chklocale.c File Reference

#include "postgres.h"
#include <langinfo.h>
#include "mb/pg_wchar.h"

Include dependency graph for chklocale.c:

Go to the source code of this file.

Data Structures
struct	encoding_match

Functions
int	pg_get_encoding_from_locale (const char *ctype, bool write_message)

Variables
static const struct encoding_match	encoding_match_list []

Function Documentation

pg_get_encoding_from_locale()

int pg_get_encoding_from_locale	(	const char *	ctype,
		bool	write_message
	)

Definition at line 301 of file chklocale.c.

{
    char       *sys;
    int         i;
 
#ifndef WIN32
    locale_t    loc;
#endif
 
    /* Get the CODESET property, and also LC_CTYPE if not passed in */
    if (!ctype)
        ctype = setlocale(LC_CTYPE, NULL);
 
 
    /* If locale is C or POSIX, we can allow all encodings */
    if (pg_strcasecmp(ctype, "C") == 0 ||
        pg_strcasecmp(ctype, "POSIX") == 0)
        return PG_SQL_ASCII;
 
 
#ifndef WIN32
    loc = newlocale(LC_CTYPE_MASK, ctype, (locale_t) 0);
    if (loc == (locale_t) 0)
        return -1;              /* bogus ctype passed in? */
 
    sys = nl_langinfo_l(CODESET, loc);
    if (sys)
        sys = strdup(sys);
 
    freelocale(loc);
#else
    sys = win32_get_codeset(ctype);
#endif
 
    if (!sys)
        return -1;              /* out of memory; unlikely */
 
    /* Check the table */
    for (i = 0; encoding_match_list[i].system_enc_name; i++)
    {
        if (pg_strcasecmp(sys, encoding_match_list[i].system_enc_name) == 0)
        {
            free(sys);
            return encoding_match_list[i].pg_enc_code;
        }
    }
 
    /* Special-case kluges for particular platforms go here */
 
#ifdef __darwin__
 
    /*
     * Current macOS has many locales that report an empty string for CODESET,
     * but they all seem to actually use UTF-8.
     */
    if (strlen(sys) == 0)
    {
        free(sys);
        return PG_UTF8;
    }
#endif
 
    /*
     * We print a warning if we got a CODESET string but couldn't recognize
     * it.  This means we need another entry in the table.
     */
    if (write_message)
    {
#ifdef FRONTEND
        fprintf(stderr, _("could not determine encoding for locale \"%s\": codeset is \"%s\""),
                ctype, sys);
        /* keep newline separate so there's only one translatable string */
        fputc('\n', stderr);
#else
        ereport(WARNING,
                (errmsg("could not determine encoding for locale \"%s\": codeset is \"%s\"",
                        ctype, sys)));
#endif
    }
 
    free(sys);
    return -1;
}

References _, encoding_match_list, ereport, errmsg(), fb(), fprintf, free, i, locale_t, encoding_match::pg_enc_code, PG_SQL_ASCII, pg_strcasecmp(), PG_UTF8, setlocale, encoding_match::system_enc_name, and WARNING.

Referenced by cache_locale_time(), check_encoding_locale_matches(), check_locale_encoding(), main(), PGLC_localeconv(), pqConnectOptions2(), PQsetClientEncoding(), and setup_locale_encoding().

Variable Documentation

encoding_match_list

const struct encoding_match encoding_match_list[]

static

Definition at line 45 of file chklocale.c.

                                                           {
    {PG_EUC_JP, "EUC-JP"},
    {PG_EUC_JP, "eucJP"},
    {PG_EUC_JP, "IBM-eucJP"},
    {PG_EUC_JP, "sdeckanji"},
    {PG_EUC_JP, "CP20932"},
 
    {PG_EUC_CN, "EUC-CN"},
    {PG_EUC_CN, "eucCN"},
    {PG_EUC_CN, "IBM-eucCN"},
    {PG_EUC_CN, "GB2312"},
    {PG_EUC_CN, "dechanzi"},
    {PG_EUC_CN, "CP20936"},
 
    {PG_EUC_KR, "EUC-KR"},
    {PG_EUC_KR, "eucKR"},
    {PG_EUC_KR, "IBM-eucKR"},
    {PG_EUC_KR, "deckorean"},
    {PG_EUC_KR, "5601"},
    {PG_EUC_KR, "CP51949"},
 
    {PG_EUC_TW, "EUC-TW"},
    {PG_EUC_TW, "eucTW"},
    {PG_EUC_TW, "IBM-eucTW"},
    {PG_EUC_TW, "cns11643"},
    /* No codepage for EUC-TW ? */
 
