL comment that identifies the trigger.)^
   **
   ** ^The callback function registered by sqlite3_profile() is invoked
   ** as each SQL statement finishes.  ^The profile callback contains
   ** the original statement text and an estimate of wall-clock time
   ** of how long that statement took to run.  ^The profile callback
   ** time is in units of nanoseconds, however the current implementation
   ** is only capable of millisecond resolution so the six least significant
   ** digits in the time are meaningless.  Future versions of SQLite
   ** might provide greater resolution on the profiler callback.  The
   ** sqlite3_profile() function is considered experimental and is
   ** subject to change in future versions of SQLite.
   */
   SQLITE_API void *sqlite3_trace(sqlite3*, void(*xTrace)(void*,const char*), void*);
   SQLITE_API SQLITE_EXPERIMENTAL void *sqlite3_profile(sqlite3*,
      void(*xProfile)(void*,const char*,sqlite3_uint64), void*);
   
   /*
   ** CAPI3REF: Query Progress Callbacks
   **
   ** ^The sqlite3_progress_handler(D,N,X,P) interface causes the callback
   ** function X to be invoked periodically during long running calls to
   ** [sqlite3_exec()], [sqlite3_step()] and [sqlite3_get_table()] for
   ** database connection D.  An example use for this
   ** interface is to keep a GUI updated during a large query.
   **
   ** ^The parameter P is passed through as the only parameter to the 
   ** callback function X.  ^The parameter N is the number of 
   ** [virtual machine instructions] that are evaluated between successive
   ** invocations of the callback X.
   **
   ** ^Only a single progress handler may be defined at one time per
   ** [database connection]; setting a new progress handler cancels the
   ** old one.  ^Setting parameter X to NULL disables the progress handler.
   ** ^The progress handler is also disabled by setting N to a value less
   ** than 1.
   **
   ** ^If the progress callback returns non-zero, the operation is
   ** interrupted.  This feature can be used to implement a
   ** "Cancel" button on a GUI progress dialog box.
   **
   ** The progress handler callback must not do anything that will modify
   ** the database connection that invoked the progress handler.
   ** Note that [sqlite3_prepare_v2()] and [sqlite3_step()] both modify their
   ** database connections for the meaning of "modify" in this paragraph.
   **
   */
   SQLITE_API void sqlite3_progress_handler(sqlite3*, int, int(*)(void*), void*);
   
   /*
   ** CAPI3REF: Opening A New Database Connection
   **
   ** ^These routines open an SQLite database file as specified by the 
   ** filename argument. ^The filename argument is interpreted as UTF-8 for
   ** sqlite3_open() and sqlite3_open_v2() and as UTF-16 in the native byte
   ** order for sqlite3_open16(). ^(A [database connection] handle is usually
   ** returned in *ppDb, even if an error occurs.  The only exception is that
   ** if SQLite is unable to allocate memory to hold the [sqlite3] object,
   ** a NULL will be written into *ppDb instead of a pointer to the [sqlite3]
   ** object.)^ ^(If the database is opened (and/or created) successfully, then
   ** [SQLITE_OK] is returned.  Otherwise an [error code] is returned.)^ ^The
   ** [sqlite3_errmsg()] or [sqlite3_errmsg16()] routines can be used to obtain
   ** an English language description of the error following a failure of any
   ** of the sqlite3_open() routines.
   **
   ** ^The default encoding for the database will be UTF-8 if
   ** sqlite3_open() or sqlite3_open_v2() is called and
   ** UTF-16 in the native byte order if sqlite3_open16() is used.
   **
   ** Whether or not an error occurs when it is opened, resources
   ** associated with the [database connection] handle should be released by
   ** passing it to [sqlite3_close()] when it is no longer required.
   **
   ** The sqlite3_open_v2() interface works like sqlite3_open()
   ** except that it accepts two additional parameters for additional control
   ** over the new database connection.  ^(The flags parameter to
   ** sqlite3_open_v2() can take one of
   ** the following three values, optionally combined with the 
   ** [SQLITE_OPEN_NOMUTEX], [SQLITE_OPEN_FULLMUTEX], [SQLITE_OPEN_SHAREDCACHE],
   ** [SQLITE_OPEN_PRIVATECACHE], and/or [SQLITE_OPEN_URI] flags:)^
   **
   ** <dl>
   ** ^(<dt>[SQLITE_OPEN_READONLY]</dt>
   ** <dd>The database is opened in read-only mode.  If the database does not
   ** already exist, an error is returned.</dd>)^
   **
   ** ^(<dt>[SQLITE_OPEN_READWRITE]</dt>
   ** <dd>The database is opened for reading and writing if possible, or reading
   ** only if the file is write protected by the operating system.  In either
   ** case the database must already exist, otherwise an error is returned.</dd>)^
   **
   ** ^(<dt>[SQLITE_OPEN_READWRITE] | [SQLITE_OPEN_CREATE]</dt>
   ** <dd>The database is opened for reading and writing, and is created if
   ** it does not already exist. This is the behavior that is always used for
   ** sqlite3_open() and sqlite3_open16().</dd>)^
   ** </dl>
   **
   ** If the 3rd parameter to sqlite3_open_v2() is not one of the
   ** combinations shown above optionally combined with other
   ** [SQLITE_OPEN_READONLY | SQLITE_OPEN_* bits]
   ** then the behavior is undefined.
   **
   ** ^If the [SQLITE_OPEN_NOMUTEX] flag is set, then the database connection
   ** opens in the multi-thread [threading mode] as long as the single-thread
   ** mode has not been set at compile-time or start-time.  ^If the
   ** [SQLITE_OPEN_FULLMUTEX] flag is set then the database connection opens
   ** in the serialized [threading mode] unless single-thread was
   ** previously selected at compile-time or start-time.
   ** ^The [SQLITE_OPEN_SHAREDCACHE] flag causes the database connection to be
   ** eligible to use [shared cache mode], regardless of whether or not shared
   ** cache is enabled using [sqlite3_enable_shared_cache()].  ^The
   ** [SQLITE_OPEN_PRIVATECACHE] flag causes the database connection to not
   ** participate in [shared cache mode] even if it is enabled.
   **
   ** ^The fourth parameter to sqlite3_open_v2() is the name of the
   ** [sqlite3_vfs] object that defines the operating system interface that
   ** the new database connection should use.  ^If the fourth parameter is
   ** a NULL pointer then the default [sqlite3_vfs] object is used.
   **
   ** ^If the filename is ":memory:", then a private, temporary in-memory database
   ** is created for the connection.  ^This in-memory database will vanish when
   ** the database connection is closed.  Future versions of SQLite might
   ** make use of additional special filenames that begin with the ":" character.
   ** It is recommended that when a database filename actually does begin with
   ** a ":" character you should prefix the filename with a pathname such as
   ** "./" to avoid ambiguity.
   **
   ** ^If the filename is an empty string, then a private, temporary
   ** on-disk database will be created.  ^This private database will be
   ** automatically deleted as soon as the database connection is closed.
   **
   ** [[URI filenames in sqlite3_open()]] <h3>URI Filenames</h3>
   **
   ** ^If [URI filename] interpretation is enabled, and the filename argument
   ** begins with "file:", then the filename is interpreted as a URI. ^URI
   ** filename interpretation is enabled if the [SQLITE_OPEN_URI] flag is
   ** set in the fourth argument to sqlite3_open_v2(), or if it has
   ** been enabled globally using the [SQLITE_CONFIG_URI] option with the
   ** [sqlite3_config()] method or by the [SQLITE_USE_URI] compile-time option.
   ** As of SQLite version 3.7.7, URI filename interpretation is turned off
   ** by default, but future releases of SQLite might enable URI filename
   ** interpretation by default.  See "[URI filenames]" for additional
   ** information.
   **
   ** URI filenames are parsed according to RFC 3986. ^If the URI contains an
   ** authority, then it must be either an empty string or the string 
   ** "localhost". ^If the authority is not an empty string or "localhost", an 
   ** error is returned to the caller. ^The fragment component of a URI, if 
   ** present, is ignored.
   **
   ** ^SQLite uses the path component of the URI as the name of the disk file
   ** which contains the database. ^If the path begins with a '/' character, 
   ** then it is interpreted as an absolute path. ^If the path does not begin 
   ** with a '/' (meaning that the authority section is omitted from the URI)
   ** then the path is interpreted as a relative path. 
   ** ^On windows, the first component of an absolute path 
   ** is a drive specification (e.g. "C:").
   **
   ** [[core URI query parameters]]
   ** The query component of a URI may contain parameters that are interpreted
   ** either by SQLite itself, or by a [VFS | custom VFS implementation].
   ** SQLite interprets the following three query parameters:
   **
   ** <ul>
   **   <li> <b>vfs</b>: ^The "vfs" parameter may be used to specify the name of
   **     a VFS object that provides the operating system interface that should
   **     be used to access the database file on disk. ^If this option is set to
   **     an empty string the default VFS object is used. ^Specifying an unknown
   **     VFS is an error. ^If sqlite3_open_v2() is used and the vfs option is
   **     present, then the VFS specified by the option takes precedence over
   **     the value passed as the fourth parameter to sqlite3_open_v2().
   **
   **   <li> <b>mode</b>: ^(The mode parameter may be set to either "ro", "rw" or
   **     "rwc". Attempting to set it to any other value is an error)^. 
   **     ^If "ro" is specified, then the database is opened for read-only 
   **     access, just as if the [SQLITE_OPEN_READONLY] flag had been set in the 
   **     third argument to sqlite3_prepare_v2(). ^If the mode option is set to 
   **     "rw", then the database is opened for read-write (but not create) 
   **     access, as if SQLITE_OPEN_READWRITE (but not SQLITE_OPEN_CREATE) had 
   **     been set. ^Value "rwc" is equivalent to setting both 
   **     SQLITE_OPEN_READWRITE and SQLITE_OPEN_CREATE. ^If sqlite3_open_v2() is 
   **     used, it is an error to specify a value for the mode parameter that is 
   **     less restrictive than that specified by the flags passed as the third 
   **     parameter.
   **
   **   <li> <b>cache</b>: ^The cache parameter may be set to either "shared" or
   **     "private". ^Setting it to "shared" is equivalent to setting the
   **     SQLITE_OPEN_SHAREDCACHE bit in the flags argument passed to
   **     sqlite3_open_v2(). ^Setting the cache parameter to "private" is 
   **     equivalent to setting the SQLITE_OPEN_PRIVATECACHE bit.
   **     ^If sqlite3_open_v2() is used and the "cache" parameter is present in
   **     a URI filename, its value overrides any behaviour requested by setting
   **     SQLITE_OPEN_PRIVATECACHE or SQLITE_OPEN_SHAREDCACHE flag.
   ** </ul>
   **
   ** ^Specifying an unknown parameter in the query component of a URI is not an
   ** error.  Future versions of SQLite might understand additional query
   ** parameters.  See "[query parameters with special meaning to SQLite]" for
   ** additional information.
   **
   ** [[URI filename examples]] <h3>URI filename examples</h3>
   **
   ** <table border="1" align=center cellpadding=5>
   ** <tr><th> URI filenames <th> Results
   ** <tr><td> file:data.db <td> 
   **          Open the file "data.db" in the current directory.
   ** <tr><td> file:/home/fred/data.db<br>
   **          file:///home/fred/data.db <br> 
   **          file://localhost/home/fred/data.db <br> <td> 
   **          Open the database file "/home/fred/data.db".
   ** <tr><td> file://darkstar/home/fred/data.db <td> 
   **          An error. "darkstar" is not a recognized authority.
   ** <tr><td style="white-space:nowrap"> 
   **          file:///C:/Documents%20and%20Settings/fred/Desktop/data.db
   **     <td> Windows only: Open the file "data.db" on fred's desktop on drive
   **          C:. Note that the %20 escaping in this example is not strictly 
   **          necessary - space characters can be used literally
   **          in URI filenames.
   ** <tr><td> file:data.db?mode=ro&cache=private <td> 
   **          Open file "data.db" in the current directory for read-only access.
   **          Regardless of whether or not shared-cache mode is enabled by
   **          default, use a private cache.
   ** <tr><td> file:/home/fred/data.db?vfs=unix-nolock <td>
   **          Open file "/home/fred/data.db". Use the special VFS "unix-nolock".
   ** <tr><td> file:data.db?mode=readonly <td> 
   **          An error. "readonly" is not a valid option for the "mode" parameter.
   ** </table>
   **
   ** ^URI hexadecimal escape sequences (%HH) are supported within the path and
   ** query components of a URI. A hexadecimal escape sequence consists of a
   ** percent sign - "%" - followed by exactly two hexadecimal digits 
   ** specifying an octet value. ^Before the path or query components of a
   ** URI filename are interpreted, they are encoded using UTF-8 and all 
   ** hexadecimal escape sequences replaced by a single byte containing the
   ** corresponding octet. If this process generates an invalid UTF-8 encoding,
   ** the results are undefined.
   **
   ** <b>Note to Windows users:</b>  The encoding used for the filename argument
   ** of sqlite3_open() and sqlite3_open_v2() must be UTF-8, not whatever
   ** codepage is currently defined.  Filenames containing international
   ** characters must be converted to UTF-8 prior to passing them into
   ** sqlite3_open() or sqlite3_open_v2().
   */
   SQLITE_API int sqlite3_open(
     const char *filename,   /* Database filename (UTF-8) */
     sqlite3 **ppDb          /* OUT: SQLite db handle */
   );
   SQLITE_API int sqlite3_open16(
     const void *filename,   /* Database filename (UTF-16) */
     sqlite3 **ppDb          /* OUT: SQLite db handle */
   );
   SQLITE_API int sqlite3_open_v2(
     const char *filename,   /* Database filename (UTF-8) */
     sqlite3 **ppDb,         /* OUT: SQLite db handle */
     int flags,              /* Flags */
     const char *zVfs        /* Name of VFS module to use */
   );
   
   /*
   ** CAPI3REF: Obtain Values For URI Parameters
   **
   ** These are utility routines, useful to VFS implementations, that check
   ** to see if a database file was a URI that contained a specific query 
   ** parameter, and if so obtains the value of that query parameter.
   **
   ** If F is the database filename pointer passed into the xOpen() method of 
   ** a VFS implementation when the flags parameter to xOpen() has one or 
   ** more of the [SQLITE_OPEN_URI] or [SQLITE_OPEN_MAIN_DB] bits set and
   ** P is the name of the query parameter, then
   ** sqlite3_uri_parameter(F,P) returns the value of the P
   ** parameter if it exists or a NULL pointer if P does not appear as a 
   ** query parameter on F.  If P is a query parameter of F
   ** has no explicit value, then sqlite3_uri_parameter(F,P) returns
   ** a pointer to an empty string.
   **
   ** The sqlite3_uri_boolean(F,P,B) routine assumes that P is a boolean
   ** parameter and returns true (1) or false (0) according to the value
   ** of P.  The value of P is true if it is "yes" or "true" or "on" or 
   ** a non-zero number and is false otherwise.  If P is not a query parameter
   ** on F then sqlite3_uri_boolean(F,P,B) returns (B!=0).
   **
   ** The sqlite3_uri_int64(F,P,D) routine converts the value of P into a
   ** 64-bit signed integer and returns that integer, or D if P does not
   ** exist.  If the value of P is something other than an integer, then
   ** zero is returned.
   ** 
   ** If F is a NULL pointer, then sqlite3_uri_parameter(F,P) returns NULL and
   ** sqlite3_uri_boolean(F,P,B) returns B.  If F is not a NULL pointer and
   ** is not a database file pathname pointer that SQLite passed into the xOpen
   ** VFS method, then the behavior of this routine is undefined and probably
   ** undesirable.
   */
   SQLITE_API const char *sqlite3_uri_parameter(const char *zFilename, const char *zParam);
   SQLITE_API int sqlite3_uri_boolean(const char *zFile, const char *zParam, int bDefault);
   SQLITE_API sqlite3_int64 sqlite3_uri_int64(const char*, const char*, sqlite3_int64);
   
   
   /*
   ** CAPI3REF: Error Codes And Messages
   **
   ** ^The sqlite3_errcode() interface returns the numeric [result code] or
   ** [extended result code] for the most recent failed sqlite3_* API call
   ** associated with a [database connection]. If a prior API call failed
   ** but the most recent API call succeeded, the return value from
   ** sqlite3_errcode() is undefined.  ^The sqlite3_extended_errcode()
   ** interface is the same except that it always returns the 
   ** [extended result code] even when extended result codes are
   ** disabled.
   **
   ** ^The sqlite3_errmsg() and sqlite3_errmsg16() return English-language
   ** text that describes the error, as either UTF-8 or UTF-16 respectively.
   ** ^(Memory to hold the error message string is managed internally.
   ** The application does not need to worry about freeing the result.
   ** However, the error string might be overwritten or deallocated by
   ** subsequent calls to other SQLite interface functions.)^
   **
   ** When the serialized [threading mode] is in use, it might be the
   ** case that a second error occurs on a separate thread in between
   ** the time of the first error and the call to these interfaces.
   ** When that happens, the second error will be reported since these
   ** interfaces always report the most recent result.  To avoid
   ** this, each thread can obtain exclusive use of the [database connection] D
   ** by invoking [sqlite3_mutex_enter]([sqlite3_db_mutex](D)) before beginning
   ** to use D and invoking [sqlite3_mutex_leave]([sqlite3_db_mutex](D)) after
   ** all calls to the interfaces listed here are completed.
   **
   ** If an interface fails with SQLITE_MISUSE, that means the interface
   ** was invoked incorrectly by the application.  In that case, the
   ** error code and message may or may not be set.
   */
   SQLITE_API int sqlite3_errcode(sqlite3 *db);
   SQLITE_API int sqlite3_extended_errcode(sqlite3 *db);
   SQLITE_API const char *sqlite3_errmsg(sqlite3*);
   SQLITE_API const void *sqlite3_errmsg16(sqlite3*);
   
   /*
   ** CAPI3REF: SQL Statement Object
   ** KEYWORDS: {prepared statement} {prepared statements}
   **
   ** An instance of this object represents a single SQL statement.
   ** This object is variously known as a "prepared statement" or a
   ** "compiled SQL statement" or simply as a "statement".
   **
   ** The life of a statement object goes something like this:
   **
   ** <ol>
   ** <li> Create the object using [sqlite3_prepare_v2()] or a related
   **      function.
   ** <li> Bind values to [host parameters] using the sqlite3_bind_*()
   **      interfaces.
   ** <li> Run the SQL by calling [sqlite3_step()] one or more times.
   ** <li> Reset the statement using [sqlite3_reset()] then go back
   **      to step 2.  Do this zero or more times.
   ** <li> Destroy the object using [sqlite3_finalize()].
   ** </ol>
   **
   ** Refer to documentation on individual methods above for additional
   ** information.
   */
   typedef struct sqlite3_stmt sqlite3_stmt;
   
   /*
   ** CAPI3REF: Run-time Limits
   **
   ** ^(This interface allows the size of various constructs to be limited
   ** on a connection by connection basis.  The first parameter is the
   ** [database connection] whose limit is to be set or queried.  The
   ** second parameter is one of the [limit categories] that define a
   ** class of constructs to be size limited.  The third parameter is the
   ** new limit for that construct.)^
   **
   ** ^If the new limit is a negative number, the limit is unchanged.
   ** ^(For each limit category SQLITE_LIMIT_<i>NAME</i> there is a 
   ** [limits | hard upper bound]
   ** set at compile-time by a C preprocessor macro called
   ** [limits | SQLITE_MAX_<i>NAME</i>].
   ** (The "_LIMIT_" in the name is changed to "_MAX_".))^
   ** ^Attempts to increase a limit above its hard upper bound are
   ** silently truncated to the hard upper bound.
   **
   ** ^Regardless of whether or not the limit was changed, the 
   ** [sqlite3_limit()] interface returns the prior value of the limit.
   ** ^Hence, to find the current value of a limit without changing it,
   ** simply invoke this interface with the third parameter set to -1.
   **
   ** Run-time limits are intended for use in applications that manage
   ** both their own internal database and also databases that are controlled
   ** by untrusted external sources.  An example application might be a
   ** web browser that has its own databases for storing history and
   ** separate databases controlled by JavaScript applications downloaded
   ** off the Internet.  The internal databases can be given the
   ** large, default limits.  Databases managed by external sources can
   ** be given much smaller limits designed to prevent a denial of service
   ** attack.  Developers might also want to use the [sqlite3_set_authorizer()]
   ** interface to further control untrusted SQL.  The size of the database
   ** created by an untrusted script can be contained using the
   ** [max_page_count] [PRAGMA].
   **
   ** New run-time limit categories may be added in future releases.
   */
   SQLITE_API int sqlite3_limit(sqlite3*, int id, int newVal);
   
   /*
   ** CAPI3REF: Run-Time Limit Categories
   ** KEYWORDS: {limit category} {*limit categories}
   **
   ** These constants define various performance limits
   ** that can be lowered at run-time using [sqlite3_limit()].
   ** The synopsis of the meanings of the various limits is shown below.
   ** Additional information is available at [limits | Limits in SQLite].
   **
   ** <dl>
   ** [[SQLITE_LIMIT_LENGTH]] ^(<dt>SQLITE_LIMIT_LENGTH</dt>
   ** <dd>The maximum size of any string or BLOB or table row, in bytes.<dd>)^
   **
   ** [[SQLITE_LIMIT_SQL_LENGTH]] ^(<dt>SQLITE_LIMIT_SQL_LENGTH</dt>
   ** <dd>The maximum length of an SQL statement, in bytes.</dd>)^
   **
   ** [[SQLITE_LIMIT_COLUMN]] ^(<dt>SQLITE_LIMIT_COLUMN</dt>
   ** <dd>The maximum number of columns in a table definition or in the
   ** result set of a [SELECT] or the maximum number of columns in an index
   ** or in an ORDER BY or GROUP BY clause.</dd>)^
   **
   ** [[SQLITE_LIMIT_EXPR_DEPTH]] ^(<dt>SQLITE_LIMIT_EXPR_DEPTH</dt>
   ** <dd>The maximum depth of the parse tree on any expression.</dd>)^
   **
   ** [[SQLITE_LIMIT_COMPOUND_SELECT]] ^(<dt>SQLITE_LIMIT_COMPOUND_SELECT</dt>
   ** <dd>The maximum number of terms in a compound SELECT statement.</dd>)^
   **
   ** [[SQLITE_LIMIT_VDBE_OP]] ^(<dt>SQLITE_LIMIT_VDBE_OP</dt>
   ** <dd>The maximum number of instructions in a virtual machine program
   ** used to implement an SQL statement.  This limit is not currently
   ** enforced, though that might be added in some future release of
   ** SQLite.</dd>)^
   **
   ** [[SQLITE_LIMIT_FUNCTION_ARG]] ^(<dt>SQLITE_LIMIT_FUNCTION_ARG</dt>
   ** <dd>The maximum number of arguments on a function.</dd>)^
   **
   ** [[SQLITE_LIMIT_ATTACHED]] ^(<dt>SQLITE_LIMIT_ATTACHED</dt>
   ** <dd>The maximum number of [ATTACH | attached databases].)^</dd>
   **
   ** [[SQLITE_LIMIT_LIKE_PATTERN_LENGTH]]
   ** ^(<dt>SQLITE_LIMIT_LIKE_PATTERN_LENGTH</dt>
   ** <dd>The maximum length of the pattern argument to the [LIKE] or
   ** [GLOB] operators.</dd>)^
   **
   ** [[SQLITE_LIMIT_VARIABLE_NUMBER]]
   ** ^(<dt>SQLITE_LIMIT_VARIABLE_NUMBER</dt>
   ** <dd>The maximum index number of any [parameter] in an SQL statement.)^
   **
   ** [[SQLITE_LIMIT_TRIGGER_DEPTH]] ^(<dt>SQLITE_LIMIT_TRIGGER_DEPTH</dt>
   ** <dd>The maximum depth of recursion for triggers.</dd>)^
   ** </dl>
   */
   #define SQLITE_LIMIT_LENGTH                    0
   #define SQLITE_LIMIT_SQL_LENGTH                1
   #define SQLITE_LIMIT_COLUMN                    2
   #define SQLITE_LIMIT_EXPR_DEPTH                3
   #define SQLITE_LIMIT_COMPOUND_SELECT           4
   #define SQLITE_LIMIT_VDBE_OP                   5
   #define SQLITE_LIMIT_FUNCTION_ARG              6
   #define SQLITE_LIMIT_ATTACHED                  7
   #define SQLITE_LIMIT_LIKE_PATTERN_LENGTH       8
   #define SQLITE_LIMIT_VARIABLE_NUMBER           9
   #define SQLITE_LIMIT_TRIGGER_DEPTH            10
   
   /*
   ** CAPI3REF: Compiling An SQL Statement
   ** KEYWORDS: {SQL statement compiler}
   **
   ** To execute an SQL query, it must first be compiled into a byte-code
   ** program using one of these routines.
   **
   ** The first argument, "db", is a [database connection] obtained from a
   ** prior successful call to [sqlite3_open()], [sqlite3_open_v2()] or
   ** [sqlite3_open16()].  The database connection must not have been closed.
   **
   ** The second argument, "zSql", is the statement to be compiled, encoded
   ** as either UTF-8 or UTF-16.  The sqlite3_prepare() and sqlite3_prepare_v2()
   ** interfaces use UTF-8, and sqlite3_prepare16() and sqlite3_prepare16_v2()
   ** use UTF-16.
   **
   ** ^If the nByte argument is less than zero, then zSql is read up to the
   ** first zero terminator. ^If nByte is non-negative, then it is the maximum
   ** number of  bytes read from zSql.  ^When nByte is non-negative, the
   ** zSql string ends at either the first '\000' or '\u0000' character or
   ** the nByte-th byte, whichever comes first. If the caller knows
   ** that the supplied string is nul-terminated, then there is a small
   ** performance advantage to be gained by passing an nByte parameter that
   ** is equal to the number of bytes in the input string <i>including</i>
   ** the nul-terminator bytes as this saves SQLite from having to
   ** make a copy of the input string.
   **
   ** ^If pzTail is not NULL then *pzTail is made to point to the first byte
   ** past the end of the first SQL statement in zSql.  These routines only
   ** compile the first statement in zSql, so *pzTail is left pointing to
   ** what remains uncompiled.
   **
   ** ^*ppStmt is left pointing to a compiled [prepared statement] that can be
   ** executed using [sqlite3_step()].  ^If there is an error, *ppStmt is set
   ** to NULL.  ^If the input text contains no SQL (if the input is an empty
   ** string or a comment) then *ppStmt is set to NULL.
   ** The calling procedure is responsible for deleting the compiled
   ** SQL statement using [sqlite3_finalize()] after it has finished with it.
   ** ppStmt may not be NULL.
   **
   ** ^On success, the sqlite3_prepare() family of routines return [SQLITE_OK];
   ** otherwise an [error code] is returned.
   **
   ** The sqlite3_prepare_v2() and sqlite3_prepare16_v2() interfaces are
   ** recommended for all new programs. The two older interfaces are retained
   ** for backwards compatibility, but their use is discouraged.
   ** ^In the "v2" interfaces, the prepared statement
   ** that is returned (the [sqlite3_stmt] object) contains a copy of the
   ** original SQL text. This causes the [sqlite3_step()] interface to
   ** behave differently in three ways:
   **
   ** <ol>
   ** <li>
   ** ^If the database schema changes, instead of returning [SQLITE_SCHEMA] as it
   ** always used to do, [sqlite3_step()] will automatically recompile the SQL
   ** statement and try to run it again.
   ** </li>
   **
   ** <li>
   ** ^When an error occurs, [sqlite3_step()] will return one of the detailed
   ** [error codes] or [extended error codes].  ^The legacy behavior was that
   ** [sqlite3_step()] would only return a generic [SQLITE_ERROR] result code
   ** and the application would have to make a second call to [sqlite3_reset()]
   ** in order to find the underlying cause of the problem. With the "v2" prepare
   ** interfaces, the underlying reason for the error is returned immediately.
   ** </li>
   **
   ** <li>
   ** ^If the specific value bound to [parameter | host parameter] in the 
   ** WHERE clause might influence the choice of query plan for a statement,
   ** then the statement will be automatically recompiled, as if there had been 
   ** a schema change, on the first  [sqlite3_step()] call following any change
   ** to the [sqlite3_bind_text | bindings] of that [parameter]. 
   ** ^The specific value of WHERE-clause [parameter] might influence the 
   ** choice of query plan if the parameter is the left-hand side of a [LIKE]
   ** or [GLOB] operator or if the parameter is compared to an indexed column
   ** and the [SQLITE_ENABLE_STAT3] compile-time option is enabled.
   ** the 
   ** </li>
   ** </ol>
   */
   SQLITE_API int sqlite3_prepare(
     sqlite3 *db,            /* Database handle */
     const char *zSql,       /* SQL statement, UTF-8 encoded */
     int nByte,              /* Maximum length of zSql in bytes. */
     sqlite3_stmt **ppStmt,  /* OUT: Statement handle */
     const char **pzTail     /* OUT: Pointer to unused portion of zSql */
   );
   SQLITE_API int sqlite3_prepare_v2(
     sqlite3 *db,            /* Database handle */
     const char *zSql,       /* SQL statement, UTF-8 encoded */
     int nByte,              /* Maximum length of zSql in bytes. */
     sqlite3_stmt **ppStmt,  /* OUT: Statement handle */
     const char **pzTail     /* OUT: Pointer to unused portion of zSql */
   );
   SQLITE_API int sqlite3_prepare16(
     sqlite3 *db,            /* Database handle */
     const void *zSql,       /* SQL statement, UTF-16 encoded */
     int nByte,              /* Maximum length of zSql in bytes. */
     sqlite3_stmt **ppStmt,  /* OUT: Statement handle */
     const void **pzTail     /* OUT: Pointer to unused portion of zSql */
   );
   SQLITE_API int sqlite3_prepare16_v2(
     sqlite3 *db,            /* Database handle */
     const void *zSql,       /* SQL statement, UTF-16 encoded */
     int nByte,              /* Maximum length of zSql in bytes. */
     sqlite3_stmt **ppStmt,  /* OUT: Statement handle */
     const void **pzTail     /* OUT: Pointer to unused portion of zSql */
   );
   
