[perl5.git] / locale.c

/*    locale.c
 *
 *    Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999,
 *    2000, 2001, 2002, 2003, 2005, 2006, by Larry Wall and others
 *
 *    You may distribute under the terms of either the GNU General Public
 *    License or the Artistic License, as specified in the README file.
 *
 */

/*
 * A Elbereth Gilthoniel,
 * silivren penna míriel
 * o menel aglar elenath!
 * Na-chaered palan-díriel
 * o galadhremmin ennorath,
 * Fanuilos, le linnathon
 * nef aear, si nef aearon!
 */

/* utility functions for handling locale-specific stuff like what
 * character represents the decimal point.
 */

#include "EXTERN.h"
#define PERL_IN_LOCALE_C
#include "perl.h"

#ifdef I_LOCALE
#  include <locale.h>
#endif

#ifdef I_LANGINFO
#   include <langinfo.h>
#endif

#include "reentr.h"

#if defined(USE_LOCALE_NUMERIC) || defined(USE_LOCALE_COLLATE)
/*
 * Standardize the locale name from a string returned by 'setlocale'.
 *
 * The standard return value of setlocale() is either
 * (1) "xx_YY" if the first argument of setlocale() is not LC_ALL
 * (2) "xa_YY xb_YY ..." if the first argument of setlocale() is LC_ALL
 *     (the space-separated values represent the various sublocales,
 *      in some unspecificed order)
 *
 * In some platforms it has a form like "LC_SOMETHING=Lang_Country.866\n",
 * which is harmful for further use of the string in setlocale().
 *
 */
STATIC char *
S_stdize_locale(pTHX_ char *locs)
{
    const char * const s = strchr(locs, '=');
    bool okay = TRUE;

    if (s) {
	const char * const t = strchr(s, '.');
	okay = FALSE;
	if (t) {
	    const char * const u = strchr(t, '\n');
	    if (u && (u[1] == 0)) {
		const STRLEN len = u - s;
		Move(s + 1, locs, len, char);
		locs[len] = 0;
		okay = TRUE;
	    }
	}
    }

    if (!okay)
	Perl_croak(aTHX_ "Can't fix broken locale name \"%s\"", locs);

    return locs;
}
#endif

void
Perl_set_numeric_radix(pTHX)
{
#ifdef USE_LOCALE_NUMERIC
    dVAR;
# ifdef HAS_LOCALECONV
    const struct lconv* const lc = localeconv();

    if (lc && lc->decimal_point) {
	if (lc->decimal_point[0] == '.' && lc->decimal_point[1] == 0) {
	    SvREFCNT_dec(PL_numeric_radix_sv);
	    PL_numeric_radix_sv = NULL;
	}
	else {
	    if (PL_numeric_radix_sv)
		sv_setpv(PL_numeric_radix_sv, lc->decimal_point);
	    else
		PL_numeric_radix_sv = newSVpv(lc->decimal_point, 0);
	}
    }
    else
	PL_numeric_radix_sv = NULL;
# endif /* HAS_LOCALECONV */
#endif /* USE_LOCALE_NUMERIC */
}

/*
 * Set up for a new numeric locale.
 */
void
Perl_new_numeric(pTHX_ const char *newnum)
{
#ifdef USE_LOCALE_NUMERIC
    dVAR;

    if (! newnum) {
	Safefree(PL_numeric_name);
	PL_numeric_name = NULL;
	PL_numeric_standard = TRUE;
	PL_numeric_local = TRUE;
	return;
    }

    if (! PL_numeric_name || strNE(PL_numeric_name, newnum)) {
	Safefree(PL_numeric_name);
	PL_numeric_name = stdize_locale(savepv(newnum));
	PL_numeric_standard = ((*newnum == 'C' && newnum[1] == '\0')
			       || strEQ(newnum, "POSIX"));
	PL_numeric_local = TRUE;
	set_numeric_radix();
    }

#endif /* USE_LOCALE_NUMERIC */
}

void
Perl_set_numeric_standard(pTHX)
{
#ifdef USE_LOCALE_NUMERIC
    dVAR;

    if (! PL_numeric_standard) {
	setlocale(LC_NUMERIC, "C");
	PL_numeric_standard = TRUE;
	PL_numeric_local = FALSE;
	set_numeric_radix();
    }

#endif /* USE_LOCALE_NUMERIC */
}

void
Perl_set_numeric_local(pTHX)
{
#ifdef USE_LOCALE_NUMERIC
    dVAR;

    if (! PL_numeric_local) {
	setlocale(LC_NUMERIC, PL_numeric_name);
	PL_numeric_standard = FALSE;
	PL_numeric_local = TRUE;
	set_numeric_radix();
    }

#endif /* USE_LOCALE_NUMERIC */
}

/*
 * Set up for a new ctype locale.
 */
void
Perl_new_ctype(pTHX_ const char *newctype)
{
#ifdef USE_LOCALE_CTYPE
    dVAR;
    int i;

    for (i = 0; i < 256; i++) {
	if (isUPPER_LC(i))
	    PL_fold_locale[i] = toLOWER_LC(i);
	else if (isLOWER_LC(i))
	    PL_fold_locale[i] = toUPPER_LC(i);
	else
	    PL_fold_locale[i] = i;
    }

#endif /* USE_LOCALE_CTYPE */
    PERL_UNUSED_ARG(newctype);
}

/*
 * Set up for a new collation locale.
 */
void
Perl_new_collate(pTHX_ const char *newcoll)
{
#ifdef USE_LOCALE_COLLATE
    dVAR;

    if (! newcoll) {
	if (PL_collation_name) {
	    ++PL_collation_ix;
	    Safefree(PL_collation_name);
	    PL_collation_name = NULL;
	}
	PL_collation_standard = TRUE;
	PL_collxfrm_base = 0;
	PL_collxfrm_mult = 2;
	return;
    }

    if (! PL_collation_name || strNE(PL_collation_name, newcoll)) {
	++PL_collation_ix;
	Safefree(PL_collation_name);
	PL_collation_name = stdize_locale(savepv(newcoll));
	PL_collation_standard = ((*newcoll == 'C' && newcoll[1] == '\0')
				 || strEQ(newcoll, "POSIX"));

