iconv.c

来自「在Linux/Unix下面访问WINDOWS SQLSERVER 的ODBC驱动」· C语言 代码 · 共 1,453 行 · 第 1/3 页

	int canonic_charset_num = tds_canonical_charset(charset);
	const char *canonic_charset;

	if (tds->major_version >= 7 && canonic_charset_num == TDS_CHARSET_ISO_8859_1)
		canonic_charset_num = TDS_CHARSET_CP1252;

	/* ignore request to change to unknown charset */
	if (canonic_charset_num < 0) {
		tdsdump_log(TDS_DBG_FUNC, "tds_srv_charset_changed: what is charset \"%s\"?\n", charset);
		return;
	}
	canonic_charset = canonic_charsets[canonic_charset_num].name;

	tdsdump_log(TDS_DBG_FUNC, "setting server single-byte charset to \"%s\"\n", canonic_charset);

	if (strcmp(canonic_charset, char_conv->server_charset.name) == 0)
		return;

	/* find and set conversion */
	char_conv = tds_iconv_get_info(tds, canonic_charset);
	if (char_conv)
		tds->char_convs[client2server_chardata] = char_conv;

	/* if sybase change also server conversions */
	if (tds->major_version >= 7)
		return;

	char_conv = tds->char_convs[iso2server_metadata];

	tds_iconv_info_close(char_conv);

	tds_iconv_info_init(char_conv, "ISO-8859-1", charset);
#endif
}

/* change singlebyte conversions according to server */
void
tds7_srv_charset_changed(TDSSOCKET * tds, int sql_collate, int lcid)
{
	tds_srv_charset_changed(tds, collate2charset(sql_collate, lcid));
}

#if !HAVE_ICONV_ALWAYS
/**
 * Determine byte/char for an iconv character set.  
 * \retval 0 failed, no such charset.
 * \retval 1 succeeded, fixed byte/char.
 * \retval 2 succeeded, variable byte/char.
 */
static int
bytes_per_char(TDS_ENCODING * charset)
{
	int i;

	assert(charset && strlen(charset->name) < sizeof(charset->name));

	for (i = 0; i < sizeof(canonic_charsets) / sizeof(TDS_ENCODING); i++) {
		if (canonic_charsets[i].min_bytes_per_char == 0)
			break;

		if (0 == strcmp(charset->name, canonic_charsets[i].name)) {
			charset->min_bytes_per_char = canonic_charsets[i].min_bytes_per_char;
			charset->max_bytes_per_char = canonic_charsets[i].max_bytes_per_char;

			return (charset->max_bytes_per_char == charset->min_bytes_per_char) ? 1 : 2;
		}
	}

	return 0;
}
#endif

/**
 * Move the input sequence pointer to the next valid position.
 * Used when an input character cannot be converted.  
 * \returns number of bytes to skip.
 */
/* FIXME possible buffer reading overflow ?? */
static int
skip_one_input_sequence(iconv_t cd, const TDS_ENCODING * charset, const char **input, size_t * input_size)
{
	int charsize = CHARSIZE(charset);
	char ib[16];
	char ob[16];
	ICONV_CONST char *pib;
	char *pob;
	size_t il, ol, l;
	iconv_t cd2;


	/* usually fixed size and UTF-8 do not have state, so do not reset it */
	if (charsize) {
		*input += charsize;
		*input_size -= charsize;
		return charsize;
	}

	if (0 == strcmp(charset->name, "UTF-8")) {
		/*
		 * Deal with UTF-8.  
		 * bytes | bits | representation
		 *     1 |    7 | 0vvvvvvv
		 *     2 |   11 | 110vvvvv 10vvvvvv
		 *     3 |   16 | 1110vvvv 10vvvvvv 10vvvvvv
		 *     4 |   21 | 11110vvv 10vvvvvv 10vvvvvv 10vvvvvv
		 */
		int c = **input;

		c = c & (c >> 1);
		do {
			++charsize;
		} while ((c <<= 1) & 0x80);
		*input += charsize;
		*input_size -= charsize;
		return charsize;
	}

	/* handle state encoding */

	/* extract state from iconv */
	pob = ib;
	ol = sizeof(ib);
	tds_sys_iconv(cd, NULL, NULL, &pob, &ol);

	/* init destination conversion */
	/* TODO use largest fixed size for this platform */
	cd2 = tds_sys_iconv_open("UCS-4", charset->name);
	if (cd2 == (iconv_t) - 1)
		return 0;

	/* add part of input */
	il = ol;
	if (il > *input_size)
		il = *input_size;
	l = sizeof(ib) - ol;
	memcpy(ib + l, *input, il);
	il += l;

	/* translate a single character */
	pib = ib;
	pob = ob;
	/* TODO use size of largest fixed charset */
	ol = 4;
	tds_sys_iconv(cd2, &pib, &il, &pob, &ol);

	/* adjust input */
	l = (pib - ib) - l;
	*input += l;
	*input_size -= l;