   /*
   ** CAPI3REF: Retrieving Statement SQL
   **
   ** ^This interface can be used to retrieve a saved copy of the original
   ** SQL text used to create a [prepared statement] if that statement was
   ** compiled using either [sqlite3_prepare_v2()] or [sqlite3_prepare16_v2()].
   */
   SQLITE_API const char *sqlite3_sql(sqlite3_stmt *pStmt);
   
   /*
   ** CAPI3REF: Determine If An SQL Statement Writes The Database
   **
   ** ^The sqlite3_stmt_readonly(X) interface returns true (non-zero) if
   ** and only if the [prepared statement] X makes no direct changes to
   ** the content of the database file.
   **
   ** Note that [application-defined SQL functions] or
   ** [virtual tables] might change the database indirectly as a side effect.  
   ** ^(For example, if an application defines a function "eval()" that 
   ** calls [sqlite3_exec()], then the following SQL statement would
   ** change the database file through side-effects:
   **
   ** <blockquote><pre>
   **    SELECT eval('DELETE FROM t1') FROM t2;
   ** </pre></blockquote>
   **
   ** But because the [SELECT] statement does not change the database file
   ** directly, sqlite3_stmt_readonly() would still return true.)^
   **
   ** ^Transaction control statements such as [BEGIN], [COMMIT], [ROLLBACK],
   ** [SAVEPOINT], and [RELEASE] cause sqlite3_stmt_readonly() to return true,
   ** since the statements themselves do not actually modify the database but
   ** rather they control the timing of when other statements modify the 
   ** database.  ^The [ATTACH] and [DETACH] statements also cause
   ** sqlite3_stmt_readonly() to return true since, while those statements
   ** change the configuration of a database connection, they do not make 
   ** changes to the content of the database files on disk.
   */
   SQLITE_API int sqlite3_stmt_readonly(sqlite3_stmt *pStmt);
   
   /*
   ** CAPI3REF: Determine If A Prepared Statement Has Been Reset
   **
   ** ^The sqlite3_stmt_busy(S) interface returns true (non-zero) if the
   ** [prepared statement] S has been stepped at least once using 
   ** [sqlite3_step(S)] but has not run to completion and/or has not 
   ** been reset using [sqlite3_reset(S)].  ^The sqlite3_stmt_busy(S)
   ** interface returns false if S is a NULL pointer.  If S is not a 
   ** NULL pointer and is not a pointer to a valid [prepared statement]
   ** object, then the behavior is undefined and probably undesirable.
   **
   ** This interface can be used in combination [sqlite3_next_stmt()]
   ** to locate all prepared statements associated with a database 
   ** connection that are in need of being reset.  This can be used,
   ** for example, in diagnostic routines to search for prepared 
   ** statements that are holding a transaction open.
   */
   SQLITE_API int sqlite3_stmt_busy(sqlite3_stmt*);
   
   /*
   ** CAPI3REF: Dynamically Typed Value Object
   ** KEYWORDS: {protected sqlite3_value} {unprotected sqlite3_value}
   **
   ** SQLite uses the sqlite3_value object to represent all values
   ** that can be stored in a database table. SQLite uses dynamic typing
   ** for the values it stores.  ^Values stored in sqlite3_value objects
   ** can be integers, floating point values, strings, BLOBs, or NULL.
   **
   ** An sqlite3_value object may be either "protected" or "unprotected".
   ** Some interfaces require a protected sqlite3_value.  Other interfaces
   ** will accept either a protected or an unprotected sqlite3_value.
   ** Every interface that accepts sqlite3_value arguments specifies
   ** whether or not it requires a protected sqlite3_value.
   **
   ** The terms "protected" and "unprotected" refer to whether or not
   ** a mutex is held.  An internal mutex is held for a protected
   ** sqlite3_value object but no mutex is held for an unprotected
   ** sqlite3_value object.  If SQLite is compiled to be single-threaded
   ** (with [SQLITE_THREADSAFE=0] and with [sqlite3_threadsafe()] returning 0)
   ** or if SQLite is run in one of reduced mutex modes 
   ** [SQLITE_CONFIG_SINGLETHREAD] or [SQLITE_CONFIG_MULTITHREAD]
   ** then there is no distinction between protected and unprotected
   ** sqlite3_value objects and they can be used interchangeably.  However,
   ** for maximum code portability it is recommended that applications
   ** still make the distinction between protected and unprotected
   ** sqlite3_value objects even when not strictly required.
   **
   ** ^The sqlite3_value objects that are passed as parameters into the
   ** implementation of [application-defined SQL functions] are protected.
   ** ^The sqlite3_value object returned by
   ** [sqlite3_column_value()] is unprotected.
   ** Unprotected sqlite3_value objects may only be used with
   ** [sqlite3_result_value()] and [sqlite3_bind_value()].
   ** The [sqlite3_value_blob | sqlite3_value_type()] family of
   ** interfaces require protected sqlite3_value objects.
   */
   typedef struct Mem sqlite3_value;
   
   /*
   ** CAPI3REF: SQL Function Context Object
   **
   ** The context in which an SQL function executes is stored in an
   ** sqlite3_context object.  ^A pointer to an sqlite3_context object
   ** is always first parameter to [application-defined SQL functions].
   ** The application-defined SQL function implementation will pass this
   ** pointer through into calls to [sqlite3_result_int | sqlite3_result()],
   ** [sqlite3_aggregate_context()], [sqlite3_user_data()],
   ** [sqlite3_context_db_handle()], [sqlite3_get_auxdata()],
   ** and/or [sqlite3_set_auxdata()].
   */
   typedef struct sqlite3_context sqlite3_context;
   
   /*
   ** CAPI3REF: Binding Values To Prepared Statements
   ** KEYWORDS: {host parameter} {host parameters} {host parameter name}
   ** KEYWORDS: {SQL parameter} {SQL parameters} {parameter binding}
   **
   ** ^(In the SQL statement text input to [sqlite3_prepare_v2()] and its variants,
   ** literals may be replaced by a [parameter] that matches one of following
   ** templates:
   **
   ** <ul>
   ** <li>  ?
   ** <li>  ?NNN
   ** <li>  :VVV
   ** <li>  @VVV
   ** <li>  $VVV
   ** </ul>
   **
   ** In the templates above, NNN represents an integer literal,
   ** and VVV represents an alphanumeric identifier.)^  ^The values of these
   ** parameters (also called "host parameter names" or "SQL parameters")
   ** can be set using the sqlite3_bind_*() routines defined here.
   **
   ** ^The first argument to the sqlite3_bind_*() routines is always
   ** a pointer to the [sqlite3_stmt] object returned from
   ** [sqlite3_prepare_v2()] or its variants.
   **
   ** ^The second argument is the index of the SQL parameter to be set.
   ** ^The leftmost SQL parameter has an index of 1.  ^When the same named
   ** SQL parameter is used more than once, second and subsequent
   ** occurrences have the same index as the first occurrence.
   ** ^The index for named parameters can be looked up using the
   ** [sqlite3_bind_parameter_index()] API if desired.  ^The index
   ** for "?NNN" parameters is the value of NNN.
   ** ^The NNN value must be between 1 and the [sqlite3_limit()]
   ** parameter [SQLITE_LIMIT_VARIABLE_NUMBER] (default value: 999).
   **
   ** ^The third argument is the value to bind to the parameter.
   **
   ** ^(In those routines that have a fourth argument, its value is the
   ** number of bytes in the parameter.  To be clear: the value is the
   ** number of <u>bytes</u> in the value, not the number of characters.)^
   ** ^If the fourth parameter is negative, the length of the string is
   ** the number of bytes up to the first zero terminator.
   ** If a non-negative fourth parameter is provided to sqlite3_bind_text()
   ** or sqlite3_bind_text16() then that parameter must be the byte offset
   ** where the NUL terminator would occur assuming the string were NUL
   ** terminated.  If any NUL characters occur at byte offsets less than 
   ** the value of the fourth parameter then the resulting string value will
   ** contain embedded NULs.  The result of expressions involving strings
   ** with embedded NULs is undefined.
   **
   ** ^The fifth argument to sqlite3_bind_blob(), sqlite3_bind_text(), and
   ** sqlite3_bind_text16() is a destructor used to dispose of the BLOB or
   ** string after SQLite has finished with it.  ^The destructor is called
   ** to dispose of the BLOB or string even if the call to sqlite3_bind_blob(),
   ** sqlite3_bind_text(), or sqlite3_bind_text16() fails.  
   ** ^If the fifth argument is
   ** the special value [SQLITE_STATIC], then SQLite assumes that the
   ** information is in static, unmanaged space and does not need to be freed.
   ** ^If the fifth argument has the value [SQLITE_TRANSIENT], then
   ** SQLite makes its own private copy of the data immediately, before
   ** the sqlite3_bind_*() routine returns.
   **
   ** ^The sqlite3_bind_zeroblob() routine binds a BLOB of length N that
   ** is filled with zeroes.  ^A zeroblob uses a fixed amount of memory
   ** (just an integer to hold its size) while it is being processed.
   ** Zeroblobs are intended to serve as placeholders for BLOBs whose
   ** content is later written using
   ** [sqlite3_blob_open | incremental BLOB I/O] routines.
   ** ^A negative value for the zeroblob results in a zero-length BLOB.
   **
   ** ^If any of the sqlite3_bind_*() routines are called with a NULL pointer
   ** for the [prepared statement] or with a prepared statement for which
   ** [sqlite3_step()] has been called more recently than [sqlite3_reset()],
   ** then the call will return [SQLITE_MISUSE].  If any sqlite3_bind_()
   ** routine is passed a [prepared statement] that has been finalized, the
   ** result is undefined and probably harmful.
   **
   ** ^Bindings are not cleared by the [sqlite3_reset()] routine.
   ** ^Unbound parameters are interpreted as NULL.
   **
   ** ^The sqlite3_bind_* routines return [SQLITE_OK] on success or an
   ** [error code] if anything goes wrong.
   ** ^[SQLITE_RANGE] is returned if the parameter
   ** index is out of range.  ^[SQLITE_NOMEM] is returned if malloc() fails.
   **
   ** See also: [sqlite3_bind_parameter_count()],
   ** [sqlite3_bind_parameter_name()], and [sqlite3_bind_parameter_index()].
   */
   SQLITE_API int sqlite3_bind_blob(sqlite3_stmt*, int, const void*, int n, void(*)(void*));
   SQLITE_API int sqlite3_bind_double(sqlite3_stmt*, int, double);
   SQLITE_API int sqlite3_bind_int(sqlite3_stmt*, int, int);
   SQLITE_API int sqlite3_bind_int64(sqlite3_stmt*, int, sqlite3_int64);
   SQLITE_API int sqlite3_bind_null(sqlite3_stmt*, int);
   SQLITE_API int sqlite3_bind_text(sqlite3_stmt*, int, const char*, int n, void(*)(void*));
   SQLITE_API int sqlite3_bind_text16(sqlite3_stmt*, int, const void*, int, void(*)(void*));
   SQLITE_API int sqlite3_bind_value(sqlite3_stmt*, int, const sqlite3_value*);
   SQLITE_API int sqlite3_bind_zeroblob(sqlite3_stmt*, int, int n);
   
   /*
   ** CAPI3REF: Number Of SQL Parameters
   **
   ** ^This routine can be used to find the number of [SQL parameters]
   ** in a [prepared statement].  SQL parameters are tokens of the
   ** form "?", "?NNN", ":AAA", "$AAA", or "@AAA" that serve as
   ** placeholders for values that are [sqlite3_bind_blob | bound]
   ** to the parameters at a later time.
   **
   ** ^(This routine actually returns the index of the largest (rightmost)
   ** parameter. For all forms except ?NNN, this will correspond to the
   ** number of unique parameters.  If parameters of the ?NNN form are used,
   ** there may be gaps in the list.)^
   **
   ** See also: [sqlite3_bind_blob|sqlite3_bind()],
   ** [sqlite3_bind_parameter_name()], and
   ** [sqlite3_bind_parameter_index()].
   */
   SQLITE_API int sqlite3_bind_parameter_count(sqlite3_stmt*);
   
   /*
   ** CAPI3REF: Name Of A Host Parameter
   **
   ** ^The sqlite3_bind_parameter_name(P,N) interface returns
   ** the name of the N-th [SQL parameter] in the [prepared statement] P.
   ** ^(SQL parameters of the form "?NNN" or ":AAA" or "@AAA" or "$AAA"
   ** have a name which is the string "?NNN" or ":AAA" or "@AAA" or "$AAA"
   ** respectively.
   ** In other words, the initial ":" or "$" or "@" or "?"
   ** is included as part of the name.)^
   ** ^Parameters of the form "?" without a following integer have no name
   ** and are referred to as "nameless" or "anonymous parameters".
   **
   ** ^The first host parameter has an index of 1, not 0.
   **
   ** ^If the value N is out of range or if the N-th parameter is
   ** nameless, then NULL is returned.  ^The returned string is
   ** always in UTF-8 encoding even if the named parameter was
   ** originally specified as UTF-16 in [sqlite3_prepare16()] or
   ** [sqlite3_prepare16_v2()].
   **
   ** See also: [sqlite3_bind_blob|sqlite3_bind()],
   ** [sqlite3_bind_parameter_count()], and
   ** [sqlite3_bind_parameter_index()].
   */
   SQLITE_API const char *sqlite3_bind_parameter_name(sqlite3_stmt*, int);
   
   /*
   ** CAPI3REF: Index Of A Parameter With A Given Name
   **
   ** ^Return the index of an SQL parameter given its name.  ^The
   ** index value returned is suitable for use as the second
   ** parameter to [sqlite3_bind_blob|sqlite3_bind()].  ^A zero
   ** is returned if no matching parameter is found.  ^The parameter
   ** name must be given in UTF-8 even if the original statement
   ** was prepared from UTF-16 text using [sqlite3_prepare16_v2()].
   **
   ** See also: [sqlite3_bind_blob|sqlite3_bind()],
   ** [sqlite3_bind_parameter_count()], and
   ** [sqlite3_bind_parameter_index()].
   */
   SQLITE_API int sqlite3_bind_parameter_index(sqlite3_stmt*, const char *zName);
   
   /*
   ** CAPI3REF: Reset All Bindings On A Prepared Statement
   **
   ** ^Contrary to the intuition of many, [sqlite3_reset()] does not reset
   ** the [sqlite3_bind_blob | bindings] on a [prepared statement].
   ** ^Use this routine to reset all host parameters to NULL.
   */
   SQLITE_API int sqlite3_clear_bindings(sqlite3_stmt*);
   
   /*
   ** CAPI3REF: Number Of Columns In A Result Set
   **
   ** ^Return the number of columns in the result set returned by the
   ** [prepared statement]. ^This routine returns 0 if pStmt is an SQL
   ** statement that does not return data (for example an [UPDATE]).
   **
   ** See also: [sqlite3_data_count()]
   */
   SQLITE_API int sqlite3_column_count(sqlite3_stmt *pStmt);
   
   /*
   ** CAPI3REF: Column Names In A Result Set
   **
   ** ^These routines return the name assigned to a particular column
   ** in the result set of a [SELECT] statement.  ^The sqlite3_column_name()
   ** interface returns a pointer to a zero-terminated UTF-8 string
   ** and sqlite3_column_name16() returns a pointer to a zero-terminated
   ** UTF-16 string.  ^The first parameter is the [prepared statement]
   ** that implements the [SELECT] statement. ^The second parameter is the
   ** column number.  ^The leftmost column is number 0.
   **
   ** ^The returned string pointer is valid until either the [prepared statement]
   ** is destroyed by [sqlite3_finalize()] or until the statement is automatically
   ** reprepared by the first call to [sqlite3_step()] for a particular run
   ** or until the next call to
   ** sqlite3_column_name() or sqlite3_column_name16() on the same column.
   **
   ** ^If sqlite3_malloc() fails during the processing of either routine
   ** (for example during a conversion from UTF-8 to UTF-16) then a
   ** NULL pointer is returned.
   **
   ** ^The name of a result column is the value of the "AS" clause for
   ** that column, if there is an AS clause.  If there is no AS clause
   ** then the name of the column is unspecified and may change from
   ** one release of SQLite to the next.
   */
   SQLITE_API const char *sqlite3_column_name(sqlite3_stmt*, int N);
   SQLITE_API const void *sqlite3_column_name16(sqlite3_stmt*, int N);
   
   /*
   ** CAPI3REF: Source Of Data In A Query Result
   **
   ** ^These routines provide a means to determine the database, table, and
   ** table column that is the origin of a particular result column in
   ** [SELECT] statement.
   ** ^The name of the database or table or column can be returned as
   ** either a UTF-8 or UTF-16 string.  ^The _database_ routines return
   ** the database name, the _table_ routines return the table name, and
   ** the origin_ routines return the column name.
   ** ^The returned string is valid until the [prepared statement] is destroyed
   ** using [sqlite3_finalize()] or until the statement is automatically
   ** reprepared by the first call to [sqlite3_step()] for a particular run
   ** or until the same information is requested
   ** again in a different encoding.
   **
   ** ^The names returned are the original un-aliased names of the
   ** database, table, and column.
   **
   ** ^The first argument to these interfaces is a [prepared statement].
   ** ^These functions return information about the Nth result column returned by
   ** the statement, where N is the second function argument.
   ** ^The left-most column is column 0 for these routines.
   **
   ** ^If the Nth column returned by the statement is an expression or
   ** subquery and is not a column value, then all of these functions return
   ** NULL.  ^These routine might also return NULL if a memory allocation error
   ** occurs.  ^Otherwise, they return the name of the attached database, table,
   ** or column that query result column was extracted from.
   **
   ** ^As with all other SQLite APIs, those whose names end with "16" return
   ** UTF-16 encoded strings and the other functions return UTF-8.
   **
   ** ^These APIs are only available if the library was compiled with the
   ** [SQLITE_ENABLE_COLUMN_METADATA] C-preprocessor symbol.
   **
   ** If two or more threads call one or more of these routines against the same
   ** prepared statement and column at the same time then the results are
   ** undefined.
   **
   ** If two or more threads call one or more
   ** [sqlite3_column_database_name | column metadata interfaces]
   ** for the same [prepared statement] and result column
   ** at the same time then the results are undefined.
   */
   SQLITE_API const char *sqlite3_column_database_name(sqlite3_stmt*,int);
   SQLITE_API const void *sqlite3_column_database_name16(sqlite3_stmt*,int);
   SQLITE_API const char *sqlite3_column_table_name(sqlite3_stmt*,int);
   SQLITE_API const void *sqlite3_column_table_name16(sqlite3_stmt*,int);
   SQLITE_API const char *sqlite3_column_origin_name(sqlite3_stmt*,int);
   SQLITE_API const void *sqlite3_column_origin_name16(sqlite3_stmt*,int);
   
   /*
   ** CAPI3REF: Declared Datatype Of A Query Result
   **
   ** ^(The first parameter is a [prepared statement].
   ** If this statement is a [SELECT] statement and the Nth column of the
   ** returned result set of that [SELECT] is a table column (not an
   ** expression or subquery) then the declared type of the table
   ** column is returned.)^  ^If the Nth column of the result set is an
   ** expression or subquery, then a NULL pointer is returned.
   ** ^The returned string is always UTF-8 encoded.
   **
   ** ^(For example, given the database schema:
   **
   ** CREATE TABLE t1(c1 VARIANT);
   **
   ** and the following statement to be compiled:
   **
   ** SELECT c1 + 1, c1 FROM t1;
   **
   ** this routine would return the string "VARIANT" for the second result
   ** column (i==1), and a NULL pointer for the first result column (i==0).)^
   **
   ** ^SQLite uses dynamic run-time typing.  ^So just because a column
   ** is declared to contain a particular type does not mean that the
   ** data stored in that column is of the declared type.  SQLite is
   ** strongly typed, but the typing is dynamic not static.  ^Type
   ** is associated with individual values, not with the containers
   ** used to hold those values.
   */
   SQLITE_API const char *sqlite3_column_decltype(sqlite3_stmt*,int);
   SQLITE_API const void *sqlite3_column_decltype16(sqlite3_stmt*,int);
   
   /*
   ** CAPI3REF: Evaluate An SQL Statement
   **
   ** After a [prepared statement] has been prepared using either
   ** [sqlite3_prepare_v2()] or [sqlite3_prepare16_v2()] or one of the legacy
   ** interfaces [sqlite3_prepare()] or [sqlite3_prepare16()], this function
   ** must be called one or more times to evaluate the statement.
   **
   ** The details of the behavior of the sqlite3_step() interface depend
   ** on whether the statement was prepared using the newer "v2" interface
   ** [sqlite3_prepare_v2()] and [sqlite3_prepare16_v2()] or the older legacy
   ** interface [sqlite3_prepare()] and [sqlite3_prepare16()].  The use of the
   ** new "v2" interface is recommended for new applications but the legacy
   ** interface will continue to be supported.
   **
   ** ^In the legacy interface, the return value will be either [SQLITE_BUSY],
   ** [SQLITE_DONE], [SQLITE_ROW], [SQLITE_ERROR], or [SQLITE_MISUSE].
   ** ^With the "v2" interface, any of the other [result codes] or
   ** [extended result codes] might be returned as well.
   **
   ** ^[SQLITE_BUSY] means that the database engine was unable to acquire the
   ** database locks it needs to do its job.  ^If the statement is a [COMMIT]
   ** or occurs outside of an explicit transaction, then you can retry the
   ** statement.  If the statement is not a [COMMIT] and occurs within an
   ** explicit transaction then you should rollback the transaction before
   ** continuing.
   **
   ** ^[SQLITE_DONE] means that the statement has finished executing
   ** successfully.  sqlite3_step() should not be called again on this virtual
   ** machine without first calling [sqlite3_reset()] to reset the virtual
   ** machine back to its initial state.
   **
   ** ^If the SQL statement being executed returns any data, then [SQLITE_ROW]
   ** is returned each time a new row of data is ready for processing by the
   ** caller. The values may be accessed using the [column access functions].
   ** sqlite3_step() is called again to retrieve the next row of data.
   **
   ** ^[SQLITE_ERROR] means that a run-time error (such as a constraint
   ** violation) has occurred.  sqlite3_step() should not be called again on
   ** the VM. More information may be found by calling [sqlite3_errmsg()].
   ** ^With the legacy interface, a more specific error code (for example,
   ** [SQLITE_INTERRUPT], [SQLITE_SCHEMA], [SQLITE_CORRUPT], and so forth)
   ** can be obtained by calling [sqlite3_reset()] on the
   ** [prepared statement].  ^In the "v2" interface,
   ** the more specific error code is returned directly by sqlite3_step().
   **
   ** [SQLITE_MISUSE] means that the this routine was called inappropriately.
   ** Perhaps it was called on a [prepared statement] that has
   ** already been [sqlite3_finalize | finalized] or on one that had
   ** previously returned [SQLITE_ERROR] or [SQLITE_DONE].  Or it could
   ** be the case that the same database connection is being used by two or
   ** more threads at the same moment in time.
   **
   ** For all versions of SQLite up to and including 3.6.23.1, a call to
   ** [sqlite3_reset()] was required after sqlite3_step() returned anything
   ** other than [SQLITE_ROW] before any subsequent invocation of
   ** sqlite3_step().  Failure to reset the prepared statement using 
   ** [sqlite3_reset()] would result in an [SQLITE_MISUSE] return from
   ** sqlite3_step().  But after version 3.6.23.1, sqlite3_step() began
   ** calling [sqlite3_reset()] automatically in this circumstance rather
   ** than returning [SQLITE_MISUSE].  This is not considered a compatibility
   ** break because any application that ever receives an SQLITE_MISUSE error
   ** is broken by definition.  The [SQLITE_OMIT_AUTORESET] compile-time option
   ** can be used to restore the legacy behavior.
   **
   ** <b>Goofy Interface Alert:</b> In the legacy interface, the sqlite3_step()
   ** API always returns a generic error code, [SQLITE_ERROR], following any
   ** error other than [SQLITE_BUSY] and [SQLITE_MISUSE].  You must call
   ** [sqlite3_reset()] or [sqlite3_finalize()] in order to find one of the
   ** specific [error codes] that better describes the error.
   ** We admit that this is a goofy design.  The problem has been fixed
   ** with the "v2" interface.  If you prepare all of your SQL statements
   ** using either [sqlite3_prepare_v2()] or [sqlite3_prepare16_v2()] instead
   ** of the legacy [sqlite3_prepare()] and [sqlite3_prepare16()] interfaces,
   ** then the more specific [error codes] are returned directly
   ** by sqlite3_step().  The use of the "v2" interface is recommended.
   */
   SQLITE_API int sqlite3_step(sqlite3_stmt*);
   
   /*
   ** CAPI3REF: Number of columns in a result set
   **
   ** ^The sqlite3_data_count(P) interface returns the number of columns in the
   ** current row of the result set of [prepared statement] P.
   ** ^If prepared statement P does not have results ready to return
   ** (via calls to the [sqlite3_column_int | sqlite3_column_*()] of
   ** interfaces) then sqlite3_data_count(P) returns 0.
   ** ^The sqlite3_data_count(P) routine also returns 0 if P is a NULL pointer.
   ** ^The sqlite3_data_count(P) routine returns 0 if the previous call to
   ** [sqlite3_step](P) returned [SQLITE_DONE].  ^The sqlite3_data_count(P)
   ** will return non-zero if previous call to [sqlite3_step](P) returned
   ** [SQLITE_ROW], except in the case of the [PRAGMA incremental_vacuum]
   ** where it always returns zero since each step of that multi-step
   ** pragma returns 0 columns of data.
   **
   ** See also: [sqlite3_column_count()]
   */
   SQLITE_API int sqlite3_data_count(sqlite3_stmt *pStmt);
   
   /*
   ** CAPI3REF: Fundamental Datatypes
   ** KEYWORDS: SQLITE_TEXT
   **
   ** ^(Every value in SQLite has one of five fundamental datatypes:
   **
   ** <ul>
   ** <li> 64-bit signed integer
   ** <li> 64-bit IEEE floating point number
   ** <li> string
   ** <li> BLOB
   ** <li> NULL
   ** </ul>)^
   **
   ** These constants are codes for each of those types.
   **
   ** Note that the SQLITE_TEXT constant was also used in SQLite version 2
   ** for a completely different meaning.  Software that links against both
   ** SQLite version 2 and SQLite version 3 should use SQLITE3_TEXT, not
   ** SQLITE_TEXT.
   */
   #define SQLITE_INTEGER  1
   #define SQLITE_FLOAT    2
   #define SQLITE_BLOB     4
   #define SQLITE_NULL     5
   #ifdef SQLITE_TEXT
   # undef SQLITE_TEXT
   #else
   # define SQLITE_TEXT     3
   #endif
   #define SQLITE3_TEXT     3
   