	{
	  /*  2: at most so many chars ('a', 'b'). */
	  /* 50: surely no system expands a char more. */
#define XFRMBUFSIZE  (2 * 50)
	  char xbuf[XFRMBUFSIZE];
	  const Size_t fa = strxfrm(xbuf, "a",  XFRMBUFSIZE);
	  const Size_t fb = strxfrm(xbuf, "ab", XFRMBUFSIZE);
	  const SSize_t mult = fb - fa;
	  if (mult < 1)
	      Perl_croak(aTHX_ "strxfrm() gets absurd");
	  PL_collxfrm_base = (fa > (Size_t)mult) ? (fa - mult) : 0;
	  PL_collxfrm_mult = mult;
	}
    }

#endif /* USE_LOCALE_COLLATE */
}

/*
 * Initialize locale awareness.
 */
int
Perl_init_i18nl10n(pTHX_ int printwarn)
{
    int ok = 1;
    /* returns
     *    1 = set ok or not applicable,
     *    0 = fallback to C locale,
     *   -1 = fallback to C locale failed
     */

#if defined(USE_LOCALE)
    dVAR;

#ifdef USE_LOCALE_CTYPE
    char *curctype   = NULL;
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
    char *curcoll    = NULL;
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
    char *curnum     = NULL;
#endif /* USE_LOCALE_NUMERIC */
#ifdef __GLIBC__
    char * const language   = PerlEnv_getenv("LANGUAGE");
#endif
    char * const lc_all     = PerlEnv_getenv("LC_ALL");
    char * const lang       = PerlEnv_getenv("LANG");
    bool setlocale_failure = FALSE;

#ifdef LOCALE_ENVIRON_REQUIRED

    /*
     * Ultrix setlocale(..., "") fails if there are no environment
     * variables from which to get a locale name.
     */

    bool done = FALSE;

#ifdef LC_ALL
    if (lang) {
	if (setlocale(LC_ALL, ""))
	    done = TRUE;
	else
	    setlocale_failure = TRUE;
    }
    if (!setlocale_failure) {
#ifdef USE_LOCALE_CTYPE
	if (! (curctype =
	       setlocale(LC_CTYPE,
			 (!done && (lang || PerlEnv_getenv("LC_CTYPE")))
				    ? "" : NULL)))
	    setlocale_failure = TRUE;
	else
	    curctype = savepv(curctype);
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	if (! (curcoll =
	       setlocale(LC_COLLATE,
			 (!done && (lang || PerlEnv_getenv("LC_COLLATE")))
				   ? "" : NULL)))
	    setlocale_failure = TRUE;
	else
	    curcoll = savepv(curcoll);
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	if (! (curnum =
	       setlocale(LC_NUMERIC,
			 (!done && (lang || PerlEnv_getenv("LC_NUMERIC")))
				  ? "" : NULL)))
	    setlocale_failure = TRUE;
	else
	    curnum = savepv(curnum);
#endif /* USE_LOCALE_NUMERIC */
    }

#endif /* LC_ALL */

#endif /* !LOCALE_ENVIRON_REQUIRED */

#ifdef LC_ALL
    if (! setlocale(LC_ALL, ""))
	setlocale_failure = TRUE;
#endif /* LC_ALL */

    if (!setlocale_failure) {
#ifdef USE_LOCALE_CTYPE
	if (! (curctype = setlocale(LC_CTYPE, "")))
	    setlocale_failure = TRUE;
	else
	    curctype = savepv(curctype);
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	if (! (curcoll = setlocale(LC_COLLATE, "")))
	    setlocale_failure = TRUE;
	else
	    curcoll = savepv(curcoll);
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	if (! (curnum = setlocale(LC_NUMERIC, "")))
	    setlocale_failure = TRUE;
	else
	    curnum = savepv(curnum);
#endif /* USE_LOCALE_NUMERIC */
    }

    if (setlocale_failure) {
	char *p;
	const bool locwarn = (printwarn > 1 ||
			(printwarn &&
			 (!(p = PerlEnv_getenv("PERL_BADLANG")) || atoi(p))));

	if (locwarn) {
#ifdef LC_ALL

	    PerlIO_printf(Perl_error_log,
	       "perl: warning: Setting locale failed.\n");

#else /* !LC_ALL */

	    PerlIO_printf(Perl_error_log,
	       "perl: warning: Setting locale failed for the categories:\n\t");
#ifdef USE_LOCALE_CTYPE
	    if (! curctype)
		PerlIO_printf(Perl_error_log, "LC_CTYPE ");
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	    if (! curcoll)
		PerlIO_printf(Perl_error_log, "LC_COLLATE ");
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	    if (! curnum)
		PerlIO_printf(Perl_error_log, "LC_NUMERIC ");
#endif /* USE_LOCALE_NUMERIC */
	    PerlIO_printf(Perl_error_log, "\n");

#endif /* LC_ALL */

	    PerlIO_printf(Perl_error_log,
		"perl: warning: Please check that your locale settings:\n");

#ifdef __GLIBC__
	    PerlIO_printf(Perl_error_log,
			  "\tLANGUAGE = %c%s%c,\n",
			  language ? '"' : '(',
			  language ? language : "unset",
			  language ? '"' : ')');
#endif

	    PerlIO_printf(Perl_error_log,
			  "\tLC_ALL = %c%s%c,\n",
			  lc_all ? '"' : '(',
			  lc_all ? lc_all : "unset",
			  lc_all ? '"' : ')');