    {PG_UTF8, "UTF-8"},
    {PG_UTF8, "utf8"},
    {PG_UTF8, "CP65001"},
 
    {PG_LATIN1, "ISO-8859-1"},
    {PG_LATIN1, "ISO8859-1"},
    {PG_LATIN1, "iso88591"},
    {PG_LATIN1, "CP28591"},
 
    {PG_LATIN2, "ISO-8859-2"},
    {PG_LATIN2, "ISO8859-2"},
    {PG_LATIN2, "iso88592"},
    {PG_LATIN2, "CP28592"},
 
    {PG_LATIN3, "ISO-8859-3"},
    {PG_LATIN3, "ISO8859-3"},
    {PG_LATIN3, "iso88593"},
    {PG_LATIN3, "CP28593"},
 
    {PG_LATIN4, "ISO-8859-4"},
    {PG_LATIN4, "ISO8859-4"},
    {PG_LATIN4, "iso88594"},
    {PG_LATIN4, "CP28594"},
 
    {PG_LATIN5, "ISO-8859-9"},
    {PG_LATIN5, "ISO8859-9"},
    {PG_LATIN5, "iso88599"},
    {PG_LATIN5, "CP28599"},
 
    {PG_LATIN6, "ISO-8859-10"},
    {PG_LATIN6, "ISO8859-10"},
    {PG_LATIN6, "iso885910"},
 
    {PG_LATIN7, "ISO-8859-13"},
    {PG_LATIN7, "ISO8859-13"},
    {PG_LATIN7, "iso885913"},
 
    {PG_LATIN8, "ISO-8859-14"},
    {PG_LATIN8, "ISO8859-14"},
    {PG_LATIN8, "iso885914"},
 
    {PG_LATIN9, "ISO-8859-15"},
    {PG_LATIN9, "ISO8859-15"},
    {PG_LATIN9, "iso885915"},
    {PG_LATIN9, "CP28605"},
 
    {PG_LATIN10, "ISO-8859-16"},
    {PG_LATIN10, "ISO8859-16"},
    {PG_LATIN10, "iso885916"},
 
    {PG_KOI8R, "KOI8-R"},
    {PG_KOI8R, "CP20866"},
 
    {PG_KOI8U, "KOI8-U"},
    {PG_KOI8U, "CP21866"},
 
    {PG_WIN866, "CP866"},
    {PG_WIN874, "CP874"},
    {PG_WIN1250, "CP1250"},
    {PG_WIN1251, "CP1251"},
    {PG_WIN1251, "ansi-1251"},
    {PG_WIN1252, "CP1252"},
    {PG_WIN1253, "CP1253"},
    {PG_WIN1254, "CP1254"},
    {PG_WIN1255, "CP1255"},
    {PG_WIN1256, "CP1256"},
    {PG_WIN1257, "CP1257"},
    {PG_WIN1258, "CP1258"},
 
    {PG_ISO_8859_5, "ISO-8859-5"},
    {PG_ISO_8859_5, "ISO8859-5"},
    {PG_ISO_8859_5, "iso88595"},
    {PG_ISO_8859_5, "CP28595"},
 
    {PG_ISO_8859_6, "ISO-8859-6"},
    {PG_ISO_8859_6, "ISO8859-6"},
    {PG_ISO_8859_6, "iso88596"},
    {PG_ISO_8859_6, "CP28596"},
 
    {PG_ISO_8859_7, "ISO-8859-7"},
    {PG_ISO_8859_7, "ISO8859-7"},
    {PG_ISO_8859_7, "iso88597"},
    {PG_ISO_8859_7, "CP28597"},
 
    {PG_ISO_8859_8, "ISO-8859-8"},
    {PG_ISO_8859_8, "ISO8859-8"},
    {PG_ISO_8859_8, "iso88598"},
    {PG_ISO_8859_8, "CP28598"},
 
    {PG_SJIS, "SJIS"},
    {PG_SJIS, "PCK"},
    {PG_SJIS, "CP932"},
    {PG_SJIS, "SHIFT_JIS"},
 
    {PG_BIG5, "BIG5"},
    {PG_BIG5, "BIG5HKSCS"},
    {PG_BIG5, "Big5-HKSCS"},
    {PG_BIG5, "CP950"},
 
    {PG_GBK, "GBK"},
    {PG_GBK, "CP936"},
 
    {PG_UHC, "UHC"},
    {PG_UHC, "CP949"},
 
    {PG_JOHAB, "JOHAB"},
    {PG_JOHAB, "CP1361"},
 
    {PG_GB18030, "GB18030"},
    {PG_GB18030, "CP54936"},
 
    {PG_SHIFT_JIS_2004, "SJIS_2004"},
 
    {PG_SQL_ASCII, "US-ASCII"},
 
    {PG_SQL_ASCII, NULL}        /* end marker */
};

Referenced by pg_get_encoding_from_locale().