   /*
   ** CAPI3REF: Result Values From A Query
   ** KEYWORDS: {column access functions}
   **
   ** These routines form the "result set" interface.
   **
   ** ^These routines return information about a single column of the current
   ** result row of a query.  ^In every case the first argument is a pointer
   ** to the [prepared statement] that is being evaluated (the [sqlite3_stmt*]
   ** that was returned from [sqlite3_prepare_v2()] or one of its variants)
   ** and the second argument is the index of the column for which information
   ** should be returned. ^The leftmost column of the result set has the index 0.
   ** ^The number of columns in the result can be determined using
   ** [sqlite3_column_count()].
   **
   ** If the SQL statement does not currently point to a valid row, or if the
   ** column index is out of range, the result is undefined.
   ** These routines may only be called when the most recent call to
   ** [sqlite3_step()] has returned [SQLITE_ROW] and neither
   ** [sqlite3_reset()] nor [sqlite3_finalize()] have been called subsequently.
   ** If any of these routines are called after [sqlite3_reset()] or
   ** [sqlite3_finalize()] or after [sqlite3_step()] has returned
   ** something other than [SQLITE_ROW], the results are undefined.
   ** If [sqlite3_step()] or [sqlite3_reset()] or [sqlite3_finalize()]
   ** are called from a different thread while any of these routines
   ** are pending, then the results are undefined.
   **
   ** ^The sqlite3_column_type() routine returns the
   ** [SQLITE_INTEGER | datatype code] for the initial data type
   ** of the result column.  ^The returned value is one of [SQLITE_INTEGER],
   ** [SQLITE_FLOAT], [SQLITE_TEXT], [SQLITE_BLOB], or [SQLITE_NULL].  The value
   ** returned by sqlite3_column_type() is only meaningful if no type
   ** conversions have occurred as described below.  After a type conversion,
   ** the value returned by sqlite3_column_type() is undefined.  Future
   ** versions of SQLite may change the behavior of sqlite3_column_type()
   ** following a type conversion.
   **
   ** ^If the result is a BLOB or UTF-8 string then the sqlite3_column_bytes()
   ** routine returns the number of bytes in that BLOB or string.
   ** ^If the result is a UTF-16 string, then sqlite3_column_bytes() converts
   ** the string to UTF-8 and then returns the number of bytes.
   ** ^If the result is a numeric value then sqlite3_column_bytes() uses
   ** [sqlite3_snprintf()] to convert that value to a UTF-8 string and returns
   ** the number of bytes in that string.
   ** ^If the result is NULL, then sqlite3_column_bytes() returns zero.
   **
   ** ^If the result is a BLOB or UTF-16 string then the sqlite3_column_bytes16()
   ** routine returns the number of bytes in that BLOB or string.
   ** ^If the result is a UTF-8 string, then sqlite3_column_bytes16() converts
   ** the string to UTF-16 and then returns the number of bytes.
   ** ^If the result is a numeric value then sqlite3_column_bytes16() uses
   ** [sqlite3_snprintf()] to convert that value to a UTF-16 string and returns
   ** the number of bytes in that string.
   ** ^If the result is NULL, then sqlite3_column_bytes16() returns zero.
   **
   ** ^The values returned by [sqlite3_column_bytes()] and 
   ** [sqlite3_column_bytes16()] do not include the zero terminators at the end
   ** of the string.  ^For clarity: the values returned by
   ** [sqlite3_column_bytes()] and [sqlite3_column_bytes16()] are the number of
   ** bytes in the string, not the number of characters.
   **
   ** ^Strings returned by sqlite3_column_text() and sqlite3_column_text16(),
   ** even empty strings, are always zero-terminated.  ^The return
   ** value from sqlite3_column_blob() for a zero-length BLOB is a NULL pointer.
   **
   ** ^The object returned by [sqlite3_column_value()] is an
   ** [unprotected sqlite3_value] object.  An unprotected sqlite3_value object
   ** may only be used with [sqlite3_bind_value()] and [sqlite3_result_value()].
   ** If the [unprotected sqlite3_value] object returned by
   ** [sqlite3_column_value()] is used in any other way, including calls
   ** to routines like [sqlite3_value_int()], [sqlite3_value_text()],
   ** or [sqlite3_value_bytes()], then the behavior is undefined.
   **
   ** These routines attempt to convert the value where appropriate.  ^For
   ** example, if the internal representation is FLOAT and a text result
   ** is requested, [sqlite3_snprintf()] is used internally to perform the
   ** conversion automatically.  ^(The following table details the conversions
   ** that are applied:
   **
   ** <blockquote>
   ** <table border="1">
   ** <tr><th> Internal<br>Type <th> Requested<br>Type <th>  Conversion
   **
   ** <tr><td>  NULL    <td> INTEGER   <td> Result is 0
   ** <tr><td>  NULL    <td>  FLOAT    <td> Result is 0.0
   ** <tr><td>  NULL    <td>   TEXT    <td> Result is NULL pointer
   ** <tr><td>  NULL    <td>   BLOB    <td> Result is NULL pointer
   ** <tr><td> INTEGER  <td>  FLOAT    <td> Convert from integer to float
   ** <tr><td> INTEGER  <td>   TEXT    <td> ASCII rendering of the integer
   ** <tr><td> INTEGER  <td>   BLOB    <td> Same as INTEGER->TEXT
   ** <tr><td>  FLOAT   <td> INTEGER   <td> Convert from float to integer
   ** <tr><td>  FLOAT   <td>   TEXT    <td> ASCII rendering of the float
   ** <tr><td>  FLOAT   <td>   BLOB    <td> Same as FLOAT->TEXT
   ** <tr><td>  TEXT    <td> INTEGER   <td> Use atoi()
   ** <tr><td>  TEXT    <td>  FLOAT    <td> Use atof()
   ** <tr><td>  TEXT    <td>   BLOB    <td> No change
   ** <tr><td>  BLOB    <td> INTEGER   <td> Convert to TEXT then use atoi()
   ** <tr><td>  BLOB    <td>  FLOAT    <td> Convert to TEXT then use atof()
   ** <tr><td>  BLOB    <td>   TEXT    <td> Add a zero terminator if needed
   ** </table>
   ** </blockquote>)^
   **
   ** The table above makes reference to standard C library functions atoi()
   ** and atof().  SQLite does not really use these functions.  It has its
   ** own equivalent internal routines.  The atoi() and atof() names are
   ** used in the table for brevity and because they are familiar to most
   ** C programmers.
   **
   ** Note that when type conversions occur, pointers returned by prior
   ** calls to sqlite3_column_blob(), sqlite3_column_text(), and/or
   ** sqlite3_column_text16() may be invalidated.
   ** Type conversions and pointer invalidations might occur
   ** in the following cases:
   **
   ** <ul>
   ** <li> The initial content is a BLOB and sqlite3_column_text() or
   **      sqlite3_column_text16() is called.  A zero-terminator might
   **      need to be added to the string.</li>
   ** <li> The initial content is UTF-8 text and sqlite3_column_bytes16() or
   **      sqlite3_column_text16() is called.  The content must be converted
   **      to UTF-16.</li>
   ** <li> The initial content is UTF-16 text and sqlite3_column_bytes() or
   **      sqlite3_column_text() is called.  The content must be converted
   **      to UTF-8.</li>
   ** </ul>
   **
   ** ^Conversions between UTF-16be and UTF-16le are always done in place and do
   ** not invalidate a prior pointer, though of course the content of the buffer
   ** that the prior pointer references will have been modified.  Other kinds
   ** of conversion are done in place when it is possible, but sometimes they
   ** are not possible and in those cases prior pointers are invalidated.
   **
   ** The safest and easiest to remember policy is to invoke these routines
   ** in one of the following ways:
   **
   ** <ul>
   **  <li>sqlite3_column_text() followed by sqlite3_column_bytes()</li>
   **  <li>sqlite3_column_blob() followed by sqlite3_column_bytes()</li>
   **  <li>sqlite3_column_text16() followed by sqlite3_column_bytes16()</li>
   ** </ul>
   **
   ** In other words, you should call sqlite3_column_text(),
   ** sqlite3_column_blob(), or sqlite3_column_text16() first to force the result
   ** into the desired format, then invoke sqlite3_column_bytes() or
   ** sqlite3_column_bytes16() to find the size of the result.  Do not mix calls
   ** to sqlite3_column_text() or sqlite3_column_blob() with calls to
   ** sqlite3_column_bytes16(), and do not mix calls to sqlite3_column_text16()
   ** with calls to sqlite3_column_bytes().
   **
   ** ^The pointers returned are valid until a type conversion occurs as
   ** described above, or until [sqlite3_step()] or [sqlite3_reset()] or
   ** [sqlite3_finalize()] is called.  ^The memory space used to hold strings
   ** and BLOBs is freed automatically.  Do <b>not</b> pass the pointers returned
   ** [sqlite3_column_blob()], [sqlite3_column_text()], etc. into
   ** [sqlite3_free()].
   **
   ** ^(If a memory allocation error occurs during the evaluation of any
   ** of these routines, a default value is returned.  The default value
   ** is either the integer 0, the floating point number 0.0, or a NULL
   ** pointer.  Subsequent calls to [sqlite3_errcode()] will return
   ** [SQLITE_NOMEM].)^
   */
   SQLITE_API const void *sqlite3_column_blob(sqlite3_stmt*, int iCol);
   SQLITE_API int sqlite3_column_bytes(sqlite3_stmt*, int iCol);
   SQLITE_API int sqlite3_column_bytes16(sqlite3_stmt*, int iCol);
   SQLITE_API double sqlite3_column_double(sqlite3_stmt*, int iCol);
   SQLITE_API int sqlite3_column_int(sqlite3_stmt*, int iCol);
   SQLITE_API sqlite3_int64 sqlite3_column_int64(sqlite3_stmt*, int iCol);
   SQLITE_API const unsigned char *sqlite3_column_text(sqlite3_stmt*, int iCol);
   SQLITE_API const void *sqlite3_column_text16(sqlite3_stmt*, int iCol);
   SQLITE_API int sqlite3_column_type(sqlite3_stmt*, int iCol);
   SQLITE_API sqlite3_value *sqlite3_column_value(sqlite3_stmt*, int iCol);
   
   /*
   ** CAPI3REF: Destroy A Prepared Statement Object
   **
   ** ^The sqlite3_finalize() function is called to delete a [prepared statement].
   ** ^If the most recent evaluation of the statement encountered no errors
   ** or if the statement is never been evaluated, then sqlite3_finalize() returns
   ** SQLITE_OK.  ^If the most recent evaluation of statement S failed, then
   ** sqlite3_finalize(S) returns the appropriate [error code] or
   ** [extended error code].
   **
   ** ^The sqlite3_finalize(S) routine can be called at any point during
   ** the life cycle of [prepared statement] S:
   ** before statement S is ever evaluated, after
   ** one or more calls to [sqlite3_reset()], or after any call
   ** to [sqlite3_step()] regardless of whether or not the statement has
   ** completed execution.
   **
   ** ^Invoking sqlite3_finalize() on a NULL pointer is a harmless no-op.
   **
   ** The application must finalize every [prepared statement] in order to avoid
   ** resource leaks.  It is a grievous error for the application to try to use
   ** a prepared statement after it has been finalized.  Any use of a prepared
   ** statement after it has been finalized can result in undefined and
   ** undesirable behavior such as segfaults and heap corruption.
   */
   SQLITE_API int sqlite3_finalize(sqlite3_stmt *pStmt);
   
   /*
   ** CAPI3REF: Reset A Prepared Statement Object
   **
   ** The sqlite3_reset() function is called to reset a [prepared statement]
   ** object back to its initial state, ready to be re-executed.
   ** ^Any SQL statement variables that had values bound to them using
   ** the [sqlite3_bind_blob | sqlite3_bind_*() API] retain their values.
   ** Use [sqlite3_clear_bindings()] to reset the bindings.
   **
   ** ^The [sqlite3_reset(S)] interface resets the [prepared statement] S
   ** back to the beginning of its program.
   **
   ** ^If the most recent call to [sqlite3_step(S)] for the
   ** [prepared statement] S returned [SQLITE_ROW] or [SQLITE_DONE],
   ** or if [sqlite3_step(S)] has never before been called on S,
   ** then [sqlite3_reset(S)] returns [SQLITE_OK].
   **
   ** ^If the most recent call to [sqlite3_step(S)] for the
   ** [prepared statement] S indicated an error, then
   ** [sqlite3_reset(S)] returns an appropriate [error code].
   **
   ** ^The [sqlite3_reset(S)] interface does not change the values
   ** of any [sqlite3_bind_blob|bindings] on the [prepared statement] S.
   */
   SQLITE_API int sqlite3_reset(sqlite3_stmt *pStmt);
   
   /*
   ** CAPI3REF: Create Or Redefine SQL Functions
   ** KEYWORDS: {function creation routines}
   ** KEYWORDS: {application-defined SQL function}
   ** KEYWORDS: {application-defined SQL functions}
   **
   ** ^These functions (collectively known as "function creation routines")
   ** are used to add SQL functions or aggregates or to redefine the behavior
   ** of existing SQL functions or aggregates.  The only differences between
   ** these routines are the text encoding expected for
   ** the second parameter (the name of the function being created)
   ** and the presence or absence of a destructor callback for
   ** the application data pointer.
   **
   ** ^The first parameter is the [database connection] to which the SQL
   ** function is to be added.  ^If an application uses more than one database
   ** connection then application-defined SQL functions must be added
   ** to each database connection separately.
   **
   ** ^The second parameter is the name of the SQL function to be created or
   ** redefined.  ^The length of the name is limited to 255 bytes in a UTF-8
   ** representation, exclusive of the zero-terminator.  ^Note that the name
   ** length limit is in UTF-8 bytes, not characters nor UTF-16 bytes.  
   ** ^Any attempt to create a function with a longer name
   ** will result in [SQLITE_MISUSE] being returned.
   **
   ** ^The third parameter (nArg)
   ** is the number of arguments that the SQL function or
   ** aggregate takes. ^If this parameter is -1, then the SQL function or
   ** aggregate may take any number of arguments between 0 and the limit
   ** set by [sqlite3_limit]([SQLITE_LIMIT_FUNCTION_ARG]).  If the third
   ** parameter is less than -1 or greater than 127 then the behavior is
   ** undefined.
   **
   ** ^The fourth parameter, eTextRep, specifies what
   ** [SQLITE_UTF8 | text encoding] this SQL function prefers for
   ** its parameters.  Every SQL function implementation must be able to work
   ** with UTF-8, UTF-16le, or UTF-16be.  But some implementations may be
   ** more efficient with one encoding than another.  ^An application may
   ** invoke sqlite3_create_function() or sqlite3_create_function16() multiple
   ** times with the same function but with different values of eTextRep.
   ** ^When multiple implementations of the same function are available, SQLite
   ** will pick the one that involves the least amount of data conversion.
   ** If there is only a single implementation which does not care what text
   ** encoding is used, then the fourth argument should be [SQLITE_ANY].
   **
   ** ^(The fifth parameter is an arbitrary pointer.  The implementation of the
   ** function can gain access to this pointer using [sqlite3_user_data()].)^
   **
   ** ^The sixth, seventh and eighth parameters, xFunc, xStep and xFinal, are
   ** pointers to C-language functions that implement the SQL function or
   ** aggregate. ^A scalar SQL function requires an implementation of the xFunc
   ** callback only; NULL pointers must be passed as the xStep and xFinal
   ** parameters. ^An aggregate SQL function requires an implementation of xStep
   ** and xFinal and NULL pointer must be passed for xFunc. ^To delete an existing
   ** SQL function or aggregate, pass NULL pointers for all three function
   ** callbacks.
   **
   ** ^(If the ninth parameter to sqlite3_create_function_v2() is not NULL,
   ** then it is destructor for the application data pointer. 
   ** The destructor is invoked when the function is deleted, either by being
   ** overloaded or when the database connection closes.)^
   ** ^The destructor is also invoked if the call to
   ** sqlite3_create_function_v2() fails.
   ** ^When the destructor callback of the tenth parameter is invoked, it
   ** is passed a single argument which is a copy of the application data 
   ** pointer which was the fifth parameter to sqlite3_create_function_v2().
   **
   ** ^It is permitted to register multiple implementations of the same
   ** functions with the same name but with either differing numbers of
   ** arguments or differing preferred text encodings.  ^SQLite will use
   ** the implementation that most closely matches the way in which the
   ** SQL function is used.  ^A function implementation with a non-negative
   ** nArg parameter is a better match than a function implementation with
   ** a negative nArg.  ^A function where the preferred text encoding
   ** matches the database encoding is a better
   ** match than a function where the encoding is different.  
   ** ^A function where the encoding difference is between UTF16le and UTF16be
   ** is a closer match than a function where the encoding difference is
   ** between UTF8 and UTF16.
   **
   ** ^Built-in functions may be overloaded by new application-defined functions.
   **
   ** ^An application-defined function is permitted to call other
   ** SQLite interfaces.  However, such calls must not
   ** close the database connection nor finalize or reset the prepared
   ** statement in which the function is running.
   */
   SQLITE_API int sqlite3_create_function(
     sqlite3 *db,
     const char *zFunctionName,
     int nArg,
     int eTextRep,
     void *pApp,
     void (*xFunc)(sqlite3_context*,int,sqlite3_value**),
     void (*xStep)(sqlite3_context*,int,sqlite3_value**),
     void (*xFinal)(sqlite3_context*)
   );
   SQLITE_API int sqlite3_create_function16(
     sqlite3 *db,
     const void *zFunctionName,
     int nArg,
     int eTextRep,
     void *pApp,
     void (*xFunc)(sqlite3_context*,int,sqlite3_value**),
     void (*xStep)(sqlite3_context*,int,sqlite3_value**),
     void (*xFinal)(sqlite3_context*)
   );
   SQLITE_API int sqlite3_create_function_v2(
     sqlite3 *db,
     const char *zFunctionName,
     int nArg,
     int eTextRep,
     void *pApp,
     void (*xFunc)(sqlite3_context*,int,sqlite3_value**),
     void (*xStep)(sqlite3_context*,int,sqlite3_value**),
     void (*xFinal)(sqlite3_context*),
     void(*xDestroy)(void*)
   );
   
   /*
   ** CAPI3REF: Text Encodings
   **
   ** These constant define integer codes that represent the various
   ** text encodings supported by SQLite.
   */
   #define SQLITE_UTF8           1
   #define SQLITE_UTF16LE        2
   #define SQLITE_UTF16BE        3
   #define SQLITE_UTF16          4    /* Use native byte order */
   #define SQLITE_ANY            5    /* sqlite3_create_function only */
   #define SQLITE_UTF16_ALIGNED  8    /* sqlite3_create_collation only */
   
   /*
   ** CAPI3REF: Deprecated Functions
   ** DEPRECATED
   **
   ** These functions are [deprecated].  In order to maintain
   ** backwards compatibility with older code, these functions continue 
   ** to be supported.  However, new applications should avoid
   ** the use of these functions.  To help encourage people to avoid
   ** using these functions, we are not going to tell you what they do.
   */
   #ifndef SQLITE_OMIT_DEPRECATED
   SQLITE_API SQLITE_DEPRECATED int sqlite3_aggregate_count(sqlite3_context*);
   SQLITE_API SQLITE_DEPRECATED int sqlite3_expired(sqlite3_stmt*);
   SQLITE_API SQLITE_DEPRECATED int sqlite3_transfer_bindings(sqlite3_stmt*, sqlite3_stmt*);
   SQLITE_API SQLITE_DEPRECATED int sqlite3_global_recover(void);
   SQLITE_API SQLITE_DEPRECATED void sqlite3_thread_cleanup(void);
   SQLITE_API SQLITE_DEPRECATED int sqlite3_memory_alarm(void(*)(void*,sqlite3_int64,int),void*,sqlite3_int64);
   #endif
   
   /*
   ** CAPI3REF: Obtaining SQL Function Parameter Values
   **
   ** The C-language implementation of SQL functions and aggregates uses
   ** this set of interface routines to access the parameter values on
   ** the function or aggregate.
   **
   ** The xFunc (for scalar functions) or xStep (for aggregates) parameters
   ** to [sqlite3_create_function()] and [sqlite3_create_function16()]
   ** define callbacks that implement the SQL functions and aggregates.
   ** The 3rd parameter to these callbacks is an array of pointers to
   ** [protected sqlite3_value] objects.  There is one [sqlite3_value] object for
   ** each parameter to the SQL function.  These routines are used to
   ** extract values from the [sqlite3_value] objects.
   **
   ** These routines work only with [protected sqlite3_value] objects.
   ** Any attempt to use these routines on an [unprotected sqlite3_value]
   ** object results in undefined behavior.
   **
   ** ^These routines work just like the corresponding [column access functions]
   ** except that  these routines take a single [protected sqlite3_value] object
   ** pointer instead of a [sqlite3_stmt*] pointer and an integer column number.
   **
   ** ^The sqlite3_value_text16() interface extracts a UTF-16 string
   ** in the native byte-order of the host machine.  ^The
   ** sqlite3_value_text16be() and sqlite3_value_text16le() interfaces
   ** extract UTF-16 strings as big-endian and little-endian respectively.
   **
   ** ^(The sqlite3_value_numeric_type() interface attempts to apply
   ** numeric affinity to the value.  This means that an attempt is
   ** made to convert the value to an integer or floating point.  If
   ** such a conversion is possible without loss of information (in other
   ** words, if the value is a string that looks like a number)
   ** then the conversion is performed.  Otherwise no conversion occurs.
   ** The [SQLITE_INTEGER | datatype] after conversion is returned.)^
   **
   ** Please pay particular attention to the fact that the pointer returned
   ** from [sqlite3_value_blob()], [sqlite3_value_text()], or
   ** [sqlite3_value_text16()] can be invalidated by a subsequent call to
   ** [sqlite3_value_bytes()], [sqlite3_value_bytes16()], [sqlite3_value_text()],
   ** or [sqlite3_value_text16()].
   **
   ** These routines must be called from the same thread as
   ** the SQL function that supplied the [sqlite3_value*] parameters.
   */
   SQLITE_API const void *sqlite3_value_blob(sqlite3_value*);
   SQLITE_API int sqlite3_value_bytes(sqlite3_value*);
   SQLITE_API int sqlite3_value_bytes16(sqlite3_value*);
   SQLITE_API double sqlite3_value_double(sqlite3_value*);
   SQLITE_API int sqlite3_value_int(sqlite3_value*);
   SQLITE_API sqlite3_int64 sqlite3_value_int64(sqlite3_value*);
   SQLITE_API const unsigned char *sqlite3_value_text(sqlite3_value*);
   SQLITE_API const void *sqlite3_value_text16(sqlite3_value*);
   SQLITE_API const void *sqlite3_value_text16le(sqlite3_value*);
   SQLITE_API const void *sqlite3_value_text16be(sqlite3_value*);
   SQLITE_API int sqlite3_value_type(sqlite3_value*);
   SQLITE_API int sqlite3_value_numeric_type(sqlite3_value*);
   
   /*
   ** CAPI3REF: Obtain Aggregate Function Context
   **
   ** Implementations of aggregate SQL functions use this
   ** routine to allocate memory for storing their state.
   **
   ** ^The first time the sqlite3_aggregate_context(C,N) routine is called 
   ** for a particular aggregate function, SQLite
   ** allocates N of memory, zeroes out that memory, and returns a pointer
   ** to the new memory. ^On second and subsequent calls to
   ** sqlite3_aggregate_context() for the same aggregate function instance,
   ** the same buffer is returned.  Sqlite3_aggregate_context() is normally
   ** called once for each invocation of the xStep callback and then one
   ** last time when the xFinal callback is invoked.  ^(When no rows match
   ** an aggregate query, the xStep() callback of the aggregate function
   ** implementation is never called and xFinal() is called exactly once.
   ** In those cases, sqlite3_aggregate_context() might be called for the
   ** first time from within xFinal().)^
   **
   ** ^The sqlite3_aggregate_context(C,N) routine returns a NULL pointer if N is
   ** less than or equal to zero or if a memory allocate error occurs.
   **
   ** ^(The amount of space allocated by sqlite3_aggregate_context(C,N) is
   ** determined by the N parameter on first successful call.  Changing the
   ** value of N in subsequent call to sqlite3_aggregate_context() within
   ** the same aggregate function instance will not resize the memory
   ** allocation.)^
   **
   ** ^SQLite automatically frees the memory allocated by 
   ** sqlite3_aggregate_context() when the aggregate query concludes.
   **
   ** The first parameter must be a copy of the
   ** [sqlite3_context | SQL function context] that is the first parameter
   ** to the xStep or xFinal callback routine that implements the aggregate
   ** function.
   **
   ** This routine must be called from the same thread in which
   ** the aggregate SQL function is running.
   */
   SQLITE_API void *sqlite3_aggregate_context(sqlite3_context*, int nBytes);
   
   /*
   ** CAPI3REF: User Data For Functions
   **
   ** ^The sqlite3_user_data() interface returns a copy of
   ** the pointer that was the pUserData parameter (the 5th parameter)
   ** of the [sqlite3_create_function()]
   ** and [sqlite3_create_function16()] routines that originally
   ** registered the application defined function.
   **
   ** This routine must be called from the same thread in which
   ** the application-defined function is running.
   */
   SQLITE_API void *sqlite3_user_data(sqlite3_context*);
   
   /*
   ** CAPI3REF: Database Connection For Functions
   **
   ** ^The sqlite3_context_db_handle() interface returns a copy of
   ** the pointer to the [database connection] (the 1st parameter)
   ** of the [sqlite3_create_function()]
   ** and [sqlite3_create_function16()] routines that originally
   ** registered the application defined function.
   */
   SQLITE_API sqlite3 *sqlite3_context_db_handle(sqlite3_context*);
   
   /*
   ** CAPI3REF: Function Auxiliary Data
   **
   ** The following two functions may be used by scalar SQL functions to
   ** associate metadata with argument values. If the same value is passed to
   ** multiple invocations of the same SQL function during query execution, under
   ** some circumstances the associated metadata may be preserved. This may
   ** be used, for example, to add a regular-expression matching scalar
   ** function. The compiled version of the regular expression is stored as
   ** metadata associated with the SQL value passed as the regular expression
   ** pattern.  The compiled regular expression can be reused on multiple
   ** invocations of the same function so that the original pattern string
   ** does not need to be recompiled on each invocation.
   **
   ** ^The sqlite3_get_auxdata() interface returns a pointer to the metadata
   ** associated by the sqlite3_set_auxdata() function with the Nth argument
   ** value to the application-defined function. ^If no metadata has been ever
   ** been set for the Nth argument of the function, or if the corresponding
   ** function parameter has changed since the meta-data was set,
   ** then sqlite3_get_auxdata() returns a NULL pointer.
   **
   ** ^The sqlite3_set_auxdata() interface saves the metadata
   ** pointed to by its 3rd parameter as the metadata for the N-th
   ** argument of the application-defined function.  Subsequent
   ** calls to sqlite3_get_auxdata() might return this data, if it has
   ** not been destroyed.
   ** ^If it is not NULL, SQLite will invoke the destructor
   ** function given by the 4th parameter to sqlite3_set_auxdata() on
   ** the metadata when the corresponding function parameter changes
   ** or when the SQL statement completes, whichever comes first.
   **
   ** SQLite is free to call the destructor and drop metadata on any
   ** parameter of any function at any time.  ^The only guarantee is that
   ** the destructor will be called before the metadata is dropped.
   **
   ** ^(In practice, metadata is preserved between function calls for
   ** expressions that are constant at compile time. This includes literal
   ** values and [parameters].)^
   **
   ** These routines must be called from the same thread in which
   ** the SQL function is running.
   */
   SQLITE_API void *sqlite3_get_auxdata(sqlite3_context*, int N);
   SQLITE_API void sqlite3_set_auxdata(sqlite3_context*, int N, void*, void (*)(void*));
   