#if defined(USE_ENVIRON_ARRAY)
	    {
	      char **e;
	      for (e = environ; *e; e++) {
		  if (strnEQ(*e, "LC_", 3)
			&& strnNE(*e, "LC_ALL=", 7)
			&& (p = strchr(*e, '=')))
		      PerlIO_printf(Perl_error_log, "\t%.*s = \"%s\",\n",
				    (int)(p - *e), *e, p + 1);
	      }
	    }
#else
	    PerlIO_printf(Perl_error_log,
			  "\t(possibly more locale environment variables)\n");
#endif

	    PerlIO_printf(Perl_error_log,
			  "\tLANG = %c%s%c\n",
			  lang ? '"' : '(',
			  lang ? lang : "unset",
			  lang ? '"' : ')');

	    PerlIO_printf(Perl_error_log,
			  "    are supported and installed on your system.\n");
	}

#ifdef LC_ALL

	if (setlocale(LC_ALL, "C")) {
	    if (locwarn)
		PerlIO_printf(Perl_error_log,
      "perl: warning: Falling back to the standard locale (\"C\").\n");
	    ok = 0;
	}
	else {
	    if (locwarn)
		PerlIO_printf(Perl_error_log,
      "perl: warning: Failed to fall back to the standard locale (\"C\").\n");
	    ok = -1;
	}

#else /* ! LC_ALL */

	if (0
#ifdef USE_LOCALE_CTYPE
	    || !(curctype || setlocale(LC_CTYPE, "C"))
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	    || !(curcoll || setlocale(LC_COLLATE, "C"))
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	    || !(curnum || setlocale(LC_NUMERIC, "C"))
#endif /* USE_LOCALE_NUMERIC */
	    )
	{
	    if (locwarn)
		PerlIO_printf(Perl_error_log,
      "perl: warning: Cannot fall back to the standard locale (\"C\").\n");
	    ok = -1;
	}

#endif /* ! LC_ALL */

#ifdef USE_LOCALE_CTYPE
	curctype = savepv(setlocale(LC_CTYPE, NULL));
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	curcoll = savepv(setlocale(LC_COLLATE, NULL));
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	curnum = savepv(setlocale(LC_NUMERIC, NULL));
#endif /* USE_LOCALE_NUMERIC */
    }
    else {

#ifdef USE_LOCALE_CTYPE
    new_ctype(curctype);
#endif /* USE_LOCALE_CTYPE */

#ifdef USE_LOCALE_COLLATE
    new_collate(curcoll);
#endif /* USE_LOCALE_COLLATE */

#ifdef USE_LOCALE_NUMERIC
    new_numeric(curnum);
#endif /* USE_LOCALE_NUMERIC */

    }

#endif /* USE_LOCALE */

#ifdef USE_PERLIO
    {
      /* Set PL_utf8locale to TRUE if using PerlIO _and_
	 any of the following are true:
	 - nl_langinfo(CODESET) contains /^utf-?8/i
	 - $ENV{LC_ALL}   contains /^utf-?8/i
	 - $ENV{LC_CTYPE} contains /^utf-?8/i
	 - $ENV{LANG}     contains /^utf-?8/i
	 The LC_ALL, LC_CTYPE, LANG obey the usual override
	 hierarchy of locale environment variables.  (LANGUAGE
	 affects only LC_MESSAGES only under glibc.) (If present,
	 it overrides LC_MESSAGES for GNU gettext, and it also
	 can have more than one locale, separated by spaces,
	 in case you need to know.)
	 If PL_utf8locale and PL_unicode (set by -C or by $ENV{PERL_UNICODE})
         are true, perl.c:S_parse_body() will turn on the PerlIO :utf8 layer
	 on STDIN, STDOUT, STDERR, _and_ the default open discipline.
      */
	 bool utf8locale = FALSE;
	 char *codeset = NULL;
#if defined(HAS_NL_LANGINFO) && defined(CODESET)
	 codeset = nl_langinfo(CODESET);
#endif
	 if (codeset)
	      utf8locale = (Perl_ibcmp(aTHX_ codeset, STR_WITH_LEN("UTF-8")) == 0 ||
 			    Perl_ibcmp(aTHX_ codeset, STR_WITH_LEN("UTF8") ) == 0);
#if defined(USE_LOCALE)
	 else { /* nl_langinfo(CODESET) is supposed to correctly
		 * interpret the locale environment variables,
		 * but just in case it fails, let's do this manually. */ 
	      if (lang)
		   utf8locale = (Perl_ibcmp(aTHX_ lang, STR_WITH_LEN("UTF-8")) == 0 ||
			         Perl_ibcmp(aTHX_ lang, STR_WITH_LEN("UTF8") ) == 0);
#ifdef USE_LOCALE_CTYPE
	      if (curctype)
		   utf8locale = (Perl_ibcmp(aTHX_ curctype, STR_WITH_LEN("UTF-8")) == 0 ||
			         Perl_ibcmp(aTHX_ curctype, STR_WITH_LEN("UTF8") ) == 0);
#endif
	      if (lc_all)
		   utf8locale = (Perl_ibcmp(aTHX_ lc_all, STR_WITH_LEN("UTF-8")) == 0 ||
			         Perl_ibcmp(aTHX_ lc_all, STR_WITH_LEN("UTF8") ) == 0);
	 }
#endif /* USE_LOCALE */
	 if (utf8locale)
	      PL_utf8locale = TRUE;
    }
    /* Set PL_unicode to $ENV{PERL_UNICODE} if using PerlIO.
       This is an alternative to using the -C command line switch
       (the -C if present will override this). */
    {
	 const char *p = PerlEnv_getenv("PERL_UNICODE");
	 PL_unicode = p ? parse_unicode_opts(&p) : 0;
    }
#endif

