[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);
    PERL_UNUSED_CONTEXT;
}

/*
 * 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
	Safefree(curctype);
	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
	Safefree(curcoll);
	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
	Safefree(curnum);
	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
	Safefree(curctype);
	if (! (curctype = setlocale(LC_CTYPE, "")))
	    setlocale_failure = TRUE;
	else
	    curctype = savepv(curctype);
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	Safefree(curcoll);
	if (! (curcoll = setlocale(LC_COLLATE, "")))
	    setlocale_failure = TRUE;
	else
	    curcoll = savepv(curcoll);
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	Safefree(curnum);
	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
	Safefree(curctype);
	curctype = savepv(setlocale(LC_CTYPE, NULL));
#endif /* USE_LOCALE_CTYPE */
#ifdef USE_LOCALE_COLLATE
	Safefree(curcoll);
	curcoll = savepv(setlocale(LC_COLLATE, NULL));
#endif /* USE_LOCALE_COLLATE */
#ifdef USE_LOCALE_NUMERIC
	Safefree(curnum);
	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);
96a5add6	188	PERL_UNUSED_CONTEXT;
98994639 HS	189	}
	190
	191	/*
	192	* Set up for a new collation locale.
	193	*/
	194	void
8772537c	195	Perl_new_collate(pTHX_ const char *newcoll)
98994639 HS	196	{
98994639 HS	197	#ifdef USE_LOCALE_COLLATE
97aff369	198	dVAR;
98994639 HS	199
	200	if (! newcoll) {
	201	if (PL_collation_name) {
	202	++PL_collation_ix;
	203	Safefree(PL_collation_name);
	204	PL_collation_name = NULL;
	205	}
	206	PL_collation_standard = TRUE;
	207	PL_collxfrm_base = 0;
	208	PL_collxfrm_mult = 2;
	209	return;
	210	}
	211
	212	if (! PL_collation_name \|\| strNE(PL_collation_name, newcoll)) {
	213	++PL_collation_ix;
	214	Safefree(PL_collation_name);
	215	PL_collation_name = stdize_locale(savepv(newcoll));
770526c1 NC	216	PL_collation_standard = ((*newcoll == 'C' && newcoll[1] == '\0')
770526c1 NC	217	\|\| strEQ(newcoll, "POSIX"));
98994639 HS	218
	219	{
	220	/* 2: at most so many chars ('a', 'b'). */
	221	/* 50: surely no system expands a char more. */
	222	#define XFRMBUFSIZE (2 * 50)
	223	char xbuf[XFRMBUFSIZE];
8772537c AL	224	const Size_t fa = strxfrm(xbuf, "a", XFRMBUFSIZE);
	225	const Size_t fb = strxfrm(xbuf, "ab", XFRMBUFSIZE);
	226	const SSize_t mult = fb - fa;
98994639 HS	227	if (mult < 1)
98994639 HS	228	Perl_croak(aTHX_ "strxfrm() gets absurd");
eb160463	229	PL_collxfrm_base = (fa > (Size_t)mult) ? (fa - mult) : 0;
98994639 HS	230	PL_collxfrm_mult = mult;
	231	}
	232	}
	233
	234	#endif /* USE_LOCALE_COLLATE */
	235	}
	236
	237	/*
	238	* Initialize locale awareness.
	239	*/
	240	int
	241	Perl_init_i18nl10n(pTHX_ int printwarn)
	242	{
	243	int ok = 1;
	244	/* returns
	245	* 1 = set ok or not applicable,
	246	* 0 = fallback to C locale,
	247	* -1 = fallback to C locale failed
	248	*/
	249
	250	#if defined(USE_LOCALE)
97aff369	251	dVAR;
98994639 HS	252
	253	#ifdef USE_LOCALE_CTYPE
	254	char *curctype = NULL;
	255	#endif /* USE_LOCALE_CTYPE */
	256	#ifdef USE_LOCALE_COLLATE
	257	char *curcoll = NULL;
	258	#endif /* USE_LOCALE_COLLATE */
	259	#ifdef USE_LOCALE_NUMERIC
	260	char *curnum = NULL;
	261	#endif /* USE_LOCALE_NUMERIC */
	262	#ifdef __GLIBC__
7452cf6a	263	char * const language = PerlEnv_getenv("LANGUAGE");
98994639	264	#endif
7452cf6a AL	265	char * const lc_all = PerlEnv_getenv("LC_ALL");
7452cf6a AL	266	char * const lang = PerlEnv_getenv("LANG");
98994639 HS	267	bool setlocale_failure = FALSE;
	268
	269	#ifdef LOCALE_ENVIRON_REQUIRED
	270
	271	/*
	272	* Ultrix setlocale(..., "") fails if there are no environment
	273	* variables from which to get a locale name.
