cyrillic_and_mic.c

来自「postgresql8.3.4源码,开源数据库」· C语言 代码 · 共 643 行 · 第 1/2 页

/*-------------------------------------------------------------------------
 *
 *	  Cyrillic and MULE_INTERNAL
 *
 * Portions Copyright (c) 1996-2008, PostgreSQL Global Development Group
 * Portions Copyright (c) 1994, Regents of the University of California
 *
 * IDENTIFICATION
 *	  $PostgreSQL: pgsql/src/backend/utils/mb/conversion_procs/cyrillic_and_mic/cyrillic_and_mic.c,v 1.16.2.1 2008/03/20 10:35:45 heikki Exp $
 *
 *-------------------------------------------------------------------------
 */

#include "postgres.h"
#include "fmgr.h"
#include "mb/pg_wchar.h"

#define ENCODING_GROWTH_RATE 4

PG_MODULE_MAGIC;

PG_FUNCTION_INFO_V1(koi8r_to_mic);
PG_FUNCTION_INFO_V1(mic_to_koi8r);
PG_FUNCTION_INFO_V1(iso_to_mic);
PG_FUNCTION_INFO_V1(mic_to_iso);
PG_FUNCTION_INFO_V1(win1251_to_mic);
PG_FUNCTION_INFO_V1(mic_to_win1251);
PG_FUNCTION_INFO_V1(win866_to_mic);
PG_FUNCTION_INFO_V1(mic_to_win866);
PG_FUNCTION_INFO_V1(koi8r_to_win1251);
PG_FUNCTION_INFO_V1(win1251_to_koi8r);
PG_FUNCTION_INFO_V1(koi8r_to_win866);
PG_FUNCTION_INFO_V1(win866_to_koi8r);
PG_FUNCTION_INFO_V1(win866_to_win1251);
PG_FUNCTION_INFO_V1(win1251_to_win866);
PG_FUNCTION_INFO_V1(iso_to_koi8r);
PG_FUNCTION_INFO_V1(koi8r_to_iso);
PG_FUNCTION_INFO_V1(iso_to_win1251);
PG_FUNCTION_INFO_V1(win1251_to_iso);
PG_FUNCTION_INFO_V1(iso_to_win866);
PG_FUNCTION_INFO_V1(win866_to_iso);

extern Datum koi8r_to_mic(PG_FUNCTION_ARGS);
extern Datum mic_to_koi8r(PG_FUNCTION_ARGS);
extern Datum iso_to_mic(PG_FUNCTION_ARGS);
extern Datum mic_to_iso(PG_FUNCTION_ARGS);
extern Datum win1251_to_mic(PG_FUNCTION_ARGS);
extern Datum mic_to_win1251(PG_FUNCTION_ARGS);
extern Datum win866_to_mic(PG_FUNCTION_ARGS);
extern Datum mic_to_win866(PG_FUNCTION_ARGS);
extern Datum koi8r_to_win1251(PG_FUNCTION_ARGS);
extern Datum win1251_to_koi8r(PG_FUNCTION_ARGS);
extern Datum koi8r_to_win866(PG_FUNCTION_ARGS);
extern Datum win866_to_koi8r(PG_FUNCTION_ARGS);
extern Datum win866_to_win1251(PG_FUNCTION_ARGS);
extern Datum win1251_to_win866(PG_FUNCTION_ARGS);
extern Datum iso_to_koi8r(PG_FUNCTION_ARGS);
extern Datum koi8r_to_iso(PG_FUNCTION_ARGS);
extern Datum iso_to_win1251(PG_FUNCTION_ARGS);
extern Datum win1251_to_iso(PG_FUNCTION_ARGS);
extern Datum iso_to_win866(PG_FUNCTION_ARGS);
extern Datum win866_to_iso(PG_FUNCTION_ARGS);

/* ----------
 * conv_proc(
 *		INTEGER,	-- source encoding id
 *		INTEGER,	-- destination encoding id
 *		CSTRING,	-- source string (null terminated C string)
 *		CSTRING,	-- destination string (null terminated C string)
 *		INTEGER		-- source string length
 * ) returns VOID;
 * ----------
 */

static void koi8r2mic(const unsigned char *l, unsigned char *p, int len);
static void mic2koi8r(const unsigned char *mic, unsigned char *p, int len);
static void iso2mic(const unsigned char *l, unsigned char *p, int len);
static void mic2iso(const unsigned char *mic, unsigned char *p, int len);
static void win12512mic(const unsigned char *l, unsigned char *p, int len);
static void mic2win1251(const unsigned char *mic, unsigned char *p, int len);
static void win8662mic(const unsigned char *l, unsigned char *p, int len);
static void mic2win866(const unsigned char *mic, unsigned char *p, int len);

Datum
koi8r_to_mic(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);

	Assert(PG_GETARG_INT32(0) == PG_KOI8R);
	Assert(PG_GETARG_INT32(1) == PG_MULE_INTERNAL);
	Assert(len >= 0);

	koi8r2mic(src, dest, len);

	PG_RETURN_VOID();
}

Datum
mic_to_koi8r(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);