#ifdef USE_LOCALE_CTYPE
    Safefree(curctype);
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
    Safefree(curcoll);
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
    Safefree(curnum);
#endif /* USE_LOCALE_NUMERIC */
    return ok;
}

#ifdef USE_LOCALE_COLLATE

/*
 * mem_collxfrm() is a bit like strxfrm() but with two important
 * differences. First, it handles embedded NULs. Second, it allocates
 * a bit more memory than needed for the transformed data itself.
 * The real transformed data begins at offset sizeof(collationix).
 * Please see sv_collxfrm() to see how this is used.
 */

char *
Perl_mem_collxfrm(pTHX_ const char *s, STRLEN len, STRLEN *xlen)
{
    dVAR;
    char *xbuf;
    STRLEN xAlloc, xin, xout; /* xalloc is a reserved word in VC */

    /* the first sizeof(collationix) bytes are used by sv_collxfrm(). */
    /* the +1 is for the terminating NUL. */

    xAlloc = sizeof(PL_collation_ix) + PL_collxfrm_base + (PL_collxfrm_mult * len) + 1;
    Newx(xbuf, xAlloc, char);
    if (! xbuf)
	goto bad;

    *(U32*)xbuf = PL_collation_ix;
    xout = sizeof(PL_collation_ix);
    for (xin = 0; xin < len; ) {
	SSize_t xused;

	for (;;) {
	    xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
	    if (xused == -1)
		goto bad;
	    if ((STRLEN)xused < xAlloc - xout)
		break;
	    xAlloc = (2 * xAlloc) + 1;
	    Renew(xbuf, xAlloc, char);
	    if (! xbuf)
		goto bad;
	}

	xin += strlen(s + xin) + 1;
	xout += xused;

	/* Embedded NULs are understood but silently skipped
	 * because they make no sense in locale collation. */
    }

    xbuf[xout] = '\0';
    *xlen = xout - sizeof(PL_collation_ix);
    return xbuf;