   
   /*
   ** CAPI3REF: Constants Defining Special Destructor Behavior
   **
   ** These are special values for the destructor that is passed in as the
   ** final argument to routines like [sqlite3_result_blob()].  ^If the destructor
   ** argument is SQLITE_STATIC, it means that the content pointer is constant
   ** and will never change.  It does not need to be destroyed.  ^The
   ** SQLITE_TRANSIENT value means that the content will likely change in
   ** the near future and that SQLite should make its own private copy of
   ** the content before returning.
   **
   ** The typedef is necessary to work around problems in certain
   ** C++ compilers.  See ticket #2191.
   */
   typedef void (*sqlite3_destructor_type)(void*);
   #define SQLITE_STATIC      ((sqlite3_destructor_type)0)
   #define SQLITE_TRANSIENT   ((sqlite3_destructor_type)-1)
   
   /*
   ** CAPI3REF: Setting The Result Of An SQL Function
   **
   ** These routines are used by the xFunc or xFinal callbacks that
   ** implement SQL functions and aggregates.  See
   ** [sqlite3_create_function()] and [sqlite3_create_function16()]
   ** for additional information.
   **
   ** These functions work very much like the [parameter binding] family of
   ** functions used to bind values to host parameters in prepared statements.
   ** Refer to the [SQL parameter] documentation for additional information.
   **
   ** ^The sqlite3_result_blob() interface sets the result from
   ** an application-defined function to be the BLOB whose content is pointed
   ** to by the second parameter and which is N bytes long where N is the
   ** third parameter.
   **
   ** ^The sqlite3_result_zeroblob() interfaces set the result of
   ** the application-defined function to be a BLOB containing all zero
   ** bytes and N bytes in size, where N is the value of the 2nd parameter.
   **
   ** ^The sqlite3_result_double() interface sets the result from
   ** an application-defined function to be a floating point value specified
   ** by its 2nd argument.
   **
   ** ^The sqlite3_result_error() and sqlite3_result_error16() functions
   ** cause the implemented SQL function to throw an exception.
   ** ^SQLite uses the string pointed to by the
   ** 2nd parameter of sqlite3_result_error() or sqlite3_result_error16()
   ** as the text of an error message.  ^SQLite interprets the error
   ** message string from sqlite3_result_error() as UTF-8. ^SQLite
   ** interprets the string from sqlite3_result_error16() as UTF-16 in native
   ** byte order.  ^If the third parameter to sqlite3_result_error()
   ** or sqlite3_result_error16() is negative then SQLite takes as the error
   ** message all text up through the first zero character.
   ** ^If the third parameter to sqlite3_result_error() or
   ** sqlite3_result_error16() is non-negative then SQLite takes that many
   ** bytes (not characters) from the 2nd parameter as the error message.
   ** ^The sqlite3_result_error() and sqlite3_result_error16()
   ** routines make a private copy of the error message text before
   ** they return.  Hence, the calling function can deallocate or
   ** modify the text after they return without harm.
   ** ^The sqlite3_result_error_code() function changes the error code
   ** returned by SQLite as a result of an error in a function.  ^By default,
   ** the error code is SQLITE_ERROR.  ^A subsequent call to sqlite3_result_error()
   ** or sqlite3_result_error16() resets the error code to SQLITE_ERROR.
   **
   ** ^The sqlite3_result_toobig() interface causes SQLite to throw an error
   ** indicating that a string or BLOB is too long to represent.
   **
   ** ^The sqlite3_result_nomem() interface causes SQLite to throw an error
   ** indicating that a memory allocation failed.
   **
   ** ^The sqlite3_result_int() interface sets the return value
   ** of the application-defined function to be the 32-bit signed integer
   ** value given in the 2nd argument.
   ** ^The sqlite3_result_int64() interface sets the return value
   ** of the application-defined function to be the 64-bit signed integer
   ** value given in the 2nd argument.
   **
   ** ^The sqlite3_result_null() interface sets the return value
   ** of the application-defined function to be NULL.
   **
   ** ^The sqlite3_result_text(), sqlite3_result_text16(),
   ** sqlite3_result_text16le(), and sqlite3_result_text16be() interfaces
   ** set the return value of the application-defined function to be
   ** a text string which is represented as UTF-8, UTF-16 native byte order,
   ** UTF-16 little endian, or UTF-16 big endian, respectively.
   ** ^SQLite takes the text result from the application from
   ** the 2nd parameter of the sqlite3_result_text* interfaces.
   ** ^If the 3rd parameter to the sqlite3_result_text* interfaces
   ** is negative, then SQLite takes result text from the 2nd parameter
   ** through the first zero character.
   ** ^If the 3rd parameter to the sqlite3_result_text* interfaces
   ** is non-negative, then as many bytes (not characters) of the text
   ** pointed to by the 2nd parameter are taken as the application-defined
   ** function result.  If the 3rd parameter is non-negative, then it
   ** must be the byte offset into the string where the NUL terminator would
   ** appear if the string where NUL terminated.  If any NUL characters occur
   ** in the string at a byte offset that is less than the value of the 3rd
   ** parameter, then the resulting string will contain embedded NULs and the
   ** result of expressions operating on strings with embedded NULs is undefined.
   ** ^If the 4th parameter to the sqlite3_result_text* interfaces
   ** or sqlite3_result_blob is a non-NULL pointer, then SQLite calls that
   ** function as the destructor on the text or BLOB result when it has
   ** finished using that result.
   ** ^If the 4th parameter to the sqlite3_result_text* interfaces or to
   ** sqlite3_result_blob is the special constant SQLITE_STATIC, then SQLite
   ** assumes that the text or BLOB result is in constant space and does not
   ** copy the content of the parameter nor call a destructor on the content
   ** when it has finished using that result.
   ** ^If the 4th parameter to the sqlite3_result_text* interfaces
   ** or sqlite3_result_blob is the special constant SQLITE_TRANSIENT
   ** then SQLite makes a copy of the result into space obtained from
   ** from [sqlite3_malloc()] before it returns.
   **
   ** ^The sqlite3_result_value() interface sets the result of
   ** the application-defined function to be a copy the
   ** [unprotected sqlite3_value] object specified by the 2nd parameter.  ^The
   ** sqlite3_result_value() interface makes a copy of the [sqlite3_value]
   ** so that the [sqlite3_value] specified in the parameter may change or
   ** be deallocated after sqlite3_result_value() returns without harm.
   ** ^A [protected sqlite3_value] object may always be used where an
   ** [unprotected sqlite3_value] object is required, so either
   ** kind of [sqlite3_value] object can be used with this interface.
   **
   ** If these routines are called from within the different thread
   ** than the one containing the application-defined function that received
   ** the [sqlite3_context] pointer, the results are undefined.
   */
   SQLITE_API void sqlite3_result_blob(sqlite3_context*, const void*, int, void(*)(void*));
   SQLITE_API void sqlite3_result_double(sqlite3_context*, double);
   SQLITE_API void sqlite3_result_error(sqlite3_context*, const char*, int);
   SQLITE_API void sqlite3_result_error16(sqlite3_context*, const void*, int);
   SQLITE_API void sqlite3_result_error_toobig(sqlite3_context*);
   SQLITE_API void sqlite3_result_error_nomem(sqlite3_context*);
   SQLITE_API void sqlite3_result_error_code(sqlite3_context*, int);
   SQLITE_API void sqlite3_result_int(sqlite3_context*, int);
   SQLITE_API void sqlite3_result_int64(sqlite3_context*, sqlite3_int64);
   SQLITE_API void sqlite3_result_null(sqlite3_context*);
   SQLITE_API void sqlite3_result_text(sqlite3_context*, const char*, int, void(*)(void*));
   SQLITE_API void sqlite3_result_text16(sqlite3_context*, const void*, int, void(*)(void*));
   SQLITE_API void sqlite3_result_text16le(sqlite3_context*, const void*, int,void(*)(void*));
   SQLITE_API void sqlite3_result_text16be(sqlite3_context*, const void*, int,void(*)(void*));
   SQLITE_API void sqlite3_result_value(sqlite3_context*, sqlite3_value*);
   SQLITE_API void sqlite3_result_zeroblob(sqlite3_context*, int n);
   
   /*
   ** CAPI3REF: Define New Collating Sequences
   **
   ** ^These functions add, remove, or modify a [collation] associated
   ** with the [database connection] specified as the first argument.
   **
   ** ^The name of the collation is a UTF-8 string
   ** for sqlite3_create_collation() and sqlite3_create_collation_v2()
   ** and a UTF-16 string in native byte order for sqlite3_create_collation16().
   ** ^Collation names that compare equal according to [sqlite3_strnicmp()] are
   ** considered to be the same name.
   **
   ** ^(The third argument (eTextRep) must be one of the constants:
   ** <ul>
   ** <li> [SQLITE_UTF8],
   ** <li> [SQLITE_UTF16LE],
   ** <li> [SQLITE_UTF16BE],
   ** <li> [SQLITE_UTF16], or
   ** <li> [SQLITE_UTF16_ALIGNED].
   ** </ul>)^
   ** ^The eTextRep argument determines the encoding of strings passed
   ** to the collating function callback, xCallback.
   ** ^The [SQLITE_UTF16] and [SQLITE_UTF16_ALIGNED] values for eTextRep
   ** force strings to be UTF16 with native byte order.
   ** ^The [SQLITE_UTF16_ALIGNED] value for eTextRep forces strings to begin
   ** on an even byte address.
   **
   ** ^The fourth argument, pArg, is an application data pointer that is passed
   ** through as the first argument to the collating function callback.
   **
   ** ^The fifth argument, xCallback, is a pointer to the collating function.
   ** ^Multiple collating functions can be registered using the same name but
   ** with different eTextRep parameters and SQLite will use whichever
   ** function requires the least amount of data transformation.
   ** ^If the xCallback argument is NULL then the collating function is
   ** deleted.  ^When all collating functions having the same name are deleted,
   ** that collation is no longer usable.
   **
   ** ^The collating function callback is invoked with a copy of the pArg 
   ** application data pointer and with two strings in the encoding specified
   ** by the eTextRep argument.  The collating function must return an
   ** integer that is negative, zero, or positive
   ** if the first string is less than, equal to, or greater than the second,
   ** respectively.  A collating function must always return the same answer
   ** given the same inputs.  If two or more collating functions are registered
   ** to the same collation name (using different eTextRep values) then all
   ** must give an equivalent answer when invoked with equivalent strings.
   ** The collating function must obey the following properties for all
   ** strings A, B, and C:
   **
   ** <ol>
   ** <li> If A==B then B==A.
   ** <li> If A==B and B==C then A==C.
   ** <li> If A&lt;B THEN B&gt;A.
   ** <li> If A&lt;B and B&lt;C then A&lt;C.
   ** </ol>
   **
   ** If a collating function fails any of the above constraints and that
   ** collating function is  registered and used, then the behavior of SQLite
   ** is undefined.
   **
   ** ^The sqlite3_create_collation_v2() works like sqlite3_create_collation()
   ** with the addition that the xDestroy callback is invoked on pArg when
   ** the collating function is deleted.
   ** ^Collating functions are deleted when they are overridden by later
   ** calls to the collation creation functions or when the
   ** [database connection] is closed using [sqlite3_close()].
   **
   ** ^The xDestroy callback is <u>not</u> called if the 
   ** sqlite3_create_collation_v2() function fails.  Applications that invoke
   ** sqlite3_create_collation_v2() with a non-NULL xDestroy argument should 
   ** check the return code and dispose of the application data pointer
   ** themselves rather than expecting SQLite to deal with it for them.
   ** This is different from every other SQLite interface.  The inconsistency 
   ** is unfortunate but cannot be changed without breaking backwards 
   ** compatibility.
   **
   ** See also:  [sqlite3_collation_needed()] and [sqlite3_collation_needed16()].
   */
   SQLITE_API int sqlite3_create_collation(
     sqlite3*, 
     const char *zName, 
     int eTextRep, 
     void *pArg,
     int(*xCompare)(void*,int,const void*,int,const void*)
   );
   SQLITE_API int sqlite3_create_collation_v2(
     sqlite3*, 
     const char *zName, 
     int eTextRep, 
     void *pArg,
     int(*xCompare)(void*,int,const void*,int,const void*),
     void(*xDestroy)(void*)
   );
   SQLITE_API int sqlite3_create_collation16(
     sqlite3*, 
     const void *zName,
     int eTextRep, 
     void *pArg,
     int(*xCompare)(void*,int,const void*,int,const void*)
   );
   
   /*
   ** CAPI3REF: Collation Needed Callbacks
   **
   ** ^To avoid having to register all collation sequences before a database
   ** can be used, a single callback function may be registered with the
   ** [database connection] to be invoked whenever an undefined collation
   ** sequence is required.
   **
   ** ^If the function is registered using the sqlite3_collation_needed() API,
   ** then it is passed the names of undefined collation sequences as strings
   ** encoded in UTF-8. ^If sqlite3_collation_needed16() is used,
   ** the names are passed as UTF-16 in machine native byte order.
   ** ^A call to either function replaces the existing collation-needed callback.
   **
   ** ^(When the callback is invoked, the first argument passed is a copy
   ** of the second argument to sqlite3_collation_needed() or
   ** sqlite3_collation_needed16().  The second argument is the database
   ** connection.  The third argument is one of [SQLITE_UTF8], [SQLITE_UTF16BE],
   ** or [SQLITE_UTF16LE], indicating the most desirable form of the collation
   ** sequence function required.  The fourth parameter is the name of the
   ** required collation sequence.)^
   **
   ** The callback function should register the desired collation using
   ** [sqlite3_create_collation()], [sqlite3_create_collation16()], or
   ** [sqlite3_create_collation_v2()].
   */
   SQLITE_API int sqlite3_collation_needed(
     sqlite3*, 
     void*, 
     void(*)(void*,sqlite3*,int eTextRep,const char*)
   );
   SQLITE_API int sqlite3_collation_needed16(
     sqlite3*, 
     void*,
     void(*)(void*,sqlite3*,int eTextRep,const void*)
   );
   
   #ifdef SQLITE_HAS_CODEC
   /*
   ** Specify the key for an encrypted database.  This routine should be
   ** called right after sqlite3_open().
   **
   ** The code to implement this API is not available in the public release
   ** of SQLite.
   */
   SQLITE_API int sqlite3_key(
     sqlite3 *db,                   /* Database to be rekeyed */
     const void *pKey, int nKey     /* The key */
   );
   
   /*
   ** Change the key on an open database.  If the current database is not
   ** encrypted, this routine will encrypt it.  If pNew==0 or nNew==0, the
   ** database is decrypted.
   **
   ** The code to implement this API is not available in the public release
   ** of SQLite.
   */
   SQLITE_API int sqlite3_rekey(
     sqlite3 *db,                   /* Database to be rekeyed */
     const void *pKey, int nKey     /* The new key */
   );
   
   /*
   ** Specify the activation key for a SEE database.  Unless 
   ** activated, none of the SEE routines will work.
   */
   SQLITE_API void sqlite3_activate_see(
     const char *zPassPhrase        /* Activation phrase */
   );
   #endif
   
   #ifdef SQLITE_ENABLE_CEROD
   /*
   ** Specify the activation key for a CEROD database.  Unless 
   ** activated, none of the CEROD routines will work.
   */
   SQLITE_API void sqlite3_activate_cerod(
     const char *zPassPhrase        /* Activation phrase */
   );
   #endif
   
   /*
   ** CAPI3REF: Suspend Execution For A Short Time
   **
   ** The sqlite3_sleep() function causes the current thread to suspend execution
   ** for at least a number of milliseconds specified in its parameter.
   **
   ** If the operating system does not support sleep requests with
   ** millisecond time resolution, then the time will be rounded up to
   ** the nearest second. The number of milliseconds of sleep actually
   ** requested from the operating system is returned.
   **
   ** ^SQLite implements this interface by calling the xSleep()
   ** method of the default [sqlite3_vfs] object.  If the xSleep() method
   ** of the default VFS is not implemented correctly, or not implemented at
   ** all, then the behavior of sqlite3_sleep() may deviate from the description
   ** in the previous paragraphs.
   */
   SQLITE_API int sqlite3_sleep(int);
   
   /*
   ** CAPI3REF: Name Of The Folder Holding Temporary Files
   **
   ** ^(If this global variable is made to point to a string which is
   ** the name of a folder (a.k.a. directory), then all temporary files
   ** created by SQLite when using a built-in [sqlite3_vfs | VFS]
   ** will be placed in that directory.)^  ^If this variable
   ** is a NULL pointer, then SQLite performs a search for an appropriate
   ** temporary file directory.
   **
   ** It is not safe to read or modify this variable in more than one
   ** thread at a time.  It is not safe to read or modify this variable
   ** if a [database connection] is being used at the same time in a separate
   ** thread.
   ** It is intended that this variable be set once
   ** as part of process initialization and before any SQLite interface
   ** routines have been called and that this variable remain unchanged
   ** thereafter.
   **
   ** ^The [temp_store_directory pragma] may modify this variable and cause
   ** it to point to memory obtained from [sqlite3_malloc].  ^Furthermore,
   ** the [temp_store_directory pragma] always assumes that any string
   ** that this variable points to is held in memory obtained from 
   ** [sqlite3_malloc] and the pragma may attempt to free that memory
   ** using [sqlite3_free].
   ** Hence, if this variable is modified directly, either it should be
   ** made NULL or made to point to memory obtained from [sqlite3_malloc]
   ** or else the use of the [temp_store_directory pragma] should be avoided.
   */
   SQLITE_API SQLITE_EXTERN char *sqlite3_temp_directory;
   
   /*
   ** CAPI3REF: Test For Auto-Commit Mode
   ** KEYWORDS: {autocommit mode}
   **
   ** ^The sqlite3_get_autocommit() interface returns non-zero or
   ** zero if the given database connection is or is not in autocommit mode,
   ** respectively.  ^Autocommit mode is on by default.
   ** ^Autocommit mode is disabled by a [BEGIN] statement.
   ** ^Autocommit mode is re-enabled by a [COMMIT] or [ROLLBACK].
   **
   ** If certain kinds of errors occur on a statement within a multi-statement
   ** transaction (errors including [SQLITE_FULL], [SQLITE_IOERR],
   ** [SQLITE_NOMEM], [SQLITE_BUSY], and [SQLITE_INTERRUPT]) then the
   ** transaction might be rolled back automatically.  The only way to
   ** find out whether SQLite automatically rolled back the transaction after
   ** an error is to use this function.
   **
   ** If another thread changes the autocommit status of the database
   ** connection while this routine is running, then the return value
   ** is undefined.
   */
   SQLITE_API int sqlite3_get_autocommit(sqlite3*);
   
   /*
   ** CAPI3REF: Find The Database Handle Of A Prepared Statement
   **
   ** ^The sqlite3_db_handle interface returns the [database connection] handle
   ** to which a [prepared statement] belongs.  ^The [database connection]
   ** returned by sqlite3_db_handle is the same [database connection]
   ** that was the first argument
   ** to the [sqlite3_prepare_v2()] call (or its variants) that was used to
   ** create the statement in the first place.
   */
   SQLITE_API sqlite3 *sqlite3_db_handle(sqlite3_stmt*);
   
   /*
   ** CAPI3REF: Return The Filename For A Database Connection
   **
   ** ^The sqlite3_db_filename(D,N) interface returns a pointer to a filename
   ** associated with database N of connection D.  ^The main database file
   ** has the name "main".  If there is no attached database N on the database
   ** connection D, or if database N is a temporary or in-memory database, then
   ** a NULL pointer is returned.
   **
   ** ^The filename returned by this function is the output of the
   ** xFullPathname method of the [VFS].  ^In other words, the filename
   ** will be an absolute pathname, even if the filename used
   ** to open the database originally was a URI or relative pathname.
   */
   SQLITE_API const char *sqlite3_db_filename(sqlite3 *db, const char *zDbName);
   
   /*
   ** CAPI3REF: Find the next prepared statement
   **
   ** ^This interface returns a pointer to the next [prepared statement] after
   ** pStmt associated with the [database connection] pDb.  ^If pStmt is NULL
   ** then this interface returns a pointer to the first prepared statement
   ** associated with the database connection pDb.  ^If no prepared statement
   ** satisfies the conditions of this routine, it returns NULL.
   **
   ** The [database connection] pointer D in a call to
   ** [sqlite3_next_stmt(D,S)] must refer to an open database
   ** connection and in particular must not be a NULL pointer.
   */
   SQLITE_API sqlite3_stmt *sqlite3_next_stmt(sqlite3 *pDb, sqlite3_stmt *pStmt);
   
   /*
   ** CAPI3REF: Commit And Rollback Notification Callbacks
   **
   ** ^The sqlite3_commit_hook() interface registers a callback
   ** function to be invoked whenever a transaction is [COMMIT | committed].
   ** ^Any callback set by a previous call to sqlite3_commit_hook()
   ** for the same database connection is overridden.
   ** ^The sqlite3_rollback_hook() interface registers a callback
   ** function to be invoked whenever a transaction is [ROLLBACK | rolled back].
   ** ^Any callback set by a previous call to sqlite3_rollback_hook()
   ** for the same database connection is overridden.
   ** ^The pArg argument is passed through to the callback.
   ** ^If the callback on a commit hook function returns non-zero,
   ** then the commit is converted into a rollback.
   **
   ** ^The sqlite3_commit_hook(D,C,P) and sqlite3_rollback_hook(D,C,P) functions
   ** return the P argument from the previous call of the same function
   ** on the same [database connection] D, or NULL for
   ** the first call for each function on D.
   **
   ** The commit and rollback hook callbacks are not reentrant.
   ** The callback implementation must not do anything that will modify
   ** the database connection that invoked the callback.  Any actions
   ** to modify the database connection must be deferred until after the
   ** completion of the [sqlite3_step()] call that triggered the commit
   ** or rollback hook in the first place.
   ** Note that running any other SQL statements, including SELECT statements,
   ** or merely calling [sqlite3_prepare_v2()] and [sqlite3_step()] will modify
   ** the database connections for the meaning of "modify" in this paragraph.
   **
   ** ^Registering a NULL function disables the callback.
   **
   ** ^When the commit hook callback routine returns zero, the [COMMIT]
   ** operation is allowed to continue normally.  ^If the commit hook
   ** returns non-zero, then the [COMMIT] is converted into a [ROLLBACK].
   ** ^The rollback hook is invoked on a rollback that results from a commit
   ** hook returning non-zero, just as it would be with any other rollback.
   **
   ** ^For the purposes of this API, a transaction is said to have been
   ** rolled back if an explicit "ROLLBACK" statement is executed, or
   ** an error or constraint causes an implicit rollback to occur.
   ** ^The rollback callback is not invoked if a transaction is
   ** automatically rolled back because the database connection is closed.
   **
   ** See also the [sqlite3_update_hook()] interface.
   */
   SQLITE_API void *sqlite3_commit_hook(sqlite3*, int(*)(void*), void*);
   SQLITE_API void *sqlite3_rollback_hook(sqlite3*, void(*)(void *), void*);
   
   /*
   ** CAPI3REF: Data Change Notification Callbacks
   **
   ** ^The sqlite3_update_hook() interface registers a callback function
   ** with the [database connection] identified by the first argument
   ** to be invoked whenever a row is updated, inserted or deleted.
   ** ^Any callback set by a previous call to this function
   ** for the same database connection is overridden.
   **
   ** ^The second argument is a pointer to the function to invoke when a
   ** row is updated, inserted or deleted.
   ** ^The first argument to the callback is a copy of the third argument
   ** to sqlite3_update_hook().
   ** ^The second callback argument is one of [SQLITE_INSERT], [SQLITE_DELETE],
   ** or [SQLITE_UPDATE], depending on the operation that caused the callback
   ** to be invoked.
   ** ^The third and fourth arguments to the callback contain pointers to the
   ** database and table name containing the affected row.
   ** ^The final callback parameter is the [rowid] of the row.
   ** ^In the case of an update, this is the [rowid] after the update takes place.
   **
   ** ^(The update hook is not invoked when internal system tables are
   ** modified (i.e. sqlite_master and sqlite_sequence).)^
   **
   ** ^In the current implementation, the update hook
   ** is not invoked when duplication rows are deleted because of an
   ** [ON CONFLICT | ON CONFLICT REPLACE] clause.  ^Nor is the update hook
   ** invoked when rows are deleted using the [truncate optimization].
   ** The exceptions defined in this paragraph might change in a future
   ** release of SQLite.
   **
   ** The update hook implementation must not do anything that will modify
   ** the database connection that invoked the update hook.  Any actions
   ** to modify the database connection must be deferred until after the
   ** completion of the [sqlite3_step()] call that triggered the update hook.
   ** Note that [sqlite3_prepare_v2()] and [sqlite3_step()] both modify their
   ** database connections for the meaning of "modify" in this paragraph.
   **
   ** ^The sqlite3_update_hook(D,C,P) function
   ** returns the P argument from the previous call
   ** on the same [database connection] D, or NULL for
   ** the first call on D.
   **
   ** See also the [sqlite3_commit_hook()] and [sqlite3_rollback_hook()]
   ** interfaces.
   */
   SQLITE_API void *sqlite3_update_hook(
     sqlite3*, 
     void(*)(void *,int ,char const *,char const *,sqlite3_int64),
     void*
   );
   
   /*
   ** CAPI3REF: Enable Or Disable Shared Pager Cache
   ** KEYWORDS: {shared cache}
   **
   ** ^(This routine enables or disables the sharing of the database cache
   ** and schema data structures between [database connection | connections]
   ** to the same database. Sharing is enabled if the argument is true
   ** and disabled if the argument is false.)^
   **
   ** ^Cache sharing is enabled and disabled for an entire process.
   ** This is a change as of SQLite version 3.5.0. In prior versions of SQLite,
   ** sharing was enabled or disabled for each thread separately.
   **
   ** ^(The cache sharing mode set by this interface effects all subsequent
   ** calls to [sqlite3_open()], [sqlite3_open_v2()], and [sqlite3_open16()].
   ** Existing database connections continue use the sharing mode
   ** that was in effect at the time they were opened.)^
   **
   ** ^(This routine returns [SQLITE_OK] if shared cache was enabled or disabled
   ** successfully.  An [error code] is returned otherwise.)^
   **
   ** ^Shared cache is disabled by default. But this might change in
   ** future releases of SQLite.  Applications that care about shared
   ** cache setting should set it explicitly.
   **
   ** See Also:  [SQLite Shared-Cache Mode]
   */
   SQLITE_API int sqlite3_enable_shared_cache(int);
   
   /*
   ** CAPI3REF: Attempt To Free Heap Memory
   **
   ** ^The sqlite3_release_memory() interface attempts to free N bytes
   ** of heap memory by deallocating non-essential memory allocations
   ** held by the database library.   Memory used to cache database
   ** pages to improve performance is an example of non-essential memory.
   ** ^sqlite3_release_memory() returns the number of bytes actually freed,
   ** which might be more or less than the amount requested.
   ** ^The sqlite3_release_memory() routine is a no-op returning zero
   ** if SQLite is not compiled with [SQLITE_ENABLE_MEMORY_MANAGEMENT].
   **
   ** See also: [sqlite3_db_release_memory()]
   */
   SQLITE_API int sqlite3_release_memory(int);
   
   /*
   ** CAPI3REF: Free Memory Used By A Database Connection
   **
   ** ^The sqlite3_db_release_memory(D) interface attempts to free as much heap
   ** memory as possible from database connection D. Unlike the
   ** [sqlite3_release_memory()] interface, this interface is effect even
   ** when then [SQLITE_ENABLE_MEMORY_MANAGEMENT] compile-time option is
   ** omitted.
   **
   ** See also: [sqlite3_release_memory()]
   */
   SQLITE_API int sqlite3_db_release_memory(sqlite3*);
   