	274	*/
	275
	276	bool done = FALSE;
	277
	278	#ifdef LC_ALL
	279	if (lang) {
	280	if (setlocale(LC_ALL, ""))
	281	done = TRUE;
	282	else
	283	setlocale_failure = TRUE;
	284	}
	285	if (!setlocale_failure) {
	286	#ifdef USE_LOCALE_CTYPE
56279a21	287	Safefree(curctype);
98994639 HS	288	if (! (curctype =
	289	setlocale(LC_CTYPE,
	290	(!done && (lang \|\| PerlEnv_getenv("LC_CTYPE")))
bd61b366	291	? "" : NULL)))
98994639 HS	292	setlocale_failure = TRUE;
	293	else
	294	curctype = savepv(curctype);
	295	#endif /* USE_LOCALE_CTYPE */
	296	#ifdef USE_LOCALE_COLLATE
56279a21	297	Safefree(curcoll);
98994639 HS	298	if (! (curcoll =
	299	setlocale(LC_COLLATE,
	300	(!done && (lang \|\| PerlEnv_getenv("LC_COLLATE")))
bd61b366	301	? "" : NULL)))
98994639 HS	302	setlocale_failure = TRUE;
	303	else
	304	curcoll = savepv(curcoll);
	305	#endif /* USE_LOCALE_COLLATE */
	306	#ifdef USE_LOCALE_NUMERIC
56279a21	307	Safefree(curnum);
98994639 HS	308	if (! (curnum =
	309	setlocale(LC_NUMERIC,
	310	(!done && (lang \|\| PerlEnv_getenv("LC_NUMERIC")))
bd61b366	311	? "" : NULL)))
98994639 HS	312	setlocale_failure = TRUE;
	313	else
	314	curnum = savepv(curnum);
	315	#endif /* USE_LOCALE_NUMERIC */
	316	}
	317
	318	#endif /* LC_ALL */
	319
	320	#endif /* !LOCALE_ENVIRON_REQUIRED */
	321
	322	#ifdef LC_ALL
	323	if (! setlocale(LC_ALL, ""))
	324	setlocale_failure = TRUE;
	325	#endif /* LC_ALL */
	326
	327	if (!setlocale_failure) {
	328	#ifdef USE_LOCALE_CTYPE
6cb43dbf	329	Safefree(curctype);
98994639 HS	330	if (! (curctype = setlocale(LC_CTYPE, "")))
	331	setlocale_failure = TRUE;
	332	else
	333	curctype = savepv(curctype);
	334	#endif /* USE_LOCALE_CTYPE */
	335	#ifdef USE_LOCALE_COLLATE
6cb43dbf	336	Safefree(curcoll);
98994639 HS	337	if (! (curcoll = setlocale(LC_COLLATE, "")))
	338	setlocale_failure = TRUE;
	339	else
	340	curcoll = savepv(curcoll);
	341	#endif /* USE_LOCALE_COLLATE */
	342	#ifdef USE_LOCALE_NUMERIC
6cb43dbf	343	Safefree(curnum);
98994639 HS	344	if (! (curnum = setlocale(LC_NUMERIC, "")))
	345	setlocale_failure = TRUE;
	346	else
	347	curnum = savepv(curnum);
	348	#endif /* USE_LOCALE_NUMERIC */
	349	}
	350
	351	if (setlocale_failure) {
	352	char *p;
0bd48802	353	const bool locwarn = (printwarn > 1 \|\|
98994639 HS	354	(printwarn &&
	355	(!(p = PerlEnv_getenv("PERL_BADLANG")) \|\| atoi(p))));
	356
	357	if (locwarn) {
	358	#ifdef LC_ALL
	359
	360	PerlIO_printf(Perl_error_log,
	361	"perl: warning: Setting locale failed.\n");
	362
	363	#else /* !LC_ALL */
	364
	365	PerlIO_printf(Perl_error_log,
	366	"perl: warning: Setting locale failed for the categories:\n\t");
	367	#ifdef USE_LOCALE_CTYPE
	368	if (! curctype)
	369	PerlIO_printf(Perl_error_log, "LC_CTYPE ");
	370	#endif /* USE_LOCALE_CTYPE */
	371	#ifdef USE_LOCALE_COLLATE
	372	if (! curcoll)
	373	PerlIO_printf(Perl_error_log, "LC_COLLATE ");
	374	#endif /* USE_LOCALE_COLLATE */
	375	#ifdef USE_LOCALE_NUMERIC
	376	if (! curnum)
	377	PerlIO_printf(Perl_error_log, "LC_NUMERIC ");
	378	#endif /* USE_LOCALE_NUMERIC */
	379	PerlIO_printf(Perl_error_log, "\n");
	380
	381	#endif /* LC_ALL */
	382
	383	PerlIO_printf(Perl_error_log,