  bad:
    Safefree(xbuf);
    *xlen = 0;
    return NULL;
}

#endif /* USE_LOCALE_COLLATE */

/*
 * Local variables:
 * c-indentation-style: bsd
 * c-basic-offset: 4
 * indent-tabs-mode: t
 * End:
 *
 * ex: set ts=8 sts=4 sw=4 noet:
 */
Commit	Line	Data
98994639 HS	1	/* locale.c
98994639 HS	2	*
4bb101f2	3	* Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999,
acde74e1	4	* 2000, 2001, 2002, 2003, 2005, 2006, by Larry Wall and others
98994639 HS	5	*
	6	* You may distribute under the terms of either the GNU General Public
	7	* License or the Artistic License, as specified in the README file.
	8	*
	9	*/
	10
	11	/*
	12	* A Elbereth Gilthoniel,
	13	* silivren penna míriel
	14	* o menel aglar elenath!
	15	* Na-chaered palan-díriel
	16	* o galadhremmin ennorath,
	17	* Fanuilos, le linnathon
	18	* nef aear, si nef aearon!
	19	*/
	20
166f8a29 DM	21	/* utility functions for handling locale-specific stuff like what
	22	* character represents the decimal point.
	23	*/
	24
98994639 HS	25	#include "EXTERN.h"
	26	#define PERL_IN_LOCALE_C
	27	#include "perl.h"
	28
	29	#ifdef I_LOCALE
	30	# include <locale.h>
	31	#endif
	32
b310b053 JH	33	#ifdef I_LANGINFO
	34	# include <langinfo.h>
	35	#endif
	36
a4af207c JH	37	#include "reentr.h"
a4af207c JH	38
27da23d5	39	#if defined(USE_LOCALE_NUMERIC) \|\| defined(USE_LOCALE_COLLATE)
98994639 HS	40	/*
	41	* Standardize the locale name from a string returned by 'setlocale'.
	42	*
	43	* The standard return value of setlocale() is either
	44	* (1) "xx_YY" if the first argument of setlocale() is not LC_ALL
	45	* (2) "xa_YY xb_YY ..." if the first argument of setlocale() is LC_ALL
	46	* (the space-separated values represent the various sublocales,
	47	* in some unspecificed order)
	48	*
	49	* In some platforms it has a form like "LC_SOMETHING=Lang_Country.866\n",
	50	* which is harmful for further use of the string in setlocale().
	51	*
	52	*/
	53	STATIC char *
	54	S_stdize_locale(pTHX_ char *locs)
	55	{
7452cf6a	56	const char * const s = strchr(locs, '=');
98994639 HS	57	bool okay = TRUE;
98994639 HS	58
8772537c AL	59	if (s) {
8772537c AL	60	const char * const t = strchr(s, '.');
98994639	61	okay = FALSE;
8772537c AL	62	if (t) {
	63	const char * const u = strchr(t, '\n');
	64	if (u && (u[1] == 0)) {
	65	const STRLEN len = u - s;
	66	Move(s + 1, locs, len, char);
	67	locs[len] = 0;
	68	okay = TRUE;
98994639 HS	69	}
	70	}
	71	}
	72
	73	if (!okay)
	74	Perl_croak(aTHX_ "Can't fix broken locale name \"%s\"", locs);
	75
	76	return locs;
	77	}
27da23d5	78	#endif
98994639 HS	79
	80	void
	81	Perl_set_numeric_radix(pTHX)
	82	{
	83	#ifdef USE_LOCALE_NUMERIC
97aff369	84	dVAR;
98994639	85	# ifdef HAS_LOCALECONV
7452cf6a	86	const struct lconv* const lc = localeconv();
98994639	87
98994639 HS	88	if (lc && lc->decimal_point) {
	89	if (lc->decimal_point[0] == '.' && lc->decimal_point[1] == 0) {
	90	SvREFCNT_dec(PL_numeric_radix_sv);
a0714e2c	91	PL_numeric_radix_sv = NULL;
98994639 HS	92	}
	93	else {
	94	if (PL_numeric_radix_sv)
	95	sv_setpv(PL_numeric_radix_sv, lc->decimal_point);
	96	else
	97	PL_numeric_radix_sv = newSVpv(lc->decimal_point, 0);
	98	}
	99	}
	100	else
a0714e2c	101	PL_numeric_radix_sv = NULL;
98994639 HS	102	# endif /* HAS_LOCALECONV */
	103	#endif /* USE_LOCALE_NUMERIC */
	104	}
	105
	106	/*
	107	* Set up for a new numeric locale.
	108	*/
	109	void
8772537c	110	Perl_new_numeric(pTHX_ const char *newnum)
98994639 HS	111	{
98994639 HS	112	#ifdef USE_LOCALE_NUMERIC
97aff369	113	dVAR;
98994639 HS	114
98994639 HS	115	if (! newnum) {
43c5f42d NC	116	Safefree(PL_numeric_name);
43c5f42d NC	117	PL_numeric_name = NULL;
98994639 HS	118	PL_numeric_standard = TRUE;
	119	PL_numeric_local = TRUE;
	120	return;
	121	}
	122
	123	if (! PL_numeric_name \|\| strNE(PL_numeric_name, newnum)) {
	124	Safefree(PL_numeric_name);
	125	PL_numeric_name = stdize_locale(savepv(newnum));
770526c1 NC	126	PL_numeric_standard = ((*newnum == 'C' && newnum[1] == '\0')
770526c1 NC	127	\|\| strEQ(newnum, "POSIX"));
98994639 HS	128	PL_numeric_local = TRUE;
	129	set_numeric_radix();
	130	}
	131
	132	#endif /* USE_LOCALE_NUMERIC */
	133	}
	134
	135	void
	136	Perl_set_numeric_standard(pTHX)
	137	{
	138	#ifdef USE_LOCALE_NUMERIC
97aff369	139	dVAR;
98994639 HS	140
	141	if (! PL_numeric_standard) {
	142	setlocale(LC_NUMERIC, "C");
	143	PL_numeric_standard = TRUE;
	144	PL_numeric_local = FALSE;
	145	set_numeric_radix();
	146	}
	147
	148	#endif /* USE_LOCALE_NUMERIC */
	149	}
	150
	151	void
	152	Perl_set_numeric_local(pTHX)
	153	{
	154	#ifdef USE_LOCALE_NUMERIC
97aff369	155	dVAR;
98994639 HS	156
	157	if (! PL_numeric_local) {