   /*
   ** CAPI3REF: Impose A Limit On Heap Size
   **
   ** ^The sqlite3_soft_heap_limit64() interface sets and/or queries the
   ** soft limit on the amount of heap memory that may be allocated by SQLite.
   ** ^SQLite strives to keep heap memory utilization below the soft heap
   ** limit by reducing the number of pages held in the page cache
   ** as heap memory usages approaches the limit.
   ** ^The soft heap limit is "soft" because even though SQLite strives to stay
   ** below the limit, it will exceed the limit rather than generate
   ** an [SQLITE_NOMEM] error.  In other words, the soft heap limit 
   ** is advisory only.
   **
   ** ^The return value from sqlite3_soft_heap_limit64() is the size of
   ** the soft heap limit prior to the call, or negative in the case of an
   ** error.  ^If the argument N is negative
   ** then no change is made to the soft heap limit.  Hence, the current
   ** size of the soft heap limit can be determined by invoking
   ** sqlite3_soft_heap_limit64() with a negative argument.
   **
   ** ^If the argument N is zero then the soft heap limit is disabled.
   **
   ** ^(The soft heap limit is not enforced in the current implementation
   ** if one or more of following conditions are true:
   **
   ** <ul>
   ** <li> The soft heap limit is set to zero.
   ** <li> Memory accounting is disabled using a combination of the
   **      [sqlite3_config]([SQLITE_CONFIG_MEMSTATUS],...) start-time option and
   **      the [SQLITE_DEFAULT_MEMSTATUS] compile-time option.
   ** <li> An alternative page cache implementation is specified using
   **      [sqlite3_config]([SQLITE_CONFIG_PCACHE2],...).
   ** <li> The page cache allocates from its own memory pool supplied
   **      by [sqlite3_config]([SQLITE_CONFIG_PAGECACHE],...) rather than
   **      from the heap.
   ** </ul>)^
   **
   ** Beginning with SQLite version 3.7.3, the soft heap limit is enforced
   ** regardless of whether or not the [SQLITE_ENABLE_MEMORY_MANAGEMENT]
   ** compile-time option is invoked.  With [SQLITE_ENABLE_MEMORY_MANAGEMENT],
   ** the soft heap limit is enforced on every memory allocation.  Without
   ** [SQLITE_ENABLE_MEMORY_MANAGEMENT], the soft heap limit is only enforced
   ** when memory is allocated by the page cache.  Testing suggests that because
   ** the page cache is the predominate memory user in SQLite, most
   ** applications will achieve adequate soft heap limit enforcement without
   ** the use of [SQLITE_ENABLE_MEMORY_MANAGEMENT].
   **
   ** The circumstances under which SQLite will enforce the soft heap limit may
   ** changes in future releases of SQLite.
   */
   SQLITE_API sqlite3_int64 sqlite3_soft_heap_limit64(sqlite3_int64 N);
   
   /*
   ** CAPI3REF: Deprecated Soft Heap Limit Interface
   ** DEPRECATED
   **
   ** This is a deprecated version of the [sqlite3_soft_heap_limit64()]
   ** interface.  This routine is provided for historical compatibility
   ** only.  All new applications should use the
   ** [sqlite3_soft_heap_limit64()] interface rather than this one.
   */
   SQLITE_API SQLITE_DEPRECATED void sqlite3_soft_heap_limit(int N);
   
   
   /*
   ** CAPI3REF: Extract Metadata About A Column Of A Table
   **
   ** ^This routine returns metadata about a specific column of a specific
   ** database table accessible using the [database connection] handle
   ** passed as the first function argument.
   **
   ** ^The column is identified by the second, third and fourth parameters to
   ** this function. ^The second parameter is either the name of the database
   ** (i.e. "main", "temp", or an attached database) containing the specified
   ** table or NULL. ^If it is NULL, then all attached databases are searched
   ** for the table using the same algorithm used by the database engine to
   ** resolve unqualified table references.
   **
   ** ^The third and fourth parameters to this function are the table and column
   ** name of the desired column, respectively. Neither of these parameters
   ** may be NULL.
   **
   ** ^Metadata is returned by writing to the memory locations passed as the 5th
   ** and subsequent parameters to this function. ^Any of these arguments may be
   ** NULL, in which case the corresponding element of metadata is omitted.
   **
   ** ^(<blockquote>
   ** <table border="1">
   ** <tr><th> Parameter <th> Output<br>Type <th>  Description
   **
   ** <tr><td> 5th <td> const char* <td> Data type
   ** <tr><td> 6th <td> const char* <td> Name of default collation sequence
   ** <tr><td> 7th <td> int         <td> True if column has a NOT NULL constraint
   ** <tr><td> 8th <td> int         <td> True if column is part of the PRIMARY KEY
   ** <tr><td> 9th <td> int         <td> True if column is [AUTOINCREMENT]
   ** </table>
   ** </blockquote>)^
   **
   ** ^The memory pointed to by the character pointers returned for the
   ** declaration type and collation sequence is valid only until the next
   ** call to any SQLite API function.
   **
   ** ^If the specified table is actually a view, an [error code] is returned.
   **
   ** ^If the specified column is "rowid", "oid" or "_rowid_" and an
   ** [INTEGER PRIMARY KEY] column has been explicitly declared, then the output
   ** parameters are set for the explicitly declared column. ^(If there is no
   ** explicitly declared [INTEGER PRIMARY KEY] column, then the output
   ** parameters are set as follows:
   **
   ** <pre>
   **     data type: "INTEGER"
   **     collation sequence: "BINARY"
   **     not null: 0
   **     primary key: 1
   **     auto increment: 0
   ** </pre>)^
   **
   ** ^(This function may load one or more schemas from database files. If an
   ** error occurs during this process, or if the requested table or column
   ** cannot be found, an [error code] is returned and an error message left
   ** in the [database connection] (to be retrieved using sqlite3_errmsg()).)^
   **
   ** ^This API is only available if the library was compiled with the
   ** [SQLITE_ENABLE_COLUMN_METADATA] C-preprocessor symbol defined.
   */
   SQLITE_API int sqlite3_table_column_metadata(
     sqlite3 *db,                /* Connection handle */
     const char *zDbName,        /* Database name or NULL */
     const char *zTableName,     /* Table name */
     const char *zColumnName,    /* Column name */
     char const **pzDataType,    /* OUTPUT: Declared data type */
     char const **pzCollSeq,     /* OUTPUT: Collation sequence name */
     int *pNotNull,              /* OUTPUT: True if NOT NULL constraint exists */
     int *pPrimaryKey,           /* OUTPUT: True if column part of PK */
     int *pAutoinc               /* OUTPUT: True if column is auto-increment */
   );
   
   /*
   ** CAPI3REF: Load An Extension
   **
   ** ^This interface loads an SQLite extension library from the named file.
   **
   ** ^The sqlite3_load_extension() interface attempts to load an
   ** SQLite extension library contained in the file zFile.
   **
   ** ^The entry point is zProc.
   ** ^zProc may be 0, in which case the name of the entry point
   ** defaults to "sqlite3_extension_init".
   ** ^The sqlite3_load_extension() interface returns
   ** [SQLITE_OK] on success and [SQLITE_ERROR] if something goes wrong.
   ** ^If an error occurs and pzErrMsg is not 0, then the
   ** [sqlite3_load_extension()] interface shall attempt to
   ** fill *pzErrMsg with error message text stored in memory
   ** obtained from [sqlite3_malloc()]. The calling function
   ** should free this memory by calling [sqlite3_free()].
   **
   ** ^Extension loading must be enabled using
   ** [sqlite3_enable_load_extension()] prior to calling this API,
   ** otherwise an error will be returned.
   **
   ** See also the [load_extension() SQL function].
   */
   SQLITE_API int sqlite3_load_extension(
     sqlite3 *db,          /* Load the extension into this database connection */
     const char *zFile,    /* Name of the shared library containing extension */
     const char *zProc,    /* Entry point.  Derived from zFile if 0 */
     char **pzErrMsg       /* Put error message here if not 0 */
   );
   
   /*
   ** CAPI3REF: Enable Or Disable Extension Loading
   **
   ** ^So as not to open security holes in older applications that are
   ** unprepared to deal with extension loading, and as a means of disabling
   ** extension loading while evaluating user-entered SQL, the following API
   ** is provided to turn the [sqlite3_load_extension()] mechanism on and off.
   **
   ** ^Extension loading is off by default. See ticket #1863.
   ** ^Call the sqlite3_enable_load_extension() routine with onoff==1
   ** to turn extension loading on and call it with onoff==0 to turn
   ** it back off again.
   */
   SQLITE_API int sqlite3_enable_load_extension(sqlite3 *db, int onoff);
   
   /*
   ** CAPI3REF: Automatically Load Statically Linked Extensions
   **
   ** ^This interface causes the xEntryPoint() function to be invoked for
   ** each new [database connection] that is created.  The idea here is that
   ** xEntryPoint() is the entry point for a statically linked SQLite extension
   ** that is to be automatically loaded into all new database connections.
   **
   ** ^(Even though the function prototype shows that xEntryPoint() takes
   ** no arguments and returns void, SQLite invokes xEntryPoint() with three
   ** arguments and expects and integer result as if the signature of the
   ** entry point where as follows:
   **
   ** <blockquote><pre>
   ** &nbsp;  int xEntryPoint(
   ** &nbsp;    sqlite3 *db,
   ** &nbsp;    const char **pzErrMsg,
   ** &nbsp;    const struct sqlite3_api_routines *pThunk
   ** &nbsp;  );
   ** </pre></blockquote>)^
   **
   ** If the xEntryPoint routine encounters an error, it should make *pzErrMsg
   ** point to an appropriate error message (obtained from [sqlite3_mprintf()])
   ** and return an appropriate [error code].  ^SQLite ensures that *pzErrMsg
   ** is NULL before calling the xEntryPoint().  ^SQLite will invoke
   ** [sqlite3_free()] on *pzErrMsg after xEntryPoint() returns.  ^If any
   ** xEntryPoint() returns an error, the [sqlite3_open()], [sqlite3_open16()],
   ** or [sqlite3_open_v2()] call that provoked the xEntryPoint() will fail.
   **
   ** ^Calling sqlite3_auto_extension(X) with an entry point X that is already
   ** on the list of automatic extensions is a harmless no-op. ^No entry point
   ** will be called more than once for each database connection that is opened.
   **
   ** See also: [sqlite3_reset_auto_extension()].
   */
   SQLITE_API int sqlite3_auto_extension(void (*xEntryPoint)(void));
   
   /*
   ** CAPI3REF: Reset Automatic Extension Loading
   **
   ** ^This interface disables all automatic extensions previously
   ** registered using [sqlite3_auto_extension()].
   */
   SQLITE_API void sqlite3_reset_auto_extension(void);
   
   /*
   ** The interface to the virtual-table mechanism is currently considered
   ** to be experimental.  The interface might change in incompatible ways.
   ** If this is a problem for you, do not use the interface at this time.
   **
   ** When the virtual-table mechanism stabilizes, we will declare the
   ** interface fixed, support it indefinitely, and remove this comment.
   */
   
   /*
   ** Structures used by the virtual table interface
   */
   typedef struct sqlite3_vtab sqlite3_vtab;
   typedef struct sqlite3_index_info sqlite3_index_info;
   typedef struct sqlite3_vtab_cursor sqlite3_vtab_cursor;
   typedef struct sqlite3_module sqlite3_module;
   
   /*
   ** CAPI3REF: Virtual Table Object
   ** KEYWORDS: sqlite3_module {virtual table module}
   **
   ** This structure, sometimes called a "virtual table module", 
   ** defines the implementation of a [virtual tables].  
   ** This structure consists mostly of methods for the module.
   **
   ** ^A virtual table module is created by filling in a persistent
   ** instance of this structure and passing a pointer to that instance
   ** to [sqlite3_create_module()] or [sqlite3_create_module_v2()].
   ** ^The registration remains valid until it is replaced by a different
   ** module or until the [database connection] closes.  The content
   ** of this structure must not change while it is registered with
   ** any database connection.
   */
   struct sqlite3_module {
     int iVersion;
     int (*xCreate)(sqlite3*, void *pAux,
                  int argc, const char *const*argv,
                  sqlite3_vtab **ppVTab, char**);
     int (*xConnect)(sqlite3*, void *pAux,
                  int argc, const char *const*argv,
                  sqlite3_vtab **ppVTab, char**);
     int (*xBestIndex)(sqlite3_vtab *pVTab, sqlite3_index_info*);
     int (*xDisconnect)(sqlite3_vtab *pVTab);
     int (*xDestroy)(sqlite3_vtab *pVTab);
     int (*xOpen)(sqlite3_vtab *pVTab, sqlite3_vtab_cursor **ppCursor);
     int (*xClose)(sqlite3_vtab_cursor*);
     int (*xFilter)(sqlite3_vtab_cursor*, int idxNum, const char *idxStr,
                   int argc, sqlite3_value **argv);
     int (*xNext)(sqlite3_vtab_cursor*);
     int (*xEof)(sqlite3_vtab_cursor*);
     int (*xColumn)(sqlite3_vtab_cursor*, sqlite3_context*, int);
     int (*xRowid)(sqlite3_vtab_cursor*, sqlite3_int64 *pRowid);
     int (*xUpdate)(sqlite3_vtab *, int, sqlite3_value **, sqlite3_int64 *);
     int (*xBegin)(sqlite3_vtab *pVTab);
     int (*xSync)(sqlite3_vtab *pVTab);
     int (*xCommit)(sqlite3_vtab *pVTab);
     int (*xRollback)(sqlite3_vtab *pVTab);
     int (*xFindFunction)(sqlite3_vtab *pVtab, int nArg, const char *zName,
                          void (**pxFunc)(sqlite3_context*,int,sqlite3_value**),
                          void **ppArg);
     int (*xRename)(sqlite3_vtab *pVtab, const char *zNew);
     /* The methods above are in version 1 of the sqlite_module object. Those 
     ** below are for version 2 and greater. */
     int (*xSavepoint)(sqlite3_vtab *pVTab, int);
     int (*xRelease)(sqlite3_vtab *pVTab, int);
     int (*xRollbackTo)(sqlite3_vtab *pVTab, int);
   };
   
   /*
   ** CAPI3REF: Virtual Table Indexing Information
   ** KEYWORDS: sqlite3_index_info
   **
   ** The sqlite3_index_info structure and its substructures is used as part
   ** of the [virtual table] interface to
   ** pass information into and receive the reply from the [xBestIndex]
   ** method of a [virtual table module].  The fields under **Inputs** are the
   ** inputs to xBestIndex and are read-only.  xBestIndex inserts its
   ** results into the **Outputs** fields.
   **
   ** ^(The aConstraint[] array records WHERE clause constraints of the form:
   **
   ** <blockquote>column OP expr</blockquote>
   **
   ** where OP is =, &lt;, &lt;=, &gt;, or &gt;=.)^  ^(The particular operator is
   ** stored in aConstraint[].op using one of the
   ** [SQLITE_INDEX_CONSTRAINT_EQ | SQLITE_INDEX_CONSTRAINT_ values].)^
   ** ^(The index of the column is stored in
   ** aConstraint[].iColumn.)^  ^(aConstraint[].usable is TRUE if the
   ** expr on the right-hand side can be evaluated (and thus the constraint
   ** is usable) and false if it cannot.)^
   **
   ** ^The optimizer automatically inverts terms of the form "expr OP column"
   ** and makes other simplifications to the WHERE clause in an attempt to
   ** get as many WHERE clause terms into the form shown above as possible.
   ** ^The aConstraint[] array only reports WHERE clause terms that are
   ** relevant to the particular virtual table being queried.
   **
   ** ^Information about the ORDER BY clause is stored in aOrderBy[].
   ** ^Each term of aOrderBy records a column of the ORDER BY clause.
   **
   ** The [xBestIndex] method must fill aConstraintUsage[] with information
   ** about what parameters to pass to xFilter.  ^If argvIndex>0 then
   ** the right-hand side of the corresponding aConstraint[] is evaluated
   ** and becomes the argvIndex-th entry in argv.  ^(If aConstraintUsage[].omit
   ** is true, then the constraint is assumed to be fully handled by the
   ** virtual table and is not checked again by SQLite.)^
   **
   ** ^The idxNum and idxPtr values are recorded and passed into the
   ** [xFilter] method.
   ** ^[sqlite3_free()] is used to free idxPtr if and only if
   ** needToFreeIdxPtr is true.
   **
   ** ^The orderByConsumed means that output from [xFilter]/[xNext] will occur in
   ** the correct order to satisfy the ORDER BY clause so that no separate
   ** sorting step is required.
   **
   ** ^The estimatedCost value is an estimate of the cost of doing the
   ** particular lookup.  A full scan of a table with N entries should have
   ** a cost of N.  A binary search of a table of N entries should have a
   ** cost of approximately log(N).
   */
   struct sqlite3_index_info {
     /* Inputs */
     int nConstraint;           /* Number of entries in aConstraint */
     struct sqlite3_index_constraint {
        int iColumn;              /* Column on left-hand side of constraint */
        unsigned char op;         /* Constraint operator */
        unsigned char usable;     /* True if this constraint is usable */
        int iTermOffset;          /* Used internally - xBestIndex should ignore */
     } *aConstraint;            /* Table of WHERE clause constraints */
     int nOrderBy;              /* Number of terms in the ORDER BY clause */
     struct sqlite3_index_orderby {
        int iColumn;              /* Column number */
        unsigned char desc;       /* True for DESC.  False for ASC. */
     } *aOrderBy;               /* The ORDER BY clause */
     /* Outputs */
     struct sqlite3_index_constraint_usage {
       int argvIndex;           /* if >0, constraint is part of argv to xFilter */
       unsigned char omit;      /* Do not code a test for this constraint */
     } *aConstraintUsage;
     int idxNum;                /* Number used to identify the index */
     char *idxStr;              /* String, possibly obtained from sqlite3_malloc */
     int needToFreeIdxStr;      /* Free idxStr using sqlite3_free() if true */
     int orderByConsumed;       /* True if output is already ordered */
     double estimatedCost;      /* Estimated cost of using this index */
   };
   
   /*
   ** CAPI3REF: Virtual Table Constraint Operator Codes
   **
   ** These macros defined the allowed values for the
   ** [sqlite3_index_info].aConstraint[].op field.  Each value represents
   ** an operator that is part of a constraint term in the wHERE clause of
   ** a query that uses a [virtual table].
   */
   #define SQLITE_INDEX_CONSTRAINT_EQ    2
   #define SQLITE_INDEX_CONSTRAINT_GT    4
   #define SQLITE_INDEX_CONSTRAINT_LE    8
   #define SQLITE_INDEX_CONSTRAINT_LT    16
   #define SQLITE_INDEX_CONSTRAINT_GE    32
   #define SQLITE_INDEX_CONSTRAINT_MATCH 64
   
   /*
   ** CAPI3REF: Register A Virtual Table Implementation
   **
   ** ^These routines are used to register a new [virtual table module] name.
   ** ^Module names must be registered before
   ** creating a new [virtual table] using the module and before using a
   ** preexisting [virtual table] for the module.
   **
   ** ^The module name is registered on the [database connection] specified
   ** by the first parameter.  ^The name of the module is given by the 
   ** second parameter.  ^The third parameter is a pointer to
   ** the implementation of the [virtual table module].   ^The fourth
   ** parameter is an arbitrary client data pointer that is passed through
   ** into the [xCreate] and [xConnect] methods of the virtual table module
   ** when a new virtual table is be being created or reinitialized.
   **
   ** ^The sqlite3_create_module_v2() interface has a fifth parameter which
   ** is a pointer to a destructor for the pClientData.  ^SQLite will
   ** invoke the destructor function (if it is not NULL) when SQLite
   ** no longer needs the pClientData pointer.  ^The destructor will also
   ** be invoked if the call to sqlite3_create_module_v2() fails.
   ** ^The sqlite3_create_module()
   ** interface is equivalent to sqlite3_create_module_v2() with a NULL
   ** destructor.
   */
   SQLITE_API int sqlite3_create_module(
     sqlite3 *db,               /* SQLite connection to register module with */
     const char *zName,         /* Name of the module */
     const sqlite3_module *p,   /* Methods for the module */
     void *pClientData          /* Client data for xCreate/xConnect */
   );
   SQLITE_API int sqlite3_create_module_v2(
     sqlite3 *db,               /* SQLite connection to register module with */
     const char *zName,         /* Name of the module */
     const sqlite3_module *p,   /* Methods for the module */
     void *pClientData,         /* Client data for xCreate/xConnect */
     void(*xDestroy)(void*)     /* Module destructor function */
   );
   
   /*
   ** CAPI3REF: Virtual Table Instance Object
   ** KEYWORDS: sqlite3_vtab
   **
   ** Every [virtual table module] implementation uses a subclass
   ** of this object to describe a particular instance
   ** of the [virtual table].  Each subclass will
   ** be tailored to the specific needs of the module implementation.
   ** The purpose of this superclass is to define certain fields that are
   ** common to all module implementations.
   **
   ** ^Virtual tables methods can set an error message by assigning a
   ** string obtained from [sqlite3_mprintf()] to zErrMsg.  The method should
   ** take care that any prior string is freed by a call to [sqlite3_free()]
   ** prior to assigning a new string to zErrMsg.  ^After the error message
   ** is delivered up to the client application, the string will be automatically
   ** freed by sqlite3_free() and the zErrMsg field will be zeroed.
   */
   struct sqlite3_vtab {
     const sqlite3_module *pModule;  /* The module for this virtual table */
     int nRef;                       /* NO LONGER USED */
     char *zErrMsg;                  /* Error message from sqlite3_mprintf() */
     /* Virtual table implementations will typically add additional fields */
   };
   
   /*
   ** CAPI3REF: Virtual Table Cursor Object
   ** KEYWORDS: sqlite3_vtab_cursor {virtual table cursor}
   **
   ** Every [virtual table module] implementation uses a subclass of the
   ** following structure to describe cursors that point into the
   ** [virtual table] and are used
   ** to loop through the virtual table.  Cursors are created using the
   ** [sqlite3_module.xOpen | xOpen] method of the module and are destroyed
   ** by the [sqlite3_module.xClose | xClose] method.  Cursors are used
   ** by the [xFilter], [xNext], [xEof], [xColumn], and [xRowid] methods
   ** of the module.  Each module implementation will define
   ** the content of a cursor structure to suit its own needs.
   **
   ** This superclass exists in order to define fields of the cursor that
   ** are common to all implementations.
   */
   struct sqlite3_vtab_cursor {
     sqlite3_vtab *pVtab;      /* Virtual table of this cursor */
     /* Virtual table implementations will typically add additional fields */
   };
   
   /*
   ** CAPI3REF: Declare The Schema Of A Virtual Table
   **
   ** ^The [xCreate] and [xConnect] methods of a
   ** [virtual table module] call this interface
   ** to declare the format (the names and datatypes of the columns) of
   ** the virtual tables they implement.
   */
   SQLITE_API int sqlite3_declare_vtab(sqlite3*, const char *zSQL);
   
   /*
   ** CAPI3REF: Overload A Function For A Virtual Table
   **
   ** ^(Virtual tables can provide alternative implementations of functions
   ** using the [xFindFunction] method of the [virtual table module].  
   ** But global versions of those functions
   ** must exist in order to be overloaded.)^
   **
   ** ^(This API makes sure a global version of a function with a particular
   ** name and number of parameters exists.  If no such function exists
   ** before this API is called, a new function is created.)^  ^The implementation
   ** of the new function always causes an exception to be thrown.  So
   ** the new function is not good for anything by itself.  Its only
   ** purpose is to be a placeholder function that can be overloaded
   ** by a [virtual table].
   */
   SQLITE_API int sqlite3_overload_function(sqlite3*, const char *zFuncName, int nArg);
   
   /*
   ** The interface to the virtual-table mechanism defined above (back up
   ** to a comment remarkably similar to this one) is currently considered
   ** to be experimental.  The interface might change in incompatible ways.
   ** If this is a problem for you, do not use the interface at this time.
   **
   ** When the virtual-table mechanism stabilizes, we will declare the
   ** interface fixed, support it indefinitely, and remove this comment.
   */
   
   /*
   ** CAPI3REF: A Handle To An Open BLOB
   ** KEYWORDS: {BLOB handle} {BLOB handles}
   **
   ** An instance of this object represents an open BLOB on which
   ** [sqlite3_blob_open | incremental BLOB I/O] can be performed.
   ** ^Objects of this type are created by [sqlite3_blob_open()]
   ** and destroyed by [sqlite3_blob_close()].
   ** ^The [sqlite3_blob_read()] and [sqlite3_blob_write()] interfaces
   ** can be used to read or write small subsections of the BLOB.
   ** ^The [sqlite3_blob_bytes()] interface returns the size of the BLOB in bytes.
   */
   typedef struct sqlite3_blob sqlite3_blob;
   
   /*
   ** CAPI3REF: Open A BLOB For Incremental I/O
   **
   ** ^(This interfaces opens a [BLOB handle | handle] to the BLOB located
   ** in row iRow, column zColumn, table zTable in database zDb;
   ** in other words, the same BLOB that would be selected by:
   **
   ** <pre>
   **     SELECT zColumn FROM zDb.zTable WHERE [rowid] = iRow;
   ** </pre>)^
   **
   ** ^If the flags parameter is non-zero, then the BLOB is opened for read
   ** and write access. ^If it is zero, the BLOB is opened for read access.
   ** ^It is not possible to open a column that is part of an index or primary 
   ** key for writing. ^If [foreign key constraints] are enabled, it is 
   ** not possible to open a column that is part of a [child key] for writing.
   **
   ** ^Note that the database name is not the filename that contains
   ** the database but rather the symbolic name of the database that
   ** appears after the AS keyword when the database is connected using [ATTACH].
   ** ^For the main database file, the database name is "main".
   ** ^For TEMP tables, the database name is "temp".
   **
   ** ^(On success, [SQLITE_OK] is returned and the new [BLOB handle] is written
   ** to *ppBlob. Otherwise an [error code] is returned and *ppBlob is set
   ** to be a null pointer.)^
   ** ^This function sets the [database connection] error code and message
   ** accessible via [sqlite3_errcode()] and [sqlite3_errmsg()] and related
   ** functions. ^Note that the *ppBlob variable is always initialized in a
   ** way that makes it safe to invoke [sqlite3_blob_close()] on *ppBlob
   ** regardless of the success or failure of this routine.
   **
   ** ^(If the row that a BLOB handle points to is modified by an
   ** [UPDATE], [DELETE], or by [ON CONFLICT] side-effects
   ** then the BLOB handle is marked as "expired".
   ** This is true if any column of the row is changed, even a column
   ** other than the one the BLOB handle is open on.)^
   ** ^Calls to [sqlite3_blob_read()] and [sqlite3_blob_write()] for
   ** an expired BLOB handle fail with a return code of [SQLITE_ABORT].
   ** ^(Changes written into a BLOB prior to the BLOB expiring are not
   ** rolled back by the expiration of the BLOB.  Such changes will eventually
   ** commit if the transaction continues to completion.)^
   **
   ** ^Use the [sqlite3_blob_bytes()] interface to determine the size of
   ** the opened blob.  ^The size of a blob may not be changed by this
   ** interface.  Use the [UPDATE] SQL command to change the size of a
   ** blob.
   **
   ** ^The [sqlite3_bind_zeroblob()] and [sqlite3_result_zeroblob()] interfaces
   ** and the built-in [zeroblob] SQL function can be used, if desired,
   ** to create an empty, zero-filled blob in which to read or write using
   ** this interface.
   **
   ** To avoid a resource leak, every open [BLOB handle] should eventually
   ** be released by a call to [sqlite3_blob_close()].
   */
   SQLITE_API int sqlite3_blob_open(
     sqlite3*,
     const char *zDb,
     const char *zTable,
     const char *zColumn,
     sqlite3_int64 iRow,
     int flags,
     sqlite3_blob **ppBlob
   );
   
   /*
   ** CAPI3REF: Move a BLOB Handle to a New Row
   **
   ** ^This function is used to move an existing blob handle so that it points
   ** to a different row of the same database table. ^The new row is identified
   ** by the rowid value passed as the second argument. Only the row can be
   ** changed. ^The database, table and column on which the blob handle is open
   ** remain the same. Moving an existing blob handle to a new row can be
   ** faster than closing the existing handle and opening a new one.
   **
   ** ^(The new row must meet the same criteria as for [sqlite3_blob_open()] -
   ** it must exist and there must be either a blob or text value stored in
   ** the nominated column.)^ ^If the new row is not present in the table, or if
   ** it does not contain a blob or text value, or if another error occurs, an
   ** SQLite error code is returned and the blob handle is considered aborted.
   ** ^All subsequent calls to [sqlite3_blob_read()], [sqlite3_blob_write()] or
   ** [sqlite3_blob_reopen()] on an aborted blob handle immediately return
   ** SQLITE_ABORT. ^Calling [sqlite3_blob_bytes()] on an aborted blob handle
   ** always returns zero.
   **
   ** ^This function sets the database handle error code and message.
   */
   SQLITE_API SQLITE_EXPERIMENTAL int sqlite3_blob_reopen(sqlite3_blob *, sqlite3_int64);
   