	/* extract state */
	pob = ib;
	ol = sizeof(ib);
	tds_sys_iconv(cd, NULL, NULL, &pob, &ol);

	/* set input state */
	pib = ib;
	il = sizeof(ib) - ol;
	pob = ob;
	ol = sizeof(ob);
	tds_sys_iconv(cd, &pib, &il, &pob, &ol);

	tds_sys_iconv_close(cd2);

	return l;
}

static int
lookup_canonic(const CHARACTER_SET_ALIAS aliases[], const char *charset_name)
{
	int i;

	for (i = 0; aliases[i].alias; ++i) {
		if (0 == strcmp(charset_name, aliases[i].alias))
			return aliases[i].canonic;
	}

	return -1;
}

/**
 * Determine canonical iconv character set.
 * \returns canonical position, or -1 if lookup failed.
 * \remarks Returned name can be used in bytes_per_char(), above.
 */
static int
tds_canonical_charset(const char *charset_name)
{
	int res;

	/* search in alternative */
	res = lookup_canonic(iconv_aliases, charset_name);
	if (res >= 0)
		return res;

	/* search in sybase */
	return lookup_canonic(sybase_aliases, charset_name);
}

/**
 * Determine canonical iconv character set name.  
 * \returns canonical name, or NULL if lookup failed.
 * \remarks Returned name can be used in bytes_per_char(), above.
 */
const char *
tds_canonical_charset_name(const char *charset_name)
{
	int res;

	/* get numeric pos */
	res = tds_canonical_charset(charset_name);
	if (res >= 0)
		return canonic_charsets[res].name;

	return NULL;
}

/**
 * Determine the name Sybase uses for a character set, given a canonical iconv name.  
 * \returns Sybase name, or NULL if lookup failed.
 * \remarks Returned name can be sent to Sybase a server.
 */
const char *
tds_sybase_charset_name(const char *charset_name)
{
	int res, i;

	/* search in sybase */
	res = lookup_canonic(iconv_aliases, charset_name);
	if (res < 0)
		return NULL;

	/* special case, ignore ascii_8, take iso_1 instead, note index start from 1 */
	assert(strcmp(sybase_aliases[0].alias, "ascii_8") == 0);

	for (i = 1; sybase_aliases[i].alias; ++i) {
		if (sybase_aliases[i].canonic == res)
			return sybase_aliases[i].alias;
	}

	return NULL;
}

static const char *
collate2charset(int sql_collate, int lcid)
{
	/*
	 * The table from the MSQLServer reference "Windows Collation Designators" 
	 * and from " NLS Information for Microsoft Windows XP"
	 */

	const char *cp = NULL;

	switch (sql_collate) {
	case 30:		/* SQL_Latin1_General_CP437_BIN */
	case 31:		/* SQL_Latin1_General_CP437_CS_AS */
	case 32:		/* SQL_Latin1_General_CP437_CI_AS */
	case 33:		/* SQL_Latin1_General_Pref_CP437_CI_AS */
	case 34:		/* SQL_Latin1_General_CP437_CI_AI */
		return "CP437";
	case 40:		/* SQL_Latin1_General_CP850_BIN */
	case 41:		/* SQL_Latin1_General_CP850_CS_AS */
	case 42:		/* SQL_Latin1_General_CP850_CI_AS */
	case 43:		/* SQL_Latin1_General_Pref_CP850_CI_AS */
	case 44:		/* SQL_Latin1_General_CP850_CI_AI */
	case 49:		/* SQL_1xCompat_CP850_CI_AS */
	case 55:		/* SQL_AltDiction_CP850_CS_AS */
	case 56:		/* SQL_AltDiction_Pref_CP850_CI_AS */
	case 57:		/* SQL_AltDiction_CP850_CI_AI */
	case 58:		/* SQL_Scandinavian_Pref_CP850_CI_AS */
	case 59:		/* SQL_Scandinavian_CP850_CS_AS */
	case 60:		/* SQL_Scandinavian_CP850_CI_AS */
	case 61:		/* SQL_AltDiction_CP850_CI_AS */
		return "CP850";
	case 81:		/* SQL_Latin1_General_CP1250_CS_AS */
	case 82:		/* SQL_Latin1_General_CP1250_CI_AS */
		return "CP1250";
	case 105:		/* SQL_Latin1_General_CP1251_CS_AS */
	case 106:		/* SQL_Latin1_General_CP1251_CI_AS */
		return "CP1251";
	case 113:		/* SQL_Latin1_General_CP1253_CS_AS */
	case 114:		/* SQL_Latin1_General_CP1253_CI_AS */
	case 120:		/* SQL_MixDiction_CP1253_CS_AS */
	case 121:		/* SQL_AltDiction_CP1253_CS_AS */
	case 124:		/* SQL_Latin1_General_CP1253_CI_AI */
		return "CP1253";
	case 137:		/* SQL_Latin1_General_CP1255_CS_AS */
	case 138:		/* SQL_Latin1_General_CP1255_CI_AS */
		return "CP1255";
	case 145:		/* SQL_Latin1_General_CP1256_CS_AS */
	case 146:		/* SQL_Latin1_General_CP1256_CI_AS */
		return "CP1256";
	case 153:		/* SQL_Latin1_General_CP1257_CS_AS */
	case 154:		/* SQL_Latin1_General_CP1257_CI_AS */
		return "CP1257";
	}