	158	setlocale(LC_NUMERIC, PL_numeric_name);
	159	PL_numeric_standard = FALSE;
	160	PL_numeric_local = TRUE;
	161	set_numeric_radix();
	162	}
	163
	164	#endif /* USE_LOCALE_NUMERIC */
	165	}
	166
	167	/*
	168	* Set up for a new ctype locale.
	169	*/
	170	void
8772537c	171	Perl_new_ctype(pTHX_ const char *newctype)
98994639 HS	172	{
98994639 HS	173	#ifdef USE_LOCALE_CTYPE
27da23d5	174	dVAR;
98994639 HS	175	int i;
	176
	177	for (i = 0; i < 256; i++) {
	178	if (isUPPER_LC(i))
	179	PL_fold_locale[i] = toLOWER_LC(i);
	180	else if (isLOWER_LC(i))
	181	PL_fold_locale[i] = toUPPER_LC(i);
	182	else
	183	PL_fold_locale[i] = i;
	184	}
	185
	186	#endif /* USE_LOCALE_CTYPE */
8772537c	187	PERL_UNUSED_ARG(newctype);
98994639 HS	188	}
	189
	190	/*
	191	* Set up for a new collation locale.
	192	*/
	193	void
8772537c	194	Perl_new_collate(pTHX_ const char *newcoll)
98994639 HS	195	{
98994639 HS	196	#ifdef USE_LOCALE_COLLATE
97aff369	197	dVAR;
98994639 HS	198
	199	if (! newcoll) {
	200	if (PL_collation_name) {
	201	++PL_collation_ix;
	202	Safefree(PL_collation_name);
	203	PL_collation_name = NULL;
	204	}
	205	PL_collation_standard = TRUE;
	206	PL_collxfrm_base = 0;
	207	PL_collxfrm_mult = 2;
	208	return;
	209	}
	210
	211	if (! PL_collation_name \|\| strNE(PL_collation_name, newcoll)) {
	212	++PL_collation_ix;
	213	Safefree(PL_collation_name);
	214	PL_collation_name = stdize_locale(savepv(newcoll));
770526c1 NC	215	PL_collation_standard = ((*newcoll == 'C' && newcoll[1] == '\0')
770526c1 NC	216	\|\| strEQ(newcoll, "POSIX"));
98994639 HS	217
	218	{
	219	/* 2: at most so many chars ('a', 'b'). */
	220	/* 50: surely no system expands a char more. */
	221	#define XFRMBUFSIZE (2 * 50)
	222	char xbuf[XFRMBUFSIZE];
8772537c AL	223	const Size_t fa = strxfrm(xbuf, "a", XFRMBUFSIZE);
	224	const Size_t fb = strxfrm(xbuf, "ab", XFRMBUFSIZE);
	225	const SSize_t mult = fb - fa;
98994639 HS	226	if (mult < 1)
98994639 HS	227	Perl_croak(aTHX_ "strxfrm() gets absurd");
eb160463	228	PL_collxfrm_base = (fa > (Size_t)mult) ? (fa - mult) : 0;
98994639 HS	229	PL_collxfrm_mult = mult;
	230	}
	231	}
	232
	233	#endif /* USE_LOCALE_COLLATE */
	234	}
	235
	236	/*
	237	* Initialize locale awareness.
	238	*/
	239	int
	240	Perl_init_i18nl10n(pTHX_ int printwarn)
	241	{
	242	int ok = 1;
	243	/* returns
	244	* 1 = set ok or not applicable,
	245	* 0 = fallback to C locale,
	246	* -1 = fallback to C locale failed
	247	*/
	248
	249	#if defined(USE_LOCALE)
97aff369	250	dVAR;
98994639 HS	251
	252	#ifdef USE_LOCALE_CTYPE
	253	char *curctype = NULL;
	254	#endif /* USE_LOCALE_CTYPE */
	255	#ifdef USE_LOCALE_COLLATE
	256	char *curcoll = NULL;
	257	#endif /* USE_LOCALE_COLLATE */
	258	#ifdef USE_LOCALE_NUMERIC
	259	char *curnum = NULL;
	260	#endif /* USE_LOCALE_NUMERIC */
	261	#ifdef __GLIBC__
7452cf6a	262	char * const language = PerlEnv_getenv("LANGUAGE");
98994639	263	#endif
7452cf6a AL	264	char * const lc_all = PerlEnv_getenv("LC_ALL");
7452cf6a AL	265	char * const lang = PerlEnv_getenv("LANG");
98994639 HS	266	bool setlocale_failure = FALSE;
	267
	268	#ifdef LOCALE_ENVIRON_REQUIRED
	269
	270	/*
	271	* Ultrix setlocale(..., "") fails if there are no environment
	272	* variables from which to get a locale name.
	273	*/
	274
	275	bool done = FALSE;
	276
	277	#ifdef LC_ALL
	278	if (lang) {
	279	if (setlocale(LC_ALL, ""))
	280	done = TRUE;
	281	else
	282	setlocale_failure = TRUE;
	283	}
	284	if (!setlocale_failure) {
	285	#ifdef USE_LOCALE_CTYPE
	286	if (! (curctype =
	287	setlocale(LC_CTYPE,
	288	(!done && (lang \|\| PerlEnv_getenv("LC_CTYPE")))
bd61b366	289	? "" : NULL)))
98994639 HS	290	setlocale_failure = TRUE;
	291	else
	292	curctype = savepv(curctype);
	293	#endif /* USE_LOCALE_CTYPE */
	294	#ifdef USE_LOCALE_COLLATE
	295	if (! (curcoll =
	296	setlocale(LC_COLLATE,
	297	(!done && (lang \|\| PerlEnv_getenv("LC_COLLATE")))
bd61b366	298	? "" : NULL)))
98994639 HS	299	setlocale_failure = TRUE;
	300	else
	301	curcoll = savepv(curcoll);
	302	#endif /* USE_LOCALE_COLLATE */
	303	#ifdef USE_LOCALE_NUMERIC
	304	if (! (curnum =
	305	setlocale(LC_NUMERIC,
	306	(!done && (lang \|\| PerlEnv_getenv("LC_NUMERIC")))
bd61b366	307	? "" : NULL)))
98994639 HS	308	setlocale_failure = TRUE;
	309	else
	310	curnum = savepv(curnum);
	311	#endif /* USE_LOCALE_NUMERIC */
	312	}
	313
	314	#endif /* LC_ALL */
	315
	316	#endif /* !LOCALE_ENVIRON_REQUIRED */
	317
	318	#ifdef LC_ALL
	319	if (! setlocale(LC_ALL, ""))
	320	setlocale_failure = TRUE;
	321	#endif /* LC_ALL */
	322
	323	if (!setlocale_failure) {
	324	#ifdef USE_LOCALE_CTYPE
	325	if (! (curctype = setlocale(LC_CTYPE, "")))
	326	setlocale_failure = TRUE;
	327	else