   /*
   ** CAPI3REF: Close A BLOB Handle
   **
   ** ^Closes an open [BLOB handle].
   **
   ** ^Closing a BLOB shall cause the current transaction to commit
   ** if there are no other BLOBs, no pending prepared statements, and the
   ** database connection is in [autocommit mode].
   ** ^If any writes were made to the BLOB, they might be held in cache
   ** until the close operation if they will fit.
   **
   ** ^(Closing the BLOB often forces the changes
   ** out to disk and so if any I/O errors occur, they will likely occur
   ** at the time when the BLOB is closed.  Any errors that occur during
   ** closing are reported as a non-zero return value.)^
   **
   ** ^(The BLOB is closed unconditionally.  Even if this routine returns
   ** an error code, the BLOB is still closed.)^
   **
   ** ^Calling this routine with a null pointer (such as would be returned
   ** by a failed call to [sqlite3_blob_open()]) is a harmless no-op.
   */
   SQLITE_API int sqlite3_blob_close(sqlite3_blob *);
   
   /*
   ** CAPI3REF: Return The Size Of An Open BLOB
   **
   ** ^Returns the size in bytes of the BLOB accessible via the 
   ** successfully opened [BLOB handle] in its only argument.  ^The
   ** incremental blob I/O routines can only read or overwriting existing
   ** blob content; they cannot change the size of a blob.
   **
   ** This routine only works on a [BLOB handle] which has been created
   ** by a prior successful call to [sqlite3_blob_open()] and which has not
   ** been closed by [sqlite3_blob_close()].  Passing any other pointer in
   ** to this routine results in undefined and probably undesirable behavior.
   */
   SQLITE_API int sqlite3_blob_bytes(sqlite3_blob *);
   
   /*
   ** CAPI3REF: Read Data From A BLOB Incrementally
   **
   ** ^(This function is used to read data from an open [BLOB handle] into a
   ** caller-supplied buffer. N bytes of data are copied into buffer Z
   ** from the open BLOB, starting at offset iOffset.)^
   **
   ** ^If offset iOffset is less than N bytes from the end of the BLOB,
   ** [SQLITE_ERROR] is returned and no data is read.  ^If N or iOffset is
   ** less than zero, [SQLITE_ERROR] is returned and no data is read.
   ** ^The size of the blob (and hence the maximum value of N+iOffset)
   ** can be determined using the [sqlite3_blob_bytes()] interface.
   **
   ** ^An attempt to read from an expired [BLOB handle] fails with an
   ** error code of [SQLITE_ABORT].
   **
   ** ^(On success, sqlite3_blob_read() returns SQLITE_OK.
   ** Otherwise, an [error code] or an [extended error code] is returned.)^
   **
   ** This routine only works on a [BLOB handle] which has been created
   ** by a prior successful call to [sqlite3_blob_open()] and which has not
   ** been closed by [sqlite3_blob_close()].  Passing any other pointer in
   ** to this routine results in undefined and probably undesirable behavior.
   **
   ** See also: [sqlite3_blob_write()].
   */
   SQLITE_API int sqlite3_blob_read(sqlite3_blob *, void *Z, int N, int iOffset);
   
   /*
   ** CAPI3REF: Write Data Into A BLOB Incrementally
   **
   ** ^This function is used to write data into an open [BLOB handle] from a
   ** caller-supplied buffer. ^N bytes of data are copied from the buffer Z
   ** into the open BLOB, starting at offset iOffset.
   **
   ** ^If the [BLOB handle] passed as the first argument was not opened for
   ** writing (the flags parameter to [sqlite3_blob_open()] was zero),
   ** this function returns [SQLITE_READONLY].
   **
   ** ^This function may only modify the contents of the BLOB; it is
   ** not possible to increase the size of a BLOB using this API.
   ** ^If offset iOffset is less than N bytes from the end of the BLOB,
   ** [SQLITE_ERROR] is returned and no data is written.  ^If N is
   ** less than zero [SQLITE_ERROR] is returned and no data is written.
   ** The size of the BLOB (and hence the maximum value of N+iOffset)
   ** can be determined using the [sqlite3_blob_bytes()] interface.
   **
   ** ^An attempt to write to an expired [BLOB handle] fails with an
   ** error code of [SQLITE_ABORT].  ^Writes to the BLOB that occurred
   ** before the [BLOB handle] expired are not rolled back by the
   ** expiration of the handle, though of course those changes might
   ** have been overwritten by the statement that expired the BLOB handle
   ** or by other independent statements.
   **
   ** ^(On success, sqlite3_blob_write() returns SQLITE_OK.
   ** Otherwise, an  [error code] or an [extended error code] is returned.)^
   **
   ** This routine only works on a [BLOB handle] which has been created
   ** by a prior successful call to [sqlite3_blob_open()] and which has not
   ** been closed by [sqlite3_blob_close()].  Passing any other pointer in
   ** to this routine results in undefined and probably undesirable behavior.
   **
   ** See also: [sqlite3_blob_read()].
   */
   SQLITE_API int sqlite3_blob_write(sqlite3_blob *, const void *z, int n, int iOffset);
   
   /*
   ** CAPI3REF: Virtual File System Objects
   **
   ** A virtual filesystem (VFS) is an [sqlite3_vfs] object
   ** that SQLite uses to interact
   ** with the underlying operating system.  Most SQLite builds come with a
   ** single default VFS that is appropriate for the host computer.
   ** New VFSes can be registered and existing VFSes can be unregistered.
   ** The following interfaces are provided.
   **
   ** ^The sqlite3_vfs_find() interface returns a pointer to a VFS given its name.
   ** ^Names are case sensitive.
   ** ^Names are zero-terminated UTF-8 strings.
   ** ^If there is no match, a NULL pointer is returned.
   ** ^If zVfsName is NULL then the default VFS is returned.
   **
   ** ^New VFSes are registered with sqlite3_vfs_register().
   ** ^Each new VFS becomes the default VFS if the makeDflt flag is set.
   ** ^The same VFS can be registered multiple times without injury.
   ** ^To make an existing VFS into the default VFS, register it again
   ** with the makeDflt flag set.  If two different VFSes with the
   ** same name are registered, the behavior is undefined.  If a
   ** VFS is registered with a name that is NULL or an empty string,
   ** then the behavior is undefined.
   **
   ** ^Unregister a VFS with the sqlite3_vfs_unregister() interface.
   ** ^(If the default VFS is unregistered, another VFS is chosen as
   ** the default.  The choice for the new VFS is arbitrary.)^
   */
   SQLITE_API sqlite3_vfs *sqlite3_vfs_find(const char *zVfsName);
   SQLITE_API int sqlite3_vfs_register(sqlite3_vfs*, int makeDflt);
   SQLITE_API int sqlite3_vfs_unregister(sqlite3_vfs*);
   
   /*
   ** CAPI3REF: Mutexes
   **
   ** The SQLite core uses these routines for thread
   ** synchronization. Though they are intended for internal
   ** use by SQLite, code that links against SQLite is
   ** permitted to use any of these routines.
   **
   ** The SQLite source code contains multiple implementations
   ** of these mutex routines.  An appropriate implementation
   ** is selected automatically at compile-time.  ^(The following
   ** implementations are available in the SQLite core:
   **
   ** <ul>
   ** <li>   SQLITE_MUTEX_OS2
   ** <li>   SQLITE_MUTEX_PTHREADS
   ** <li>   SQLITE_MUTEX_W32
   ** <li>   SQLITE_MUTEX_NOOP
   ** </ul>)^
   **
   ** ^The SQLITE_MUTEX_NOOP implementation is a set of routines
   ** that does no real locking and is appropriate for use in
   ** a single-threaded application.  ^The SQLITE_MUTEX_OS2,
   ** SQLITE_MUTEX_PTHREADS, and SQLITE_MUTEX_W32 implementations
   ** are appropriate for use on OS/2, Unix, and Windows.
   **
   ** ^(If SQLite is compiled with the SQLITE_MUTEX_APPDEF preprocessor
   ** macro defined (with "-DSQLITE_MUTEX_APPDEF=1"), then no mutex
   ** implementation is included with the library. In this case the
   ** application must supply a custom mutex implementation using the
   ** [SQLITE_CONFIG_MUTEX] option of the sqlite3_config() function
   ** before calling sqlite3_initialize() or any other public sqlite3_
   ** function that calls sqlite3_initialize().)^
   **
   ** ^The sqlite3_mutex_alloc() routine allocates a new
   ** mutex and returns a pointer to it. ^If it returns NULL
   ** that means that a mutex could not be allocated.  ^SQLite
   ** will unwind its stack and return an error.  ^(The argument
   ** to sqlite3_mutex_alloc() is one of these integer constants:
   **
   ** <ul>
   ** <li>  SQLITE_MUTEX_FAST
   ** <li>  SQLITE_MUTEX_RECURSIVE
   ** <li>  SQLITE_MUTEX_STATIC_MASTER
   ** <li>  SQLITE_MUTEX_STATIC_MEM
   ** <li>  SQLITE_MUTEX_STATIC_MEM2
   ** <li>  SQLITE_MUTEX_STATIC_PRNG
   ** <li>  SQLITE_MUTEX_STATIC_LRU
   ** <li>  SQLITE_MUTEX_STATIC_LRU2
   ** </ul>)^
   **
   ** ^The first two constants (SQLITE_MUTEX_FAST and SQLITE_MUTEX_RECURSIVE)
   ** cause sqlite3_mutex_alloc() to create
   ** a new mutex.  ^The new mutex is recursive when SQLITE_MUTEX_RECURSIVE
   ** is used but not necessarily so when SQLITE_MUTEX_FAST is used.
   ** The mutex implementation does not need to make a distinction
   ** between SQLITE_MUTEX_RECURSIVE and SQLITE_MUTEX_FAST if it does
   ** not want to.  ^SQLite will only request a recursive mutex in
   ** cases where it really needs one.  ^If a faster non-recursive mutex
   ** implementation is available on the host platform, the mutex subsystem
   ** might return such a mutex in response to SQLITE_MUTEX_FAST.
   **
   ** ^The other allowed parameters to sqlite3_mutex_alloc() (anything other
   ** than SQLITE_MUTEX_FAST and SQLITE_MUTEX_RECURSIVE) each return
   ** a pointer to a static preexisting mutex.  ^Six static mutexes are
   ** used by the current version of SQLite.  Future versions of SQLite
   ** may add additional static mutexes.  Static mutexes are for internal
   ** use by SQLite only.  Applications that use SQLite mutexes should
   ** use only the dynamic mutexes returned by SQLITE_MUTEX_FAST or
   ** SQLITE_MUTEX_RECURSIVE.
   **
   ** ^Note that if one of the dynamic mutex parameters (SQLITE_MUTEX_FAST
   ** or SQLITE_MUTEX_RECURSIVE) is used then sqlite3_mutex_alloc()
   ** returns a different mutex on every call.  ^But for the static
   ** mutex types, the same mutex is returned on every call that has
   ** the same type number.
   **
   ** ^The sqlite3_mutex_free() routine deallocates a previously
   ** allocated dynamic mutex.  ^SQLite is careful to deallocate every
   ** dynamic mutex that it allocates.  The dynamic mutexes must not be in
   ** use when they are deallocated.  Attempting to deallocate a static
   ** mutex results in undefined behavior.  ^SQLite never deallocates
   ** a static mutex.
   **
   ** ^The sqlite3_mutex_enter() and sqlite3_mutex_try() routines attempt
   ** to enter a mutex.  ^If another thread is already within the mutex,
   ** sqlite3_mutex_enter() will block and sqlite3_mutex_try() will return
   ** SQLITE_BUSY.  ^The sqlite3_mutex_try() interface returns [SQLITE_OK]
   ** upon successful entry.  ^(Mutexes created using
   ** SQLITE_MUTEX_RECURSIVE can be entered multiple times by the same thread.
   ** In such cases the,
   ** mutex must be exited an equal number of times before another thread
   ** can enter.)^  ^(If the same thread tries to enter any other
   ** kind of mutex more than once, the behavior is undefined.
   ** SQLite will never exhibit
   ** such behavior in its own use of mutexes.)^
   **
   ** ^(Some systems (for example, Windows 95) do not support the operation
   ** implemented by sqlite3_mutex_try().  On those systems, sqlite3_mutex_try()
   ** will always return SQLITE_BUSY.  The SQLite core only ever uses
   ** sqlite3_mutex_try() as an optimization so this is acceptable behavior.)^
   **
   ** ^The sqlite3_mutex_leave() routine exits a mutex that was
   ** previously entered by the same thread.   ^(The behavior
   ** is undefined if the mutex is not currently entered by the
   ** calling thread or is not currently allocated.  SQLite will
   ** never do either.)^
   **
   ** ^If the argument to sqlite3_mutex_enter(), sqlite3_mutex_try(), or
   ** sqlite3_mutex_leave() is a NULL pointer, then all three routines
   ** behave as no-ops.
   **
   ** See also: [sqlite3_mutex_held()] and [sqlite3_mutex_notheld()].
   */
   SQLITE_API sqlite3_mutex *sqlite3_mutex_alloc(int);
   SQLITE_API void sqlite3_mutex_free(sqlite3_mutex*);
   SQLITE_API void sqlite3_mutex_enter(sqlite3_mutex*);
   SQLITE_API int sqlite3_mutex_try(sqlite3_mutex*);
   SQLITE_API void sqlite3_mutex_leave(sqlite3_mutex*);
   
   /*
   ** CAPI3REF: Mutex Methods Object
   **
   ** An instance of this structure defines the low-level routines
   ** used to allocate and use mutexes.
   **
   ** Usually, the default mutex implementations provided by SQLite are
   ** sufficient, however the user has the option of substituting a custom
   ** implementation for specialized deployments or systems for which SQLite
   ** does not provide a suitable implementation. In this case, the user
   ** creates and populates an instance of this structure to pass
   ** to sqlite3_config() along with the [SQLITE_CONFIG_MUTEX] option.
   ** Additionally, an instance of this structure can be used as an
   ** output variable when querying the system for the current mutex
   ** implementation, using the [SQLITE_CONFIG_GETMUTEX] option.
   **
   ** ^The xMutexInit method defined by this structure is invoked as
   ** part of system initialization by the sqlite3_initialize() function.
   ** ^The xMutexInit routine is called by SQLite exactly once for each
   ** effective call to [sqlite3_initialize()].
   **
   ** ^The xMutexEnd method defined by this structure is invoked as
   ** part of system shutdown by the sqlite3_shutdown() function. The
   ** implementation of this method is expected to release all outstanding
   ** resources obtained by the mutex methods implementation, especially
   ** those obtained by the xMutexInit method.  ^The xMutexEnd()
   ** interface is invoked exactly once for each call to [sqlite3_shutdown()].
   **
   ** ^(The remaining seven methods defined by this structure (xMutexAlloc,
   ** xMutexFree, xMutexEnter, xMutexTry, xMutexLeave, xMutexHeld and
   ** xMutexNotheld) implement the following interfaces (respectively):
   **
   ** <ul>
   **   <li>  [sqlite3_mutex_alloc()] </li>
   **   <li>  [sqlite3_mutex_free()] </li>
   **   <li>  [sqlite3_mutex_enter()] </li>
   **   <li>  [sqlite3_mutex_try()] </li>
   **   <li>  [sqlite3_mutex_leave()] </li>
   **   <li>  [sqlite3_mutex_held()] </li>
   **   <li>  [sqlite3_mutex_notheld()] </li>
   ** </ul>)^
   **
   ** The only difference is that the public sqlite3_XXX functions enumerated
   ** above silently ignore any invocations that pass a NULL pointer instead
   ** of a valid mutex handle. The implementations of the methods defined
   ** by this structure are not required to handle this case, the results
   ** of passing a NULL pointer instead of a valid mutex handle are undefined
   ** (i.e. it is acceptable to provide an implementation that segfaults if
   ** it is passed a NULL pointer).
   **
   ** The xMutexInit() method must be threadsafe.  ^It must be harmless to
   ** invoke xMutexInit() multiple times within the same process and without
   ** intervening calls to xMutexEnd().  Second and subsequent calls to
   ** xMutexInit() must be no-ops.
   **
   ** ^xMutexInit() must not use SQLite memory allocation ([sqlite3_malloc()]
   ** and its associates).  ^Similarly, xMutexAlloc() must not use SQLite memory
   ** allocation for a static mutex.  ^However xMutexAlloc() may use SQLite
   ** memory allocation for a fast or recursive mutex.
   **
   ** ^SQLite will invoke the xMutexEnd() method when [sqlite3_shutdown()] is
   ** called, but only if the prior call to xMutexInit returned SQLITE_OK.
   ** If xMutexInit fails in any way, it is expected to clean up after itself
   ** prior to returning.
   */
   typedef struct sqlite3_mutex_methods sqlite3_mutex_methods;
   struct sqlite3_mutex_methods {
     int (*xMutexInit)(void);
     int (*xMutexEnd)(void);
     sqlite3_mutex *(*xMutexAlloc)(int);
     void (*xMutexFree)(sqlite3_mutex *);
     void (*xMutexEnter)(sqlite3_mutex *);
     int (*xMutexTry)(sqlite3_mutex *);
     void (*xMutexLeave)(sqlite3_mutex *);
     int (*xMutexHeld)(sqlite3_mutex *);
     int (*xMutexNotheld)(sqlite3_mutex *);
   };
   
   /*
   ** CAPI3REF: Mutex Verification Routines
   **
   ** The sqlite3_mutex_held() and sqlite3_mutex_notheld() routines
   ** are intended for use inside assert() statements.  ^The SQLite core
   ** never uses these routines except inside an assert() and applications
   ** are advised to follow the lead of the core.  ^The SQLite core only
   ** provides implementations for these routines when it is compiled
   ** with the SQLITE_DEBUG flag.  ^External mutex implementations
   ** are only required to provide these routines if SQLITE_DEBUG is
   ** defined and if NDEBUG is not defined.
   **
   ** ^These routines should return true if the mutex in their argument
   ** is held or not held, respectively, by the calling thread.
   **
   ** ^The implementation is not required to provide versions of these
   ** routines that actually work. If the implementation does not provide working
   ** versions of these routines, it should at least provide stubs that always
   ** return true so that one does not get spurious assertion failures.
   **
   ** ^If the argument to sqlite3_mutex_held() is a NULL pointer then
   ** the routine should return 1.   This seems counter-intuitive since
   ** clearly the mutex cannot be held if it does not exist.  But
   ** the reason the mutex does not exist is because the build is not
   ** using mutexes.  And we do not want the assert() containing the
   ** call to sqlite3_mutex_held() to fail, so a non-zero return is
   ** the appropriate thing to do.  ^The sqlite3_mutex_notheld()
   ** interface should also return 1 when given a NULL pointer.
   */
   #ifndef NDEBUG
   SQLITE_API int sqlite3_mutex_held(sqlite3_mutex*);
   SQLITE_API int sqlite3_mutex_notheld(sqlite3_mutex*);
   #endif
   
   /*
   ** CAPI3REF: Mutex Types
   **
   ** The [sqlite3_mutex_alloc()] interface takes a single argument
   ** which is one of these integer constants.
   **
   ** The set of static mutexes may change from one SQLite release to the
   ** next.  Applications that override the built-in mutex logic must be
   ** prepared to accommodate additional static mutexes.
   */
   #define SQLITE_MUTEX_FAST             0
   #define SQLITE_MUTEX_RECURSIVE        1
   #define SQLITE_MUTEX_STATIC_MASTER    2
   #define SQLITE_MUTEX_STATIC_MEM       3  /* sqlite3_malloc() */
   #define SQLITE_MUTEX_STATIC_MEM2      4  /* NOT USED */
   #define SQLITE_MUTEX_STATIC_OPEN      4  /* sqlite3BtreeOpen() */
   #define SQLITE_MUTEX_STATIC_PRNG      5  /* sqlite3_random() */
   #define SQLITE_MUTEX_STATIC_LRU       6  /* lru page list */
   #define SQLITE_MUTEX_STATIC_LRU2      7  /* NOT USED */
   #define SQLITE_MUTEX_STATIC_PMEM      7  /* sqlite3PageMalloc() */
   
   /*
   ** CAPI3REF: Retrieve the mutex for a database connection
   **
   ** ^This interface returns a pointer the [sqlite3_mutex] object that 
   ** serializes access to the [database connection] given in the argument
   ** when the [threading mode] is Serialized.
   ** ^If the [threading mode] is Single-thread or Multi-thread then this
   ** routine returns a NULL pointer.
   */
   SQLITE_API sqlite3_mutex *sqlite3_db_mutex(sqlite3*);
   
   /*
   ** CAPI3REF: Low-Level Control Of Database Files
   **
   ** ^The [sqlite3_file_control()] interface makes a direct call to the
   ** xFileControl method for the [sqlite3_io_methods] object associated
   ** with a particular database identified by the second argument. ^The
   ** name of the database is "main" for the main database or "temp" for the
   ** TEMP database, or the name that appears after the AS keyword for
   ** databases that are added using the [ATTACH] SQL command.
   ** ^A NULL pointer can be used in place of "main" to refer to the
   ** main database file.
   ** ^The third and fourth parameters to this routine
   ** are passed directly through to the second and third parameters of
   ** the xFileControl method.  ^The return value of the xFileControl
   ** method becomes the return value of this routine.
   **
   ** ^The SQLITE_FCNTL_FILE_POINTER value for the op parameter causes
   ** a pointer to the underlying [sqlite3_file] object to be written into
   ** the space pointed to by the 4th parameter.  ^The SQLITE_FCNTL_FILE_POINTER
   ** case is a short-circuit path which does not actually invoke the
   ** underlying sqlite3_io_methods.xFileControl method.
   **
   ** ^If the second parameter (zDbName) does not match the name of any
   ** open database file, then SQLITE_ERROR is returned.  ^This error
   ** code is not remembered and will not be recalled by [sqlite3_errcode()]
   ** or [sqlite3_errmsg()].  The underlying xFileControl method might
   ** also return SQLITE_ERROR.  There is no way to distinguish between
   ** an incorrect zDbName and an SQLITE_ERROR return from the underlying
   ** xFileControl method.
   **
   ** See also: [SQLITE_FCNTL_LOCKSTATE]
   */
   SQLITE_API int sqlite3_file_control(sqlite3*, const char *zDbName, int op, void*);
   
   /*
   ** CAPI3REF: Testing Interface
   **
   ** ^The sqlite3_test_control() interface is used to read out internal
   ** state of SQLite and to inject faults into SQLite for testing
   ** purposes.  ^The first parameter is an operation code that determines
   ** the number, meaning, and operation of all subsequent parameters.
   **
   ** This interface is not for use by applications.  It exists solely
   ** for verifying the correct operation of the SQLite library.  Depending
   ** on how the SQLite library is compiled, this interface might not exist.
   **
   ** The details of the operation codes, their meanings, the parameters
   ** they take, and what they do are all subject to change without notice.
   ** Unlike most of the SQLite API, this function is not guaranteed to
   ** operate consistently from one release to the next.
   */
   SQLITE_API int sqlite3_test_control(int op, ...);
   
   /*
   ** CAPI3REF: Testing Interface Operation Codes
   **
   ** These constants are the valid operation code parameters used
   ** as the first argument to [sqlite3_test_control()].
   **
   ** These parameters and their meanings are subject to change
   ** without notice.  These values are for testing purposes only.
   ** Applications should not use any of these parameters or the
   ** [sqlite3_test_control()] interface.
   */
   #define SQLITE_TESTCTRL_FIRST                    5
   #define SQLITE_TESTCTRL_PRNG_SAVE                5
   #define SQLITE_TESTCTRL_PRNG_RESTORE             6
   #define SQLITE_TESTCTRL_PRNG_RESET               7
   #define SQLITE_TESTCTRL_BITVEC_TEST              8
   #define SQLITE_TESTCTRL_FAULT_INSTALL            9
   #define SQLITE_TESTCTRL_BENIGN_MALLOC_HOOKS     10
   #define SQLITE_TESTCTRL_PENDING_BYTE            11
   #define SQLITE_TESTCTRL_ASSERT                  12
   #define SQLITE_TESTCTRL_ALWAYS                  13
   #define SQLITE_TESTCTRL_RESERVE                 14
   #define SQLITE_TESTCTRL_OPTIMIZATIONS           15
   #define SQLITE_TESTCTRL_ISKEYWORD               16
   #define SQLITE_TESTCTRL_SCRATCHMALLOC           17
   #define SQLITE_TESTCTRL_LOCALTIME_FAULT         18
   #define SQLITE_TESTCTRL_EXPLAIN_STMT            19
   #define SQLITE_TESTCTRL_LAST                    19
   
   /*
   ** CAPI3REF: SQLite Runtime Status
   **
   ** ^This interface is used to retrieve runtime status information
   ** about the performance of SQLite, and optionally to reset various
   ** highwater marks.  ^The first argument is an integer code for
   ** the specific parameter to measure.  ^(Recognized integer codes
   ** are of the form [status parameters | SQLITE_STATUS_...].)^
   ** ^The current value of the parameter is returned into *pCurrent.
   ** ^The highest recorded value is returned in *pHighwater.  ^If the
   ** resetFlag is true, then the highest record value is reset after
   ** *pHighwater is written.  ^(Some parameters do not record the highest
   ** value.  For those parameters
   ** nothing is written into *pHighwater and the resetFlag is ignored.)^
   ** ^(Other parameters record only the highwater mark and not the current
   ** value.  For these latter parameters nothing is written into *pCurrent.)^
   **
   ** ^The sqlite3_status() routine returns SQLITE_OK on success and a
   ** non-zero [error code] on failure.
   **
   ** This routine is threadsafe but is not atomic.  This routine can be
   ** called while other threads are running the same or different SQLite
   ** interfaces.  However the values returned in *pCurrent and
   ** *pHighwater reflect the status of SQLite at different points in time
   ** and it is possible that another thread might change the parameter
   ** in between the times when *pCurrent and *pHighwater are written.
   **
   ** See also: [sqlite3_db_status()]
   */
   SQLITE_API int sqlite3_status(int op, int *pCurrent, int *pHighwater, int resetFlag);
   