	384	"perl: warning: Please check that your locale settings:\n");
	385
	386	#ifdef __GLIBC__
	387	PerlIO_printf(Perl_error_log,
	388	"\tLANGUAGE = %c%s%c,\n",
	389	language ? '"' : '(',
	390	language ? language : "unset",
	391	language ? '"' : ')');
	392	#endif
	393
	394	PerlIO_printf(Perl_error_log,
	395	"\tLC_ALL = %c%s%c,\n",
	396	lc_all ? '"' : '(',
	397	lc_all ? lc_all : "unset",
	398	lc_all ? '"' : ')');
	399
	400	#if defined(USE_ENVIRON_ARRAY)
	401	{
	402	char **e;
	403	for (e = environ; *e; e++) {
	404	if (strnEQ(*e, "LC_", 3)
	405	&& strnNE(*e, "LC_ALL=", 7)
	406	&& (p = strchr(*e, '=')))
	407	PerlIO_printf(Perl_error_log, "\t%.*s = \"%s\",\n",
	408	(int)(p - e), e, p + 1);
	409	}
	410	}
	411	#else
	412	PerlIO_printf(Perl_error_log,
	413	"\t(possibly more locale environment variables)\n");
	414	#endif
	415
	416	PerlIO_printf(Perl_error_log,
	417	"\tLANG = %c%s%c\n",
418	lang ? '"' : '(',
419	lang ? lang : "unset",
420	lang ? '"' : ')');
421
422	PerlIO_printf(Perl_error_log,
423	" are supported and installed on your system.\n");
424	}
425
426	#ifdef LC_ALL
427
428	if (setlocale(LC_ALL, "C")) {
429	if (locwarn)
430	PerlIO_printf(Perl_error_log,
431	"perl: warning: Falling back to the standard locale (\"C\").\n");
432	ok = 0;
433	}
434	else {
435	if (locwarn)
436	PerlIO_printf(Perl_error_log,
437	"perl: warning: Failed to fall back to the standard locale (\"C\").\n");
438	ok = -1;
439	}
440
441	#else /* ! LC_ALL */
442
443	if (0
444	#ifdef USE_LOCALE_CTYPE
445	\|\| !(curctype \|\| setlocale(LC_CTYPE, "C"))
446	#endif /* USE_LOCALE_CTYPE */
447	#ifdef USE_LOCALE_COLLATE
448	\|\| !(curcoll \|\| setlocale(LC_COLLATE, "C"))
449	#endif /* USE_LOCALE_COLLATE */
450	#ifdef USE_LOCALE_NUMERIC
451	\|\| !(curnum \|\| setlocale(LC_NUMERIC, "C"))
452	#endif /* USE_LOCALE_NUMERIC */
453	)
454	{
455	if (locwarn)
456	PerlIO_printf(Perl_error_log,
457	"perl: warning: Cannot fall back to the standard locale (\"C\").\n");
458	ok = -1;
459	}
460
461	#endif /* ! LC_ALL */
462
463	#ifdef USE_LOCALE_CTYPE
56279a21	464	Safefree(curctype);
bd61b366	465	curctype = savepv(setlocale(LC_CTYPE, NULL));
98994639 HS	466	#endif /* USE_LOCALE_CTYPE */
98994639 HS	467	#ifdef USE_LOCALE_COLLATE
56279a21	468	Safefree(curcoll);
bd61b366	469	curcoll = savepv(setlocale(LC_COLLATE, NULL));
98994639 HS	470	#endif /* USE_LOCALE_COLLATE */
98994639 HS	471	#ifdef USE_LOCALE_NUMERIC
56279a21	472	Safefree(curnum);
bd61b366	473	curnum = savepv(setlocale(LC_NUMERIC, NULL));
98994639 HS	474	#endif /* USE_LOCALE_NUMERIC */
	475	}
	476	else {
	477
	478	#ifdef USE_LOCALE_CTYPE
	479	new_ctype(curctype);
	480	#endif /* USE_LOCALE_CTYPE */
	481
	482	#ifdef USE_LOCALE_COLLATE
	483	new_collate(curcoll);
	484	#endif /* USE_LOCALE_COLLATE */
	485
	486	#ifdef USE_LOCALE_NUMERIC
	487	new_numeric(curnum);
	488	#endif /* USE_LOCALE_NUMERIC */
b310b053	489
98994639 HS	490	}
	491
	492	#endif /* USE_LOCALE */
	493
ec71e770	494	#ifdef USE_PERLIO
b310b053	495	{
fde18df1	496	/* Set PL_utf8locale to TRUE if using PerlIO _and_
ec71e770	497	any of the following are true:
085a54d9	498	- nl_langinfo(CODESET) contains /^utf-?8/i
61de9fb5	499	- $ENV{LC_ALL} contains /^utf