	328	curctype = savepv(curctype);
	329	#endif /* USE_LOCALE_CTYPE */
	330	#ifdef USE_LOCALE_COLLATE
	331	if (! (curcoll = setlocale(LC_COLLATE, "")))
	332	setlocale_failure = TRUE;
	333	else
	334	curcoll = savepv(curcoll);
	335	#endif /* USE_LOCALE_COLLATE */
	336	#ifdef USE_LOCALE_NUMERIC
	337	if (! (curnum = setlocale(LC_NUMERIC, "")))
	338	setlocale_failure = TRUE;
	339	else
	340	curnum = savepv(curnum);
	341	#endif /* USE_LOCALE_NUMERIC */
	342	}
	343
	344	if (setlocale_failure) {
	345	char *p;
0bd48802	346	const bool locwarn = (printwarn > 1 \|\|
98994639 HS	347	(printwarn &&
	348	(!(p = PerlEnv_getenv("PERL_BADLANG")) \|\| atoi(p))));
	349
	350	if (locwarn) {
	351	#ifdef LC_ALL
	352
	353	PerlIO_printf(Perl_error_log,
	354	"perl: warning: Setting locale failed.\n");
	355
	356	#else /* !LC_ALL */
	357
	358	PerlIO_printf(Perl_error_log,
	359	"perl: warning: Setting locale failed for the categories:\n\t");
	360	#ifdef USE_LOCALE_CTYPE
	361	if (! curctype)
	362	PerlIO_printf(Perl_error_log, "LC_CTYPE ");
	363	#endif /* USE_LOCALE_CTYPE */
	364	#ifdef USE_LOCALE_COLLATE
	365	if (! curcoll)
	366	PerlIO_printf(Perl_error_log, "LC_COLLATE ");
	367	#endif /* USE_LOCALE_COLLATE */
	368	#ifdef USE_LOCALE_NUMERIC
	369	if (! curnum)
	370	PerlIO_printf(Perl_error_log, "LC_NUMERIC ");
	371	#endif /* USE_LOCALE_NUMERIC */
	372	PerlIO_printf(Perl_error_log, "\n");
	373
	374	#endif /* LC_ALL */
	375
	376	PerlIO_printf(Perl_error_log,
	377	"perl: warning: Please check that your locale settings:\n");
	378
	379	#ifdef __GLIBC__
	380	PerlIO_printf(Perl_error_log,
	381	"\tLANGUAGE = %c%s%c,\n",
	382	language ? '"' : '(',
	383	language ? language : "unset",
	384	language ? '"' : ')');
	385	#endif
	386
	387	PerlIO_printf(Perl_error_log,
	388	"\tLC_ALL = %c%s%c,\n",
	389	lc_all ? '"' : '(',
	390	lc_all ? lc_all : "unset",
	391	lc_all ? '"' : ')');
	392
	393	#if defined(USE_ENVIRON_ARRAY)
	394	{
	395	char **e;
	396	for (e = environ; *e; e++) {
	397	if (strnEQ(*e, "LC_", 3)
	398	&& strnNE(*e, "LC_ALL=", 7)
	399	&& (p = strchr(*e, '=')))
	400	PerlIO_printf(Perl_error_log, "\t%.*s = \"%s\",\n",
	401	(int)(p - e), e, p + 1);
	402	}
	403	}
	404	#else
	405	PerlIO_printf(Perl_error_log,
	406	"\t(possibly more locale environment variables)\n");
	407	#endif
	408
	409	PerlIO_printf(Perl_error_log,
	410	"\tLANG = %c%s%c\n",
411	lang ? '"' : '(',
412	lang ? lang : "unset",
413	lang ? '"' : ')');
414
415	PerlIO_printf(Perl_error_log,
416	" are supported and installed on your system.\n");
417	}
418
419	#ifdef LC_ALL
420
421	if (setlocale(LC_ALL, "C")) {
422	if (locwarn)
423	PerlIO_printf(Perl_error_log,
424	"perl: warning: Falling back to the standard locale (\"C\").\n");
425	ok = 0;
426	}
427	else {
428	if (locwarn)
429	PerlIO_printf(Perl_error_log,
430	"perl: warning: Failed to fall back to the standard locale (\"C\").\n");
431	ok = -1;
432	}
433
434	#else /* ! LC_ALL */
435
436	if (0
437	#ifdef USE_LOCALE_CTYPE
438	\|\| !(curctype \|\| setlocale(LC_CTYPE, "C"))
439	#endif /* USE_LOCALE_CTYPE */
440	#ifdef USE_LOCALE_COLLATE
441	\|\| !(curcoll \|\| setlocale(LC_COLLATE, "C"))
442	#endif /* USE_LOCALE_COLLATE */
443	#ifdef USE_LOCALE_NUMERIC
444	\|\| !(curnum \|\| setlocale(LC_NUMERIC, "C"))
445	#endif /* USE_LOCALE_NUMERIC */
446	)
447	{
448	if (locwarn)
449	PerlIO_printf(Perl_error_log,
450	"perl: warning: Cannot fall back to the standard locale (\"C\").\n");
451	ok = -1;
452	}
453
454	#endif /* ! LC_ALL */
455
456	#ifdef USE_LOCALE_CTYPE
bd61b366	457	curctype = savepv(setlocale(LC_CTYPE, NULL));
98994639 HS	458	#endif /* USE_LOCALE_CTYPE */
98994639 HS	459	#ifdef USE_LOCALE_COLLATE
bd61b366	460	curcoll = savepv(setlocale(LC_COLLATE, NULL));
98994639 HS	461	#endif /* USE_LOCALE_COLLATE */
98994639 HS	462	#ifdef USE_LOCALE_NUMERIC
bd61b366	463	curnum = savepv(setlocale(LC_NUMERIC, NULL));
98994639 HS	464	#endif /* USE_LOCALE_NUMERIC */
	465	}
	466	else {
	467
	468	#ifdef USE_LOCALE_CTYPE
	469	new_ctype(curctype);
	470	#endif /* USE_LOCALE_CTYPE */
	471
	472	#ifdef USE_LOCALE_COLLATE
	473	new_collate(curcoll);
	474	#endif /* USE_LOCALE_COLLATE */
	475
	476	#ifdef USE_LOCALE_NUMERIC
	477	new_numeric(curnum);
	478	#endif /* USE_LOCALE_NUMERIC */
b310b053	479
98994639 HS	480	}
	481
	482	#endif /* USE_LOCALE */
	483
ec71e770	484	#ifdef USE_PERLIO
b310b053	485	{
fde18df1	486	/* Set PL_utf8locale to TRUE if using PerlIO _and_
ec71e770	487	any of the following are true:
085a54d9	488	- nl_langinfo(CODESET) contains /^utf-?8/i
61de9fb5	489	- $ENV{LC_ALL} contains /^utf-?8/i
085a54d9	490	- $ENV{LC_CTYPE} contains /^utf-?8/i
61de9fb5 JH	491	- $ENV{LANG} contains /^utf-?8/i