	Assert(PG_GETARG_INT32(0) == PG_MULE_INTERNAL);
	Assert(PG_GETARG_INT32(1) == PG_KOI8R);
	Assert(len >= 0);

	mic2koi8r(src, dest, len);

	PG_RETURN_VOID();
}

Datum
iso_to_mic(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);

	Assert(PG_GETARG_INT32(0) == PG_ISO_8859_5);
	Assert(PG_GETARG_INT32(1) == PG_MULE_INTERNAL);
	Assert(len >= 0);

	iso2mic(src, dest, len);

	PG_RETURN_VOID();
}

Datum
mic_to_iso(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);

	Assert(PG_GETARG_INT32(0) == PG_MULE_INTERNAL);
	Assert(PG_GETARG_INT32(1) == PG_ISO_8859_5);
	Assert(len >= 0);

	mic2iso(src, dest, len);

	PG_RETURN_VOID();
}

Datum
win1251_to_mic(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);

	Assert(PG_GETARG_INT32(0) == PG_WIN1251);
	Assert(PG_GETARG_INT32(1) == PG_MULE_INTERNAL);
	Assert(len >= 0);

	win12512mic(src, dest, len);

	PG_RETURN_VOID();
}

Datum
mic_to_win1251(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);

	Assert(PG_GETARG_INT32(0) == PG_MULE_INTERNAL);
	Assert(PG_GETARG_INT32(1) == PG_WIN1251);
	Assert(len >= 0);

	mic2win1251(src, dest, len);

	PG_RETURN_VOID();
}

Datum
win866_to_mic(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);

	Assert(PG_GETARG_INT32(0) == PG_WIN866);
	Assert(PG_GETARG_INT32(1) == PG_MULE_INTERNAL);
	Assert(len >= 0);

	win8662mic(src, dest, len);

	PG_RETURN_VOID();
}

Datum
mic_to_win866(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);

	Assert(PG_GETARG_INT32(0) == PG_MULE_INTERNAL);
	Assert(PG_GETARG_INT32(1) == PG_WIN866);
	Assert(len >= 0);

	mic2win866(src, dest, len);

	PG_RETURN_VOID();
}

Datum
koi8r_to_win1251(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);
	unsigned char *buf;

	Assert(PG_GETARG_INT32(0) == PG_KOI8R);
	Assert(PG_GETARG_INT32(1) == PG_WIN1251);
	Assert(len >= 0);

	buf = palloc(len * ENCODING_GROWTH_RATE);
	koi8r2mic(src, buf, len);
	mic2win1251(buf, dest, strlen((char *) buf));
	pfree(buf);

	PG_RETURN_VOID();
}

Datum
win1251_to_koi8r(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);
	unsigned char *buf;

	Assert(PG_GETARG_INT32(0) == PG_WIN1251);
	Assert(PG_GETARG_INT32(1) == PG_KOI8R);
	Assert(len >= 0);

	buf = palloc(len * ENCODING_GROWTH_RATE);
	win12512mic(src, buf, len);
	mic2koi8r(buf, dest, strlen((char *) buf));
	pfree(buf);

	PG_RETURN_VOID();
}

Datum
koi8r_to_win866(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);
	unsigned char *buf;

	Assert(PG_GETARG_INT32(0) == PG_KOI8R);
	Assert(PG_GETARG_INT32(1) == PG_WIN866);
	Assert(len >= 0);

	buf = palloc(len * ENCODING_GROWTH_RATE);
	koi8r2mic(src, buf, len);
	mic2win866(buf, dest, strlen((char *) buf));
	pfree(buf);

	PG_RETURN_VOID();
}

Datum
win866_to_koi8r(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);
	unsigned char *buf;

	Assert(PG_GETARG_INT32(0) == PG_WIN866);
	Assert(PG_GETARG_INT32(1) == PG_KOI8R);
	Assert(len >= 0);

	buf = palloc(len * ENCODING_GROWTH_RATE);
	win8662mic(src, buf, len);
	mic2koi8r(buf, dest, strlen((char *) buf));
	pfree(buf);

	PG_RETURN_VOID();
}

Datum
win866_to_win1251(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
	int			len = PG_GETARG_INT32(4);
	unsigned char *buf;

	Assert(PG_GETARG_INT32(0) == PG_WIN866);
	Assert(PG_GETARG_INT32(1) == PG_WIN1251);
	Assert(len >= 0);

	/*
	 * Note: There are a few characters like the "Numero" sign that exist in
	 * all the other cyrillic encodings (win1251, ISO_8859-5 and cp866), but
	 * not in KOI8R. As we use MULE_INTERNAL/KOI8R as an intermediary, we
	 * will fail to convert those characters.
	 */
	buf = palloc(len * ENCODING_GROWTH_RATE);
	win8662mic(src, buf, len);
	mic2win1251(buf, dest, strlen((char *) buf));
	pfree(buf);

	PG_RETURN_VOID();
}

Datum
win1251_to_win866(PG_FUNCTION_ARGS)
{
	unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
	unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);