	switch (lcid & 0xffff) {
	case 0x405:
	case 0x40e:		/* 0x1040e */
	case 0x415:
	case 0x418:
	case 0x41a:
	case 0x41b:
	case 0x41c:
	case 0x424:
		/* case 0x81a: seem wrong in XP table TODO check */
	case 0x104e:		/* ?? */
		cp = "CP1250";
		break;
	case 0x402:
	case 0x419:
	case 0x422:
	case 0x423:
	case 0x42f:
	case 0x43f:
	case 0x440:
	case 0x444:
	case 0x450:
	case 0x81a:		/* ?? */
	case 0x82c:
	case 0x843:
	case 0xc1a:
		cp = "CP1251";
		break;
	case 0x1007:
	case 0x1009:
	case 0x100a:
	case 0x100c:
	case 0x1407:
	case 0x1409:
	case 0x140a:
	case 0x140c:
	case 0x1809:
	case 0x180a:
	case 0x180c:
	case 0x1c09:
	case 0x1c0a:
	case 0x2009:
	case 0x200a:
	case 0x2409:
	case 0x240a:
	case 0x2809:
	case 0x280a:
	case 0x2c09:
	case 0x2c0a:
	case 0x3009:
	case 0x300a:
	case 0x3409:
	case 0x340a:
	case 0x380a:
	case 0x3c0a:
	case 0x400a:
	case 0x403:
	case 0x406:
	case 0x407:		/* 0x10407 */
	case 0x409:
	case 0x40a:
	case 0x40b:
	case 0x40c:
	case 0x40f:
	case 0x410:
	case 0x413:
	case 0x414:
	case 0x416:
	case 0x41d:
	case 0x421:
	case 0x42d:
	case 0x436:
	case 0x437:		/* 0x10437 */
	case 0x438:
		/*case 0x439:  ??? Unicode only */
	case 0x43e:
	case 0x440a:
	case 0x441:
	case 0x456:
	case 0x480a:
	case 0x4c0a:
	case 0x500a:
	case 0x807:
	case 0x809:
	case 0x80a:
	case 0x80c:
	case 0x810:
	case 0x813:
	case 0x814:
	case 0x816:
	case 0x81d:
	case 0x83e:
	case 0xc07:
	case 0xc09:
	case 0xc0a:
	case 0xc0c:
		cp = "CP1252";
		break;
	case 0x408:
		cp = "CP1253";
		break;
	case 0x41f:
	case 0x42c:
	case 0x443:
		cp = "CP1254";
		break;
	case 0x40d:
		cp = "CP1255";
		break;
	case 0x1001:
	case 0x1401:
	case 0x1801:
	case 0x1c01:
	case 0x2001:
	case 0x2401:
	case 0x2801:
	case 0x2c01:
	case 0x3001:
	case 0x3401:
	case 0x3801:
	case 0x3c01:
	case 0x4001:
	case 0x401:
	case 0x420:
	case 0x429:
	case 0x801:
	case 0xc01:
		cp = "CP1256";
		break;
	case 0x425:
	case 0x426:
	case 0x427:
	case 0x827:		/* ?? */
		cp = "CP1257";
		break;
	case 0x42a:
		cp = "CP1258";
		break;
	case 0x41e:
		cp = "CP874";
		break;
	case 0x411:		/* 0x10411 */
		cp = "CP932";
		break;
	case 0x1004:
	case 0x804:		/* 0x20804 */
		cp = "CP936";
		break;
	case 0x412:		/* 0x10412 */
		cp = "CP949";
		break;
	case 0x1404:
	case 0x404:		/* 0x30404 */
	case 0xc04:
		cp = "CP950";
		break;
	default:
		cp = "CP1252";
	}

	assert(cp);
	return cp;
}

/**
 * Get iconv information from a LCID (to support different column encoding under MSSQL2K)
 */
TDSICONV *
tds_iconv_from_collate(TDSSOCKET * tds, int sql_collate, int lcid)
{
	const char *charset = collate2charset(sql_collate, lcid);

#if ENABLE_EXTRA_CHECKS
	assert(strcmp(tds_canonical_charset_name(charset), charset) == 0);
#endif

	/* same as client (usually this is true, so this improve performance) ? */
	if (strcmp(tds->char_convs[client2server_chardata]->server_charset.name, charset) == 0)
		return tds->char_convs[client2server_chardata];

	return tds_iconv_get_info(tds, charset);
}

/** @} */