	492	The LC_ALL, LC_CTYPE, LANG obey the usual override
	493	hierarchy of locale environment variables. (LANGUAGE
	494	affects only LC_MESSAGES only under glibc.) (If present,
	495	it overrides LC_MESSAGES for GNU gettext, and it also
	496	can have more than one locale, separated by spaces,
	497	in case you need to know.)
8aa8f774	498	If PL_utf8locale and PL_unicode (set by -C or by $ENV{PERL_UNICODE})
a05d7ebb	499	are true, perl.c:S_parse_body() will turn on the PerlIO :utf8 layer
fde18df1	500	on STDIN, STDOUT, STDERR, _and_ the default open discipline.
085a54d9	501	*/
fde18df1	502	bool utf8locale = FALSE;
b310b053 JH	503	char *codeset = NULL;
	504	#if defined(HAS_NL_LANGINFO) && defined(CODESET)
	505	codeset = nl_langinfo(CODESET);
	506	#endif
61de9fb5	507	if (codeset)
89529cee GA	508	utf8locale = (Perl_ibcmp(aTHX_ codeset, STR_WITH_LEN("UTF-8")) == 0 \|\|
89529cee GA	509	Perl_ibcmp(aTHX_ codeset, STR_WITH_LEN("UTF8") ) == 0);
077a72a9	510	#if defined(USE_LOCALE)
61de9fb5 JH	511	else { /* nl_langinfo(CODESET) is supposed to correctly
	512	* interpret the locale environment variables,
	513	* but just in case it fails, let's do this manually. */
	514	if (lang)
89529cee GA	515	utf8locale = (Perl_ibcmp(aTHX_ lang, STR_WITH_LEN("UTF-8")) == 0 \|\|
89529cee GA	516	Perl_ibcmp(aTHX_ lang, STR_WITH_LEN("UTF8") ) == 0);
b310b053	517	#ifdef USE_LOCALE_CTYPE
61de9fb5	518	if (curctype)
89529cee GA	519	utf8locale = (Perl_ibcmp(aTHX_ curctype, STR_WITH_LEN("UTF-8")) == 0 \|\|
89529cee GA	520	Perl_ibcmp(aTHX_ curctype, STR_WITH_LEN("UTF8") ) == 0);
b310b053	521	#endif
61de9fb5	522	if (lc_all)
89529cee GA	523	utf8locale = (Perl_ibcmp(aTHX_ lc_all, STR_WITH_LEN("UTF-8")) == 0 \|\|
89529cee GA	524	Perl_ibcmp(aTHX_ lc_all, STR_WITH_LEN("UTF8") ) == 0);
61de9fb5	525	}
fde18df1 JH	526	#endif /* USE_LOCALE */
	527	if (utf8locale)
	528	PL_utf8locale = TRUE;
	529	}
a05d7ebb	530	/* Set PL_unicode to $ENV{PERL_UNICODE} if using PerlIO.
fde18df1 JH	531	This is an alternative to using the -C command line switch
	532	(the -C if present will override this). */
	533	{
dd374669	534	const char *p = PerlEnv_getenv("PERL_UNICODE");
a05d7ebb	535	PL_unicode = p ? parse_unicode_opts(&p) : 0;
b310b053	536	}
ec71e770	537	#endif
b310b053	538
98994639	539	#ifdef USE_LOCALE_CTYPE
43c5f42d	540	Safefree(curctype);
98994639 HS	541	#endif /* USE_LOCALE_CTYPE */
98994639 HS	542	#ifdef USE_LOCALE_COLLATE
43c5f42d	543	Safefree(curcoll);
98994639 HS	544	#endif /* USE_LOCALE_COLLATE */
98994639 HS	545	#ifdef USE_LOCALE_NUMERIC
43c5f42d	546	Safefree(curnum);
98994639 HS	547	#endif /* USE_LOCALE_NUMERIC */
	548	return ok;
	549	}
	550
98994639 HS	551	#ifdef USE_LOCALE_COLLATE
	552
	553	/*
	554	* mem_collxfrm() is a bit like strxfrm() but with two important
	555	* differences. First, it handles embedded NULs. Second, it allocates
	556	* a bit more memory than needed for the transformed data itself.
	557	* The real transformed data begins at offset sizeof(collationix).
	558	* Please see sv_collxfrm() to see how this is used.
	559	*/
166f8a29	560
98994639 HS	561	char *
	562	Perl_mem_collxfrm(pTHX_ const char s, STRLEN len, STRLEN xlen)
	563	{
97aff369	564	dVAR;
98994639 HS	565	char *xbuf;
	566	STRLEN xAlloc, xin, xout; /* xalloc is a reserved word in VC */
	567
	568	/* the first sizeof(collationix) bytes are used by sv_collxfrm(). */
	569	/* the +1 is for the terminating NUL. */
	570
	571	xAlloc = sizeof(PL_collation_ix) + PL_collxfrm_base + (PL_collxfrm_mult * len) + 1;
a02a5408	572	Newx(xbuf, xAlloc, char);
98994639 HS	573	if (! xbuf)
	574	goto bad;
	575
	576	(U32)xbuf = PL_collation_ix;
	577	xout = sizeof(PL_collation_ix);
	578	for (xin = 0; xin < len; ) {
	579	SSize_t xused;
	580
	581	for (;;) {
	582	xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
	583	if (xused == -1)
	584	goto bad;
eb160463	585	if ((STRLEN)xused < xAlloc - xout)
98994639 HS	586	break;
	587	xAlloc = (2 * xAlloc) + 1;
	588	Renew(xbuf, xAlloc, char);
	589	if (! xbuf)
	590	goto bad;
	591	}
	592
	593	xin += strlen(s + xin) + 1;
	594	xout += xused;
	595
	596	/* Embedded NULs are understood but silently skipped
	597	* because they make no sense in locale collation. */
	598	}
	599
	600	xbuf[xout] = '\0';
	601	*xlen = xout - sizeof(PL_collation_ix);
	602	return xbuf;
	603
	604	bad:
	605	Safefree(xbuf);
	606	*xlen = 0;
	607	return NULL;
	608	}
	609
	610	#endif /* USE_LOCALE_COLLATE */
	611
66610fdd RGS	612	/*
	613	* Local variables:
	614	* c-indentation-style: bsd
	615	* c-basic-offset: 4
	616	* indent-tabs-mode: t
	617	* End:
	618	*
37442d52 RGS	619	* ex: set ts=8 sts=4 sw=4 noet:
37442d52 RGS	620	*/