   
   /*
   ** CAPI3REF: Status Parameters
   ** KEYWORDS: {status parameters}
   **
   ** These integer constants designate various run-time status parameters
   ** that can be returned by [sqlite3_status()].
   **
   ** <dl>
   ** [[SQLITE_STATUS_MEMORY_USED]] ^(<dt>SQLITE_STATUS_MEMORY_USED</dt>
   ** <dd>This parameter is the current amount of memory checked out
   ** using [sqlite3_malloc()], either directly or indirectly.  The
   ** figure includes calls made to [sqlite3_malloc()] by the application
   ** and internal memory usage by the SQLite library.  Scratch memory
   ** controlled by [SQLITE_CONFIG_SCRATCH] and auxiliary page-cache
   ** memory controlled by [SQLITE_CONFIG_PAGECACHE] is not included in
   ** this parameter.  The amount returned is the sum of the allocation
   ** sizes as reported by the xSize method in [sqlite3_mem_methods].</dd>)^
   **
   ** [[SQLITE_STATUS_MALLOC_SIZE]] ^(<dt>SQLITE_STATUS_MALLOC_SIZE</dt>
   ** <dd>This parameter records the largest memory allocation request
   ** handed to [sqlite3_malloc()] or [sqlite3_realloc()] (or their
   ** internal equivalents).  Only the value returned in the
   ** *pHighwater parameter to [sqlite3_status()] is of interest.  
   ** The value written into the *pCurrent parameter is undefined.</dd>)^
   **
   ** [[SQLITE_STATUS_MALLOC_COUNT]] ^(<dt>SQLITE_STATUS_MALLOC_COUNT</dt>
   ** <dd>This parameter records the number of separate memory allocations
   ** currently checked out.</dd>)^
   **
   ** [[SQLITE_STATUS_PAGECACHE_USED]] ^(<dt>SQLITE_STATUS_PAGECACHE_USED</dt>
   ** <dd>This parameter returns the number of pages used out of the
   ** [pagecache memory allocator] that was configured using 
   ** [SQLITE_CONFIG_PAGECACHE].  The
   ** value returned is in pages, not in bytes.</dd>)^
   **
   ** [[SQLITE_STATUS_PAGECACHE_OVERFLOW]] 
   ** ^(<dt>SQLITE_STATUS_PAGECACHE_OVERFLOW</dt>
   ** <dd>This parameter returns the number of bytes of page cache
   ** allocation which could not be satisfied by the [SQLITE_CONFIG_PAGECACHE]
   ** buffer and where forced to overflow to [sqlite3_malloc()].  The
   ** returned value includes allocations that overflowed because they
   ** where too large (they were larger than the "sz" parameter to
   ** [SQLITE_CONFIG_PAGECACHE]) and allocations that overflowed because
   ** no space was left in the page cache.</dd>)^
   **
   ** [[SQLITE_STATUS_PAGECACHE_SIZE]] ^(<dt>SQLITE_STATUS_PAGECACHE_SIZE</dt>
   ** <dd>This parameter records the largest memory allocation request
   ** handed to [pagecache memory allocator].  Only the value returned in the
   ** *pHighwater parameter to [sqlite3_status()] is of interest.  
   ** The value written into the *pCurrent parameter is undefined.</dd>)^
   **
   ** [[SQLITE_STATUS_SCRATCH_USED]] ^(<dt>SQLITE_STATUS_SCRATCH_USED</dt>
   ** <dd>This parameter returns the number of allocations used out of the
   ** [scratch memory allocator] configured using
   ** [SQLITE_CONFIG_SCRATCH].  The value returned is in allocations, not
   ** in bytes.  Since a single thread may only have one scratch allocation
   ** outstanding at time, this parameter also reports the number of threads
   ** using scratch memory at the same time.</dd>)^
   **
   ** [[SQLITE_STATUS_SCRATCH_OVERFLOW]] ^(<dt>SQLITE_STATUS_SCRATCH_OVERFLOW</dt>
   ** <dd>This parameter returns the number of bytes of scratch memory
   ** allocation which could not be satisfied by the [SQLITE_CONFIG_SCRATCH]
   ** buffer and where forced to overflow to [sqlite3_malloc()].  The values
   ** returned include overflows because the requested allocation was too
   ** larger (that is, because the requested allocation was larger than the
   ** "sz" parameter to [SQLITE_CONFIG_SCRATCH]) and because no scratch buffer
   ** slots were available.
   ** </dd>)^
   **
   ** [[SQLITE_STATUS_SCRATCH_SIZE]] ^(<dt>SQLITE_STATUS_SCRATCH_SIZE</dt>
   ** <dd>This parameter records the largest memory allocation request
   ** handed to [scratch memory allocator].  Only the value returned in the
   ** *pHighwater parameter to [sqlite3_status()] is of interest.  
   ** The value written into the *pCurrent parameter is undefined.</dd>)^
   **
   ** [[SQLITE_STATUS_PARSER_STACK]] ^(<dt>SQLITE_STATUS_PARSER_STACK</dt>
   ** <dd>This parameter records the deepest parser stack.  It is only
   ** meaningful if SQLite is compiled with [YYTRACKMAXSTACKDEPTH].</dd>)^
   ** </dl>
   **
   ** New status parameters may be added from time to time.
   */
   #define SQLITE_STATUS_MEMORY_USED          0
   #define SQLITE_STATUS_PAGECACHE_USED       1
   #define SQLITE_STATUS_PAGECACHE_OVERFLOW   2
   #define SQLITE_STATUS_SCRATCH_USED         3
   #define SQLITE_STATUS_SCRATCH_OVERFLOW     4
   #define SQLITE_STATUS_MALLOC_SIZE          5
   #define SQLITE_STATUS_PARSER_STACK         6
   #define SQLITE_STATUS_PAGECACHE_SIZE       7
   #define SQLITE_STATUS_SCRATCH_SIZE         8
   #define SQLITE_STATUS_MALLOC_COUNT         9
   
   /*
   ** CAPI3REF: Database Connection Status
   **
   ** ^This interface is used to retrieve runtime status information 
   ** about a single [database connection].  ^The first argument is the
   ** database connection object to be interrogated.  ^The second argument
   ** is an integer constant, taken from the set of
   ** [SQLITE_DBSTATUS options], that
   ** determines the parameter to interrogate.  The set of 
   ** [SQLITE_DBSTATUS options] is likely
   ** to grow in future releases of SQLite.
   **
   ** ^The current value of the requested parameter is written into *pCur
   ** and the highest instantaneous value is written into *pHiwtr.  ^If
   ** the resetFlg is true, then the highest instantaneous value is
   ** reset back down to the current value.
   **
   ** ^The sqlite3_db_status() routine returns SQLITE_OK on success and a
   ** non-zero [error code] on failure.
   **
   ** See also: [sqlite3_status()] and [sqlite3_stmt_status()].
   */
   SQLITE_API int sqlite3_db_status(sqlite3*, int op, int *pCur, int *pHiwtr, int resetFlg);
   
   /*
   ** CAPI3REF: Status Parameters for database connections
   ** KEYWORDS: {SQLITE_DBSTATUS options}
   **
   ** These constants are the available integer "verbs" that can be passed as
   ** the second argument to the [sqlite3_db_status()] interface.
   **
   ** New verbs may be added in future releases of SQLite. Existing verbs
   ** might be discontinued. Applications should check the return code from
   ** [sqlite3_db_status()] to make sure that the call worked.
   ** The [sqlite3_db_status()] interface will return a non-zero error code
   ** if a discontinued or unsupported verb is invoked.
   **
   ** <dl>
   ** [[SQLITE_DBSTATUS_LOOKASIDE_USED]] ^(<dt>SQLITE_DBSTATUS_LOOKASIDE_USED</dt>
   ** <dd>This parameter returns the number of lookaside memory slots currently
   ** checked out.</dd>)^
   **
   ** [[SQLITE_DBSTATUS_LOOKASIDE_HIT]] ^(<dt>SQLITE_DBSTATUS_LOOKASIDE_HIT</dt>
   ** <dd>This parameter returns the number malloc attempts that were 
   ** satisfied using lookaside memory. Only the high-water value is meaningful;
   ** the current value is always zero.)^
   **
   ** [[SQLITE_DBSTATUS_LOOKASIDE_MISS_SIZE]]
   ** ^(<dt>SQLITE_DBSTATUS_LOOKASIDE_MISS_SIZE</dt>
   ** <dd>This parameter returns the number malloc attempts that might have
   ** been satisfied using lookaside memory but failed due to the amount of
   ** memory requested being larger than the lookaside slot size.
   ** Only the high-water value is meaningful;
   ** the current value is always zero.)^
   **
   ** [[SQLITE_DBSTATUS_LOOKASIDE_MISS_FULL]]
   ** ^(<dt>SQLITE_DBSTATUS_LOOKASIDE_MISS_FULL</dt>
   ** <dd>This parameter returns the number malloc attempts that might have
   ** been satisfied using lookaside memory but failed due to all lookaside
   ** memory already being in use.
   ** Only the high-water value is meaningful;
   ** the current value is always zero.)^
   **
   ** [[SQLITE_DBSTATUS_CACHE_USED]] ^(<dt>SQLITE_DBSTATUS_CACHE_USED</dt>
   ** <dd>This parameter returns the approximate number of of bytes of heap
   ** memory used by all pager caches associated with the database connection.)^
   ** ^The highwater mark associated with SQLITE_DBSTATUS_CACHE_USED is always 0.
   **
   ** [[SQLITE_DBSTATUS_SCHEMA_USED]] ^(<dt>SQLITE_DBSTATUS_SCHEMA_USED</dt>
   ** <dd>This parameter returns the approximate number of of bytes of heap
   ** memory used to store the schema for all databases associated
   ** with the connection - main, temp, and any [ATTACH]-ed databases.)^ 
   ** ^The full amount of memory used by the schemas is reported, even if the
   ** schema memory is shared with other database connections due to
   ** [shared cache mode] being enabled.
   ** ^The highwater mark associated with SQLITE_DBSTATUS_SCHEMA_USED is always 0.
   **
   ** [[SQLITE_DBSTATUS_STMT_USED]] ^(<dt>SQLITE_DBSTATUS_STMT_USED</dt>
   ** <dd>This parameter returns the approximate number of of bytes of heap
   ** and lookaside memory used by all prepared statements associated with
   ** the database connection.)^
   ** ^The highwater mark associated with SQLITE_DBSTATUS_STMT_USED is always 0.
   ** </dd>
   **
   ** [[SQLITE_DBSTATUS_CACHE_HIT]] ^(<dt>SQLITE_DBSTATUS_CACHE_HIT</dt>
   ** <dd>This parameter returns the number of pager cache hits that have
   ** occurred.)^ ^The highwater mark associated with SQLITE_DBSTATUS_CACHE_HIT 
   ** is always 0.
   ** </dd>
   **
   ** [[SQLITE_DBSTATUS_CACHE_MISS]] ^(<dt>SQLITE_DBSTATUS_CACHE_MISS</dt>
   ** <dd>This parameter returns the number of pager cache misses that have
   ** occurred.)^ ^The highwater mark associated with SQLITE_DBSTATUS_CACHE_MISS 
   ** is always 0.
   ** </dd>
   ** </dl>
   */
   #define SQLITE_DBSTATUS_LOOKASIDE_USED       0
   #define SQLITE_DBSTATUS_CACHE_USED           1
   #define SQLITE_DBSTATUS_SCHEMA_USED          2
   #define SQLITE_DBSTATUS_STMT_USED            3
   #define SQLITE_DBSTATUS_LOOKASIDE_HIT        4
   #define SQLITE_DBSTATUS_LOOKASIDE_MISS_SIZE  5
   #define SQLITE_DBSTATUS_LOOKASIDE_MISS_FULL  6
   #define SQLITE_DBSTATUS_CACHE_HIT            7
   #define SQLITE_DBSTATUS_CACHE_MISS           8
   #define SQLITE_DBSTATUS_MAX                  8   /* Largest defined DBSTATUS */
   
   
   /*
   ** CAPI3REF: Prepared Statement Status
   **
   ** ^(Each prepared statement maintains various
   ** [SQLITE_STMTSTATUS counters] that measure the number
   ** of times it has performed specific operations.)^  These counters can
   ** be used to monitor the performance characteristics of the prepared
   ** statements.  For example, if the number of table steps greatly exceeds
   ** the number of table searches or result rows, that would tend to indicate
   ** that the prepared statement is using a full table scan rather than
   ** an index.  
   **
   ** ^(This interface is used to retrieve and reset counter values from
   ** a [prepared statement].  The first argument is the prepared statement
   ** object to be interrogated.  The second argument
   ** is an integer code for a specific [SQLITE_STMTSTATUS counter]
   ** to be interrogated.)^
   ** ^The current value of the requested counter is returned.
   ** ^If the resetFlg is true, then the counter is reset to zero after this
   ** interface call returns.
   **
   ** See also: [sqlite3_status()] and [sqlite3_db_status()].
   */
   SQLITE_API int sqlite3_stmt_status(sqlite3_stmt*, int op,int resetFlg);
   
   /*
   ** CAPI3REF: Status Parameters for prepared statements
   ** KEYWORDS: {SQLITE_STMTSTATUS counter} {SQLITE_STMTSTATUS counters}
   **
   ** These preprocessor macros define integer codes that name counter
   ** values associated with the [sqlite3_stmt_status()] interface.
   ** The meanings of the various counters are as follows:
   **
   ** <dl>
   ** [[SQLITE_STMTSTATUS_FULLSCAN_STEP]] <dt>SQLITE_STMTSTATUS_FULLSCAN_STEP</dt>
   ** <dd>^This is the number of times that SQLite has stepped forward in
   ** a table as part of a full table scan.  Large numbers for this counter
   ** may indicate opportunities for performance improvement through 
   ** careful use of indices.</dd>
   **
   ** [[SQLITE_STMTSTATUS_SORT]] <dt>SQLITE_STMTSTATUS_SORT</dt>
   ** <dd>^This is the number of sort operations that have occurred.
   ** A non-zero value in this counter may indicate an opportunity to
   ** improvement performance through careful use of indices.</dd>
   **
   ** [[SQLITE_STMTSTATUS_AUTOINDEX]] <dt>SQLITE_STMTSTATUS_AUTOINDEX</dt>
   ** <dd>^This is the number of rows inserted into transient indices that
   ** were created automatically in order to help joins run faster.
   ** A non-zero value in this counter may indicate an opportunity to
   ** improvement performance by adding permanent indices that do not
   ** need to be reinitialized each time the statement is run.</dd>
   ** </dl>
   */
   #define SQLITE_STMTSTATUS_FULLSCAN_STEP     1
   #define SQLITE_STMTSTATUS_SORT              2
   #define SQLITE_STMTSTATUS_AUTOINDEX         3
   
   /*
   ** CAPI3REF: Custom Page Cache Object
   **
   ** The sqlite3_pcache type is opaque.  It is implemented by
   ** the pluggable module.  The SQLite core has no knowledge of
   ** its size or internal structure and never deals with the
   ** sqlite3_pcache object except by holding and passing pointers
   ** to the object.
   **
   ** See [sqlite3_pcache_methods2] for additional information.
   */
   typedef struct sqlite3_pcache sqlite3_pcache;
   
   /*
   ** CAPI3REF: Custom Page Cache Object
   **
   ** The sqlite3_pcache_page object represents a single page in the
   ** page cache.  The page cache will allocate instances of this
   ** object.  Various methods of the page cache use pointers to instances
   ** of this object as parameters or as their return value.
   **
   ** See [sqlite3_pcache_methods2] for additional information.
   */
   typedef struct sqlite3_pcache_page sqlite3_pcache_page;
   struct sqlite3_pcache_page {
     void *pBuf;        /* The content of the page */
     void *pExtra;      /* Extra information associated with the page */
   };
   
   /*
   ** CAPI3REF: Application Defined Page Cache.
   ** KEYWORDS: {page cache}
   **
   ** ^(The [sqlite3_config]([SQLITE_CONFIG_PCACHE2], ...) interface can
   ** register an alternative page cache implementation by passing in an 
   ** instance of the sqlite3_pcache_methods2 structure.)^
   ** In many applications, most of the heap memory allocated by 
   ** SQLite is used for the page cache.
   ** By implementing a 
   ** custom page cache using this API, an application can better control
   ** the amount of memory consumed by SQLite, the way in which 
   ** that memory is allocated and released, and the policies used to 
   ** determine exactly which parts of a database file are cached and for 
   ** how long.
   **
   ** The alternative page cache mechanism is an
   ** extreme measure that is only needed by the most demanding applications.
   ** The built-in page cache is recommended for most uses.
   **
   ** ^(The contents of the sqlite3_pcache_methods2 structure are copied to an
   ** internal buffer by SQLite within the call to [sqlite3_config].  Hence
   ** the application may discard the parameter after the call to
   ** [sqlite3_config()] returns.)^
   **
   ** [[the xInit() page cache method]]
   ** ^(The xInit() method is called once for each effective 
   ** call to [sqlite3_initialize()])^
   ** (usually only once during the lifetime of the process). ^(The xInit()
   ** method is passed a copy of the sqlite3_pcache_methods2.pArg value.)^
   ** The intent of the xInit() method is to set up global data structures 
   ** required by the custom page cache implementation. 
   ** ^(If the xInit() method is NULL, then the 
   ** built-in default page cache is used instead of the application defined
   ** page cache.)^
   **
   ** [[the xShutdown() page cache method]]
   ** ^The xShutdown() method is called by [sqlite3_shutdown()].
   ** It can be used to clean up 
   ** any outstanding resources before process shutdown, if required.
   ** ^The xShutdown() method may be NULL.
   **
   ** ^SQLite automatically serializes calls to the xInit method,
   ** so the xInit method need not be threadsafe.  ^The
   ** xShutdown method is only called from [sqlite3_shutdown()] so it does
   ** not need to be threadsafe either.  All other methods must be threadsafe
   ** in multithreaded applications.
   **
   ** ^SQLite will never invoke xInit() more than once without an intervening
   ** call to xShutdown().
   **
   ** [[the xCreate() page cache methods]]
   ** ^SQLite invokes the xCreate() method to construct a new cache instance.
   ** SQLite will typically create one cache instance for each open database file,
   ** though this is not guaranteed. ^The
   ** first parameter, szPage, is the size in bytes of the pages that must
   ** be allocated by the cache.  ^szPage will always a power of two.  ^The
   ** second parameter szExtra is a number of bytes of extra storage 
   ** associated with each page cache entry.  ^The szExtra parameter will
   ** a number less than 250.  SQLite will use the
   ** extra szExtra bytes on each page to store metadata about the underlying
   ** database page on disk.  The value passed into szExtra depends
   ** on the SQLite version, the target platform, and how SQLite was compiled.
   ** ^The third argument to xCreate(), bPurgeable, is true if the cache being
   ** created will be used to cache database pages of a file stored on disk, or
   ** false if it is used for an in-memory database. The cache implementation
   ** does not have to do anything special based with the value of bPurgeable;
   ** it is purely advisory.  ^On a cache where bPurgeable is false, SQLite will
   ** never invoke xUnpin() except to deliberately delete a page.
   ** ^In other words, calls to xUnpin() on a cache with bPurgeable set to
   ** false will always have the "discard" flag set to true.  
   ** ^Hence, a cache created with bPurgeable false will
   ** never contain any unpinned pages.
   **
   ** [[the xCachesize() page cache method]]
   ** ^(The xCachesize() method may be called at any time by SQLite to set the
   ** suggested maximum cache-size (number of pages stored by) the cache
   ** instance passed as the first argument. This is the value configured using
   ** the SQLite "[PRAGMA cache_size]" command.)^  As with the bPurgeable
   ** parameter, the implementation is not required to do anything with this
   ** value; it is advisory only.
   **
   ** [[the xPagecount() page cache methods]]
   ** The xPagecount() method must return the number of pages currently
   ** stored in the cache, both pinned and unpinned.
   ** 
   ** [[the xFetch() page cache methods]]
   ** The xFetch() method locates a page in the cache and returns a pointer to 
   ** an sqlite3_pcache_page object associated with that page, or a NULL pointer.
   ** The pBuf element of the returned sqlite3_pcache_page object will be a
   ** pointer to a buffer of szPage bytes used to store the content of a 
   ** single database page.  The pExtra element of sqlite3_pcache_page will be
   ** a pointer to the szExtra bytes of extra storage that SQLite has requested
   ** for each entry in the page cache.
   **
   ** The page to be fetched is determined by the key. ^The minimum key value
   ** is 1.  After it has been retrieved using xFetch, the page is considered
   ** to be "pinned".
   **
   ** If the requested page is already in the page cache, then the page cache
   ** implementation must return a pointer to the page buffer with its content
   ** intact.  If the requested page is not already in the cache, then the
   ** cache implementation should use the value of the createFlag
   ** parameter to help it determined what action to take:
   **
   ** <table border=1 width=85% align=center>
   ** <tr><th> createFlag <th> Behaviour when page is not already in cache
   ** <tr><td> 0 <td> Do not allocate a new page.  Return NULL.
   ** <tr><td> 1 <td> Allocate a new page if it easy and convenient to do so.
   **                 Otherwise return NULL.
   ** <tr><td> 2 <td> Make every effort to allocate a new page.  Only return
   **                 NULL if allocating a new page is effectively impossible.
   ** </table>
   **
   ** ^(SQLite will normally invoke xFetch() with a createFlag of 0 or 1.  SQLite
   ** will only use a createFlag of 2 after a prior call with a createFlag of 1
   ** failed.)^  In between the to xFetch() calls, SQLite may
   ** attempt to unpin one or more cache pages by spilling the content of
   ** pinned pages to disk and synching the operating system disk cache.
   **
   ** [[the xUnpin() page cache method]]
   ** ^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 must be evicted from the cache.
   ** ^If the discard parameter is
   ** zero, then the page may be discarded or retained at the discretion of
   ** page cache implementation. ^The page cache implementation
   ** may choose to evict unpinned pages at any time.
   **
   ** The cache must not perform any reference counting. A single 
   ** call to xUnpin() unpins the page regardless of the number of prior calls 
   ** to xFetch().
   **
   ** [[the xRekey() page cache methods]]
   ** The xRekey() method is used to change the key value associated with the
   ** page passed as the second argument. If the cache
   ** previously contains an entry associated with newKey, it must be
   ** discarded. ^Any prior cache entry associated with newKey is guaranteed not
   ** to be pinned.
   **
   ** When SQLite calls the xTruncate() method, the cache must discard all
   ** existing cache entries with page numbers (keys) greater than or equal
   ** to the value of the iLimit parameter passed to xTruncate(). If any
   ** of these pages are pinned, they are implicitly unpinned, meaning that
   ** they can be safely discarded.
   **
   ** [[the xDestroy() page cache method]]
   ** ^The xDestroy() method is used to delete a cache allocated by xCreate().
   ** All resources associated with the specified cache should be freed. ^After
   ** calling the xDestroy() method, SQLite considers the [sqlite3_pcache*]
   ** handle invalid, and will not use it with any other sqlite3_pcache_methods2
   ** functions.
   **
   ** [[the xShrink() page cache method]]
   ** ^SQLite invokes the xShrink() method when it wants the page cache to
   ** free up as much of heap memory as possible.  The page cache implementation
   ** is not obligated to free any memory, but well-behaved implementations should
   ** do their best.
   */
   typedef struct sqlite3_pcache_methods2 sqlite3_pcache_methods2;
   struct sqlite3_pcache_methods2 {
     int iVersion;
     void *pArg;
     int (*xInit)(void*);
     void (*xShutdown)(void*);
     sqlite3_pcache *(*xCreate)(int szPage, int szExtra, int bPurgeable);
     void (*xCachesize)(sqlite3_pcache*, int nCachesize);
     int (*xPagecount)(sqlite3_pcache*);
     sqlite3_pcache_page *(*xFetch)(sqlite3_pcache*, unsigned key, int createFlag);
     void (*xUnpin)(sqlite3_pcache*, sqlite3_pcache_page*, int discard);
     void (*xRekey)(sqlite3_pcache*, sqlite3_pcache_page*, 
         unsigned oldKey, unsigned newKey);
     void (*xTruncate)(sqlite3_pcache*, unsigned iLimit);
     void (*xDestroy)(sqlite3_pcache*);
     void (*xShrink)(sqlite3_pcache*);
   };
   
   /*
   ** This is the obsolete pcache_methods object that has now been replaced
   ** by sqlite3_pcache_methods2.  This object is not used by SQLite.  It is
   ** retained in the header file for backwards compatibility only.
   */
   typedef struct sqlite3_pcache_methods sqlite3_pcache_methods;
   struct sqlite3_pcache_methods {
     void *pArg;
     int (*xInit)(void*);
     void (*xShutdown)(void*);
     sqlite3_pcache *(*xCreate)(int szPage, int bPurgeable);
     void (*xCachesize)(sqlite3_pcache*, int nCachesize);
     int (*xPagecount)(sqlite3_pcache*);
     void *(*xFetch)(sqlite3_pcache*, unsigned key, int createFlag);
     void (*xUnpin)(sqlite3_pcache*, void*, int discard);
     void (*xRekey)(sqlite3_pcache*, void*, unsigned oldKey, unsigned newKey);
     void (*xTruncate)(sqlite3_pcache*, unsigned iLimit);
     void (*xDestroy)(sqlite3_pcache*);
   };
   
   
   /*
   ** CAPI3REF: Online Backup Object
   **
   ** The sqlite3_backup object records state information about an ongoing
   ** online backup operation.  ^The sqlite3_backup object is created by
   ** a call to [sqlite3_backup_init()] and is destroyed by a call to
   ** [sqlite3_backup_finish()].
   **
   ** See Also: [Using the SQLite Online Backup API]
   */
   typedef struct sqlite3_backup sqlite3_backup;
   
   /*
   ** CAPI3REF: Online Backup API.
   **
   ** The backup API copies the content of one database into another.
   ** It is useful either for creating backups of databases or
   ** for copying in-memory databases to or from persistent files. 
   **
   ** See Also: [Using the SQLite Online Backup API]
   **
   ** ^SQLite holds a write transaction open on the destination database file
   ** for the duration of the backup operation.
   ** ^The source database is read-locked only while it is being read;
   ** it is not locked continuously for the entire backup operation.
   ** ^Thus, the backup may be performed on a live source database without
   ** preventing other database connections from
   ** reading or writing to the source database while the backup is underway.
   ** 
   ** ^(To perform a backup operation: 
   **   <ol>
   **     <li><b>sqlite3_backup_init()</b> is called once to initialize the
   **         backup, 
   **     <li><b>sqlite3_backup_step()</b> is called one or more times to transfer 
   **         the data between the two databases, and finally
   **     <li><b>sqlite3_backup_finish()</b> is called to release all resources 
   **         associated with the backup operation. 
   **   </ol>)^
   ** There should be exactly one call to sqlite3_backup_finish() for each
   ** successful call to sqlite3_backup_init().
   **
   ** [[sqlite3_backup_init()]] <b>sqlite3_backup_init()</b>
   **
   ** ^The D and N arguments to sqlite3_backup_init(D,N,S,M) are the 
   ** [database connection] associated with the destination database 
   ** and the database name, respectively.
   ** ^The database name is "main" for the main database, "temp" for the
   ** temporary database, or the name specified after the AS keyword in
   ** an [ATTACH] statement for an attached database.
   ** ^The S and M arguments passed to 
   ** sqlite3_backup_init(D,N,S,M) identify the [database connection]
   ** and database name of the source database, respectively.
   ** ^The source and destination [database connections] (parameters S and D)
   ** must be different or else sqlite3_backup_init(D,N,S,M) will fail with
   ** an error.
   **
   ** ^If an error occurs within sqlite3_backup_init(D,N,S,M), then NULL is
   ** returned and an error code and error message are stored in the
   ** destination [database connection] D.
   ** ^The error code and message for the failed call to sqlite3_backup_init()
   ** can be retrieved using the [sqlite3_errcode()], [sqlite3_errmsg()], and/or
   ** [sqlite3_errmsg16()] functions.
   ** ^A successful call to sqlite3_backup_init() returns a pointer to an
   ** [sqlite3_backup] object.
   ** ^The [sqlite3_backup] object may be used with the sqlite3_backup_step() and
   ** sqlite3_backup_finish() functions to perform the specified backup 
   ** operation.
   **
   ** [[sqlite3_backup_step()]] <b>sqlite3_backup_step()</b>
   **
   ** ^Function sqlite3_backup_step(B,N) will copy up to N pages between 
   ** the source and destination databases specified by [sqlite3_backup] object B.
   ** ^If N is negative, all remaining source pages are copied. 
   ** ^If sqlite3_backup_step(B,N) successfully copies N pages and there
   ** are still more pages to be copied, then the function returns [SQLITE_OK].
   ** ^If sqlite3_backup_step(B,N) successfully finishes copying all pages
   ** from source to destination, then it returns [SQLITE_DONE].
   ** ^If an error occurs while running sqlite3_backup_step(B,N),
   ** then an [error code] is returned. ^As well as [SQLITE_OK] and
   ** [SQLITE_DONE], a call to sqlite3_backup_step() may return [SQLITE_READONLY],
   ** [SQLITE_NOMEM], [SQLITE_BUSY], [SQLITE_LOCKED], or an
   ** [SQLITE_IOERR_ACCESS | SQLITE_IOERR_XXX] extended error code.
   **
   ** ^(The sqlite3_backup_step() might return [SQLITE_READONLY] if
   ** <ol>
   ** <li> the destination database was opened read-only, or
   ** <li> the destination database is using write-ahead-log journaling
   ** and the destination and source page sizes differ, or
   ** <li> the destination database is an in-memory database and the
   ** destination and source page sizes differ.
   ** </ol>)^
   **
   ** ^If sqlite3_backup_step() cannot obtain a required file-system lock, then
   ** the [sqlite3_busy_handler | busy-handler function]
   ** is invoked (if one is specified). ^If the 
   ** busy-handler returns non-zero before the lock is available, then 
   ** [SQLITE_BUSY] is returned to the caller. ^In this case the call to
   ** sqlite3_backup_step() can be retried later. ^If the source
   ** [database connection]
   ** is being used to write to the source database when sqlite3_backup_step()
   ** is called, then [SQLITE_LOCKED] is returned immediately. ^Again, in this
   ** case the call to sqlite3_backup_step() can be retried later on. ^(If
   ** [SQLITE_IOERR_ACCESS | SQLITE_IOERR_XXX], [SQLITE_NOMEM], or
   ** [SQLITE_READONLY] is returned, then 
   ** there is no point in retrying the call to sqlite3_backup_step(). These 
   ** errors are considered fatal.)^  The application must accept 
   ** that the backup operation has failed and pass the backup operation handle 
   ** to the sqlite3_backup_finish() to release associated resources.
   **
   ** ^The first call to sqlite3_backup_step() obtains an exclusive lock
   ** on the destination file. ^The exclusive lock is not released until either 
   ** sqlite3_backup_finish() is called or the backup operation is complete 
   ** and sqlite3_backup_step() returns [SQLITE_DONE].  ^Every call to
   ** sqlite3_backup_step() obtains a [shared lock] on the source database that
   ** lasts for the duration of the sqlite3_backup_step() call.
   ** ^Because the source database is not locked between calls to
   ** sqlite3_backup_step(), the source database may be modified mid-way
   ** through the backup process.  ^If the source database is modified by an
   ** external process or via a database connection other than the one being
   ** used by the backup operation, then the backup will be automatically
   ** restarted by the next call to sqlite3_backup_step(). ^If the source 
   ** database is modified by the using the same database connection as is used
   ** by the backup operation, then the backup database is automatically
   ** updated at the same time.
   **
   ** [[sqlite3_backup_finish()]] <b>sqlite3_backup_finish()</b>
   **
   ** When sqlite3_backup_step() has returned [SQLITE_DONE], or when the 
   ** application wishes to abandon the backup operation, the application
   ** should destroy the [sqlite3_backup] by passing it to sqlite3_backup_finish().
   ** ^The sqlite3_backup_finish() interfaces releases all
   ** resources associated with the [sqlite3_backup] object. 
   ** ^If sqlite3_backup_step() has not yet returned [SQLITE_DONE], then any
   ** active write-transaction on the destination database is rolled back.
   ** The [sqlite3_backup] object is invalid
   ** and may not be used following a call to sqlite3_backup_finish().
   **
   ** ^The value returned by sqlite3_backup_finish is [SQLITE_OK] if no
   ** sqlite3_backup_step() errors occurred, regardless or whether or not
   ** sqlite3_backup_step() completed.
   ** ^If an out-of-memory condition or IO error occurred during any prior
   ** sqlite3_backup_step() call on the same [sqlite3_backup] object, then
   ** sqlite3_backup_finish() returns the corresponding [error code].
   **
   ** ^A return of [SQLITE_BUSY] or [SQLITE_LOCKED] from sqlite3_backup_step()
   ** is not a permanent error and does not affect the return value of
   ** sqlite3_backup_finish().
   **
   ** [[sqlite3_backup__remaining()]] [[sqlite3_backup_pagecount()]]
   ** <b>sqlite3_backup_remaining() and sqlite3_backup_pagecount()</b>
   **
   ** ^Each call to sqlite3_backup_step() sets two values inside
   ** the [sqlite3_backup] object: the number of pages still to be backed
   ** up and the total number of pages in the source database file.
   ** The sqlite3_backup_remaining() and sqlite3_backup_pagecount() interfaces
   ** retrieve these two values, respectively.
   **
   ** ^The values returned by these functions are only updated by
   ** sqlite3_backup_step(). ^If the source database is modified during a backup
   ** operation, then the values are not updated to account for any extra
   ** pages that need to be updated or the size of the source database file
   ** changing.
   **
   ** <b>Concurrent Usage of Database Handles</b>
   **
   ** ^The source [database connection] may be used by the application for other
   ** purposes while a backup operation is underway or being initialized.
   ** ^If SQLite is compiled and configured to support threadsafe database
   ** connections, then the source database connection may be used concurrently
   ** from within other threads.
   **
   ** However, the application must guarantee that the destination 
   ** [database connection] is not passed to any other API (by any thread) after 
   ** sqlite3_backup_init() is called and before the corresponding call to
   ** sqlite3_backup_finish().  SQLite does not currently check to see
   ** if the application incorrectly accesses the destination [database connection]
   ** and so no error code is reported, but the operations may malfunction
   ** nevertheless.  Use of the destination database connection while a
   ** backup is in progress might also also cause a mutex deadlock.
   **
   ** If running in [shared cache mode], the application must
   ** guarantee that the shared cache used by the destination database
   ** is not accessed while the backup is running. In practice this means
   ** that the application must guarantee that the disk file being 
   ** backed up to is not accessed by any connection within the process,
   ** not just the specific connection that was passed to sqlite3_backup_init().
   **
   ** The [sqlite3_backup] object itself is partially threadsafe. Multiple 
   ** threads may safely make multiple concurrent calls to sqlite3_backup_step().
   ** However, the sqlite3_backup_remaining() and sqlite3_backup_pagecount()
   ** APIs are not strictly speaking threadsafe. If they are invoked at the
   ** same time as another thread is invoking sqlite3_backup_step() it is
   ** possible that they return invalid values.
   */
   SQLITE_API sqlite3_backup *sqlite3_backup_init(
     sqlite3 *pDest,                        /* Destination database handle */
     const char *zDestName,                 /* Destination database name */
     sqlite3 *pSource,                      /* Source database handle */
     const char *zSourceName                /* Source database name */
   );
   SQLITE_API int sqlite3_backup_step(sqlite3_backup *p, int nPage);
   SQLITE_API int sqlite3_backup_finish(sqlite3_backup *p);
   SQLITE_API int sqlite3_backup_remaining(sqlite3_backup *p);
   SQLITE_API int sqlite3_backup_pagecount(sqlite3_backup *p);
   
   /*
   ** CAPI3REF: Unlock Notification
   **
   ** ^When running in shared-cache mode, a database operation may fail with
   ** an [SQLITE_LOCKED] error if the required locks on the shared-cache or
   ** individual tables within the shared-cache cannot be obtained. See
   ** [SQLite Shared-Cache Mode] for a description of shared-cache locking. 
   ** ^This API may be used to register a callback that SQLite will invoke 
   ** when the connection currently holding the required lock relinquishes it.
   ** ^This API is only available if the library was compiled with the
   ** [SQLITE_ENABLE_UNLOCK_NOTIFY] C-preprocessor symbol defined.
   **
   ** See Also: [Using the SQLite Unlock Notification Feature].
   **
   ** ^Shared-cache locks are released when a database connection concludes
   ** its current transaction, either by committing it or rolling it back. 
   **
   ** ^When a connection (known as the blocked connection) fails to obtain a
   ** shared-cache lock and SQLITE_LOCKED is returned to the caller, the
   ** identity of the database connection (the blocking connection) that
   ** has locked the required resource is stored internally. ^After an 
   ** application receives an SQLITE_LOCKED error, it may call the
   ** sqlite3_unlock_notify() method with the blocked connection handle as 
   ** the first argument to register for a callback that will be invoked
   ** when the blocking connections current transaction is concluded. ^The
   ** callback is invoked from within the [sqlite3_step] or [sqlite3_close]
   ** call that concludes the blocking connections transaction.
   **
   ** ^(If sqlite3_unlock_notify() is called in a multi-threaded application,
   ** there is a chance that the blocking connection will have already
   ** concluded its transaction by the time sqlite3_unlock_notify() is invoked.
   ** If this happens, then the specified callback is invoked immediately,
   ** from within the call to sqlite3_unlock_notify().)^
   **
   ** ^If the blocked connection is attempting to obtain a write-lock on a
   ** shared-cache table, and more than one other connection currently holds
   ** a read-lock on the same table, then SQLite arbitrarily selects one of 
   ** the other connections to use as the blocking connection.
   **
   ** ^(There may be at most one unlock-notify callback registered by a 
   ** blocked connection. If sqlite3_unlock_notify() is called when the
   ** blocked connection already has a registered unlock-notify callback,
   ** then the new callback replaces the old.)^ ^If sqlite3_unlock_notify() is
   ** called with a NULL pointer as its second argument, then any existing
   ** unlock-notify callback is canceled. ^The blocked connections 
   ** unlock-notify callback may also be canceled by closing the blocked
   ** connection using [sqlite3_close()].
   **
   ** The unlock-notify callback is not reentrant. If an application invokes
   ** any sqlite3_xxx API functions from within an unlock-notify callback, a
   ** crash or deadlock may be the result.
   **
   ** ^Unless deadlock is detected (see below), sqlite3_unlock_notify() always
   ** returns SQLITE_OK.
   **
   ** <b>Callback Invocation Details</b>
   **
   ** When an unlock-notify callback is registered, the application provides a 
   ** single void* pointer that is passed to the callback when it is invoked.
   ** However, the signature of the callback function allows SQLite to pass
   ** it an array of void* context pointers. The first argument passed to
   ** an unlock-notify callback is a pointer to an array of void* pointers,
   ** and the second is the number of entries in the array.
   **
   ** When a blocking connections transaction is concluded, there may be
   ** more than one blocked connection that has registered for an unlock-notify
   ** callback. ^If two or more such blocked connections have specified the
   ** same callback function, then instead of invoking the callback function
   ** multiple times, it is invoked once with the set of void* context pointers
   ** specified by the blocked connections bundled together into an array.
   ** This gives the application an opportunity to prioritize any actions 
   ** related to the set of unblocked database connections.
   **
   ** <b>Deadlock Detection</b>
   **
   ** Assuming that after registering for an unlock-notify callback a 
   ** database waits for the callback to be issued before taking any further
   ** action (a reasonable assumption), then using this API may cause the
   ** application to deadlock. For example, if connection X is waiting for
   ** connection Y's transaction to be concluded, and similarly connection
   ** Y is waiting on connection X's transaction, then neither connection
   ** will proceed and the system may remain deadlocked indefinitely.
   **
   ** To avoid this scenario, the sqlite3_unlock_notify() performs deadlock
   ** detection. ^If a given call to sqlite3_unlock_notify() would put the
   ** system in a deadlocked state, then SQLITE_LOCKED is returned and no
   ** unlock-notify callback is registered. The system is said to be in
   ** a deadlocked state if connection A has registered for an unlock-notify
   ** callback on the conclusion of connection B's transaction, and connection
   ** B has itself registered for an unlock-notify callback when connection
   ** A's transaction is concluded. ^Indirect deadlock is also detected, so
   ** the system is also considered to be deadlocked if connection B has
   ** registered for an unlock-notify callback on the conclusion of connection
   ** C's transaction, where connection C is waiting on connection A. ^Any
   ** number of levels of indirection are allowed.
   **
   ** <b>The "DROP TABLE" Exception</b>
   **
   ** When a call to [sqlite3_step()] returns SQLITE_LOCKED, it is almost 
   ** always appropriate to call sqlite3_unlock_notify(). There is however,
   ** one exception. When executing a "DROP TABLE" or "DROP INDEX" statement,
   ** SQLite checks if there are any currently executing SELECT statements
   ** that belong to the same connection. If there are, SQLITE_LOCKED is
   ** returned. In this case there is no "blocking connection", so invoking
   ** sqlite3_unlock_notify() results in the unlock-notify callback being
   ** invoked immediately. If the application then re-attempts the "DROP TABLE"
   ** or "DROP INDEX" query, an infinite loop might be the result.
   **
   ** One way around this problem is to check the extended error code returned
   ** by an sqlite3_step() call. ^(If there is a blocking connection, then the
   ** extended error code is set to SQLITE_LOCKED_SHAREDCACHE. Otherwise, in
   ** the special "DROP TABLE/INDEX" case, the extended error code is just 
   ** SQLITE_LOCKED.)^
   */
   SQLITE_API int sqlite3_unlock_notify(
     sqlite3 *pBlocked,                          /* Waiting connection */
     void (*xNotify)(void **apArg, int nArg),    /* Callback function to invoke */
     void *pNotifyArg                            /* Argument to pass to xNotify */
   );
   
   
   /*
   ** CAPI3REF: String Comparison
   **
   ** ^The [sqlite3_strnicmp()] API allows applications and extensions to
   ** compare the contents of two buffers containing UTF-8 strings in a
   ** case-independent fashion, using the same definition of case independence 
   ** that SQLite uses internally when comparing identifiers.
   */
   SQLITE_API int sqlite3_strnicmp(const char *, const char *, int);
   
   /*
   ** CAPI3REF: Error Logging Interface
   **
   ** ^The [sqlite3_log()] interface writes a message into the error log
   ** established by the [SQLITE_CONFIG_LOG] option to [sqlite3_config()].
   ** ^If logging is enabled, the zFormat string and subsequent arguments are
   ** used with [sqlite3_snprintf()] to generate the final output string.
   **
   ** The sqlite3_log() interface is intended for use by extensions such as
   ** virtual tables, collating functions, and SQL functions.  While there is
   ** nothing to prevent an application from calling sqlite3_log(), doing so
   ** is considered bad form.
   **
   ** The zFormat string must not be NULL.
   **
   ** To avoid deadlocks and other threading problems, the sqlite3_log() routine
   ** will not use dynamically allocated memory.  The log message is stored in
   ** a fixed-length buffer on the stack.  If the log message is longer than
   ** a few hundred characters, it will be truncated to the length of the
   ** buffer.
   */
   SQLITE_API void sqlite3_log(int iErrCode, const char *zFormat, ...);
   
   /*
   ** CAPI3REF: Write-Ahead Log Commit Hook
   **
   ** ^The [sqlite3_wal_hook()] function is used to register a callback that
   ** will be invoked each time a database connection commits data to a
   ** [write-ahead log] (i.e. whenever a transaction is committed in
   ** [journal_mode | journal_mode=WAL mode]). 
   **
   ** ^The callback is invoked by SQLite after the commit has taken place and 
   ** the associated write-lock on the database released, so the implementation 
   ** may read, write or [checkpoint] the database as required.
   **
   ** ^The first parameter passed to the callback function when it is invoked
   ** is a copy of the third parameter passed to sqlite3_wal_hook() when
   ** registering the callback. ^The second is a copy of the database handle.
   ** ^The third parameter is the name of the database that was written to -
   ** either "main" or the name of an [ATTACH]-ed database. ^The fourth parameter
   ** is the number of pages currently in the write-ahead log file,
   ** including those that were just committed.
   **
   ** The callback function should normally return [SQLITE_OK].  ^If an error
   ** code is returned, that error will propagate back up through the
   ** SQLite code base to cause the statement that provoked the callback
   ** to report an error, though the commit will have still occurred. If the
   ** callback returns [SQLITE_ROW] or [SQLITE_DONE], or if it returns a value
   ** that does not correspond to any valid SQLite error code, the results
   ** are undefined.
   **
   ** A single database handle may have at most a single write-ahead log callback 
   ** registered at one time. ^Calling [sqlite3_wal_hook()] replaces any
   ** previously registered write-ahead log callback. ^Note that the
   ** [sqlite3_wal_autocheckpoint()] interface and the
   ** [wal_autocheckpoint pragma] both invoke [sqlite3_wal_hook()] and will
   ** those overwrite any prior [sqlite3_wal_hook()] settings.
   */
   SQLITE_API void *sqlite3_wal_hook(
     sqlite3*, 
     int(*)(void *,sqlite3*,const char*,int),
     void*
   );
   
   /*
   ** CAPI3REF: Configure an auto-checkpoint
   **
   ** ^The [sqlite3_wal_autocheckpoint(D,N)] is a wrapper around
   ** [sqlite3_wal_hook()] that causes any database on [database connection] D
   ** to automatically [checkpoint]
   ** after committing a transaction if there are N or
   ** more frames in the [write-ahead log] file.  ^Passing zero or 
   ** a negative value as the nFrame parameter disables automatic
   ** checkpoints entirely.
   **
   ** ^The callback registered by this function replaces any existing callback
   ** registered using [sqlite3_wal_hook()].  ^Likewise, registering a callback
   ** using [sqlite3_wal_hook()] disables the automatic checkpoint mechanism
   ** configured by this function.
   **
   ** ^The [wal_autocheckpoint pragma] can be used to invoke this interface
   ** from SQL.
   **
   ** ^Every new [database connection] defaults to having the auto-checkpoint
   ** enabled with a threshold of 1000 or [SQLITE_DEFAULT_WAL_AUTOCHECKPOINT]
   ** pages.  The use of this interface
   ** is only necessary if the default setting is found to be suboptimal
   ** for a particular application.
   */
   SQLITE_API int sqlite3_wal_autocheckpoint(sqlite3 *db, int N);
   
   /*
   ** CAPI3REF: Checkpoint a database
   **
   ** ^The [sqlite3_wal_checkpoint(D,X)] interface causes database named X
   ** on [database connection] D to be [checkpointed].  ^If X is NULL or an
   ** empty string, then a checkpoint is run on all databases of
   ** connection D.  ^If the database connection D is not in
   ** [WAL | write-ahead log mode] then this interface is a harmless no-op.
   **
   ** ^The [wal_checkpoint pragma] can be used to invoke this interface
   ** from SQL.  ^The [sqlite3_wal_autocheckpoint()] interface and the
   ** [wal_autocheckpoint pragma] can be used to cause this interface to be
   ** run whenever the WAL reaches a certain size threshold.
   **
   ** See also: [sqlite3_wal_checkpoint_v2()]
   */
   SQLITE_API int sqlite3_wal_checkpoint(sqlite3 *db, const char *zDb);
   
   /*
   ** CAPI3REF: Checkpoint a database
   **
   ** Run a checkpoint operation on WAL database zDb attached to database 
   ** handle db. The specific operation is determined by the value of the 
   ** eMode parameter:
   **
   ** <dl>
   ** <dt>SQLITE_CHECKPOINT_PASSIVE<dd>
   **   Checkpoint as many frames as possible without waiting for any database 
   **   readers or writers to finish. Sync the db file if all frames in the log
   **   are checkpointed. This mode is the same as calling 
   **   sqlite3_wal_checkpoint(). The busy-handler callback is never invoked.
   **
   ** <dt>SQLITE_CHECKPOINT_FULL<dd>
   **   This mode blocks (calls the busy-handler callback) until there is no
   **   database writer and all readers are reading from the most recent database
   **   snapshot. It then checkpoints all frames in the log file and syncs the
   **   database file. This call blocks database writers while it is running,
   **   but not database readers.
   **
   ** <dt>SQLITE_CHECKPOINT_RESTART<dd>
   **   This mode works the same way as SQLITE_CHECKPOINT_FULL, except after 
   **   checkpointing the log file it blocks (calls the busy-handler callback)
   **   until all readers are reading from the database file only. This ensures 
   **   that the next client to write to the database file restarts the log file 
   **   from the beginning. This call blocks database writers while it is running,
   **   but not database readers.
   ** </dl>
   **
   ** If pnLog is not NULL, then *pnLog is set to the total number of frames in
   ** the log file before returning. If pnCkpt is not NULL, then *pnCkpt is set to
   ** the total number of checkpointed frames (including any that were already
   ** checkpointed when this function is called). *pnLog and *pnCkpt may be
   ** populated even if sqlite3_wal_checkpoint_v2() returns other than SQLITE_OK.
   ** If no values are available because of an error, they are both set to -1
   ** before returning to communicate this to the caller.
   **
   ** All calls obtain an exclusive "checkpoint" lock on the database file. If
   ** any other process is running a checkpoint operation at the same time, the 
   ** lock cannot be obtained and SQLITE_BUSY is returned. Even if there is a 
   ** busy-handler configured, it will not be invoked in this case.
   **
   ** The SQLITE_CHECKPOINT_FULL and RESTART modes also obtain the exclusive 
   ** "writer" lock on the database file. If the writer lock cannot be obtained
   ** immediately, and a busy-handler is configured, it is invoked and the writer
   ** lock retried until either the busy-handler returns 0 or the lock is
   ** successfully obtained. The busy-handler is also invoked while waiting for
   ** database readers as described above. If the busy-handler returns 0 before
   ** the writer lock is obtained or while waiting for database readers, the
   ** checkpoint operation proceeds from that point in the same way as 
   ** SQLITE_CHECKPOINT_PASSIVE - checkpointing as many frames as possible 
   ** without blocking any further. SQLITE_BUSY is returned in this case.
   **
   ** If parameter zDb is NULL or points to a zero length string, then the
   ** specified operation is attempted on all WAL databases. In this case the
   ** values written to output parameters *pnLog and *pnCkpt are undefined. If 
   ** an SQLITE_BUSY error is encountered when processing one or more of the 
   ** attached WAL databases, the operation is still attempted on any remaining 
   ** attached databases and SQLITE_BUSY is returned to the caller. If any other 
   ** error occurs while processing an attached database, processing is abandoned 
   ** and the error code returned to the caller immediately. If no error 
   ** (SQLITE_BUSY or otherwise) is encountered while processing the attached 
   ** databases, SQLITE_OK is returned.
   **
   ** If database zDb is the name of an attached database that is not in WAL
   ** mode, SQLITE_OK is returned and both *pnLog and *pnCkpt set to -1. If
   ** zDb is not NULL (or a zero length string) and is not the name of any
   ** attached database, SQLITE_ERROR is returned to the caller.
   */
   SQLITE_API int sqlite3_wal_checkpoint_v2(
     sqlite3 *db,                    /* Database handle */
     const char *zDb,                /* Name of attached database (or NULL) */
     int eMode,                      /* SQLITE_CHECKPOINT_* value */
     int *pnLog,                     /* OUT: Size of WAL log in frames */
     int *pnCkpt                     /* OUT: Total number of frames checkpointed */
   );
   
   /*
   ** CAPI3REF: Checkpoint operation parameters
   **
   ** These constants can be used as the 3rd parameter to
   ** [sqlite3_wal_checkpoint_v2()].  See the [sqlite3_wal_checkpoint_v2()]
   ** documentation for additional information about the meaning and use of
   ** each of these values.
   */
   #define SQLITE_CHECKPOINT_PASSIVE 0
   #define SQLITE_CHECKPOINT_FULL    1
   #define SQLITE_CHECKPOINT_RESTART 2
   
   /*
   ** CAPI3REF: Virtual Table Interface Configuration
   **
   ** This function may be called by either the [xConnect] or [xCreate] method
   ** of a [virtual table] implementation to configure
   ** various facets of the virtual table interface.
   **
   ** If this interface is invoked outside the context of an xConnect or
   ** xCreate virtual table method then the behavior is undefined.
   **
   ** At present, there is only one option that may be configured using
   ** this function. (See [SQLITE_VTAB_CONSTRAINT_SUPPORT].)  Further options
   ** may be added in the future.
   */
   SQLITE_API int sqlite3_vtab_config(sqlite3*, int op, ...);
   
   /*
   ** CAPI3REF: Virtual Table Configuration Options
   **
   ** These macros define the various options to the
   ** [sqlite3_vtab_config()] interface that [virtual table] implementations
   ** can use to customize and optimize their behavior.
   **
   ** <dl>
   ** <dt>SQLITE_VTAB_CONSTRAINT_SUPPORT
   ** <dd>Calls of the form
   ** [sqlite3_vtab_config](db,SQLITE_VTAB_CONSTRAINT_SUPPORT,X) are supported,
   ** where X is an integer.  If X is zero, then the [virtual table] whose
   ** [xCreate] or [xConnect] method invoked [sqlite3_vtab_config()] does not
   ** support constraints.  In this configuration (which is the default) if
   ** a call to the [xUpdate] method returns [SQLITE_CONSTRAINT], then the entire
   ** statement is rolled back as if [ON CONFLICT | OR ABORT] had been
   ** specified as part of the users SQL statement, regardless of the actual
   ** ON CONFLICT mode specified.
   **
   ** If X is non-zero, then the virtual table implementation guarantees
   ** that if [xUpdate] returns [SQLITE_CONSTRAINT], it will do so before
   ** any modifications to internal or persistent data structures have been made.
   ** If the [ON CONFLICT] mode is ABORT, FAIL, IGNORE or ROLLBACK, SQLite 
   ** is able to roll back a statement or database transaction, and abandon
   ** or continue processing the current SQL statement as appropriate. 
   ** If the ON CONFLICT mode is REPLACE and the [xUpdate] method returns
   ** [SQLITE_CONSTRAINT], SQLite handles this as if the ON CONFLICT mode
   ** had been ABORT.
   **
   ** Virtual table implementations that are required to handle OR REPLACE
   ** must do so within the [xUpdate] method. If a call to the 
   ** [sqlite3_vtab_on_conflict()] function indicates that the current ON 
   ** CONFLICT policy is REPLACE, the virtual table implementation should 
   ** silently replace the appropriate rows within the xUpdate callback and
   ** return SQLITE_OK. Or, if this is not possible, it may return
   ** SQLITE_CONSTRAINT, in which case SQLite falls back to OR ABORT 
   ** constraint handling.
   ** </dl>
   */
   #define SQLITE_VTAB_CONSTRAINT_SUPPORT 1
   
   /*
   ** CAPI3REF: Determine The Virtual Table Conflict Policy
   **
   ** This function may only be called from within a call to the [xUpdate] method
   ** of a [virtual table] implementation for an INSERT or UPDATE operation. ^The
   ** value returned is one of [SQLITE_ROLLBACK], [SQLITE_IGNORE], [SQLITE_FAIL],
   ** [SQLITE_ABORT], or [SQLITE_REPLACE], according to the [ON CONFLICT] mode
   ** of the SQL statement that triggered the call to the [xUpdate] method of the
   ** [virtual table].
   */
   SQLITE_API int sqlite3_vtab_on_conflict(sqlite3 *);
   
   /*
   ** CAPI3REF: Conflict resolution modes
   **
   ** These constants are returned by [sqlite3_vtab_on_conflict()] to
   ** inform a [virtual table] implementation what the [ON CONFLICT] mode
   ** is for the SQL statement being evaluated.
   **
   ** Note that the [SQLITE_IGNORE] constant is also used as a potential
   ** return value from the [sqlite3_set_authorizer()] callback and that
   ** [SQLITE_ABORT] is also a [result code].
   */
   #define SQLITE_ROLLBACK 1
   /* #define SQLITE_IGNORE 2 // Also used by sqlite3_authorizer() callback */
   #define SQLITE_FAIL     3
   /* #define SQLITE_ABORT 4  // Also an error code */
   #define SQLITE_REPLACE  5
   
   
   
   /*
   ** Undo the hack that converts floating point types to integer for
   ** builds on processors without floating point support.
   */
   #ifdef SQLITE_OMIT_FLOATING_POINT
   # undef double
   #endif
   
   #ifdef __cplusplus
   }  /* End of the 'extern "C"' block */
   #endif
   #endif
   
   /*
   ** 2010 August 30
   **
   ** The author disclaims copyright to this source code.  In place of
   ** a legal notice, here is a blessing:
   **
   **    May you do good and not evil.
   **    May you find forgiveness for yourself and forgive others.
   **    May you share freely, never taking more than you give.
   **
   *************************************************************************
   */
   
   #ifndef _SQLITE3RTREE_H_
   #define _SQLITE3RTREE_H_
   
   
   #ifdef __cplusplus
   extern "C" {
   #endif
   
   typedef struct sqlite3_rtree_geometry sqlite3_rtree_geometry;
   
   /*
   ** Register a geometry callback named zGeom that can be used as part of an
   ** R-Tree geometry query as follows:
   **
   **   SELECT ... FROM <rtree> WHERE <rtree col> MATCH $zGeom(... params ...)
   */
   SQLITE_API int sqlite3_rtree_geometry_callback(
     sqlite3 *db,
     const char *zGeom,
     int (*xGeom)(sqlite3_rtree_geometry *, int nCoord, double *aCoord, int *pRes),
     void *pContext
   );
   
   
   /*
   ** A pointer to a structure of the following type is passed as the first
   ** argument to callbacks registered using rtree_geometry_callback().
   */
   struct sqlite3_rtree_geometry {
     void *pContext;                 /* Copy of pContext passed to s_r_g_c() */
     int nParam;                     /* Size of array aParam[] */
     double *aParam;                 /* Parameters passed to SQL geom function */
     void *pUser;                    /* Callback implementation user data */
     void (*xDelUser)(void *);       /* Called by SQLite to clean up pUser */
   };
   
   
   #ifdef __cplusplus
   }  /* end of the 'extern "C"' block */
   #endif
   
   #endif  /* ifndef _SQLITE3RTREE_H_ */
   
Removed from v.1.1  
changed lines
  Added in v.1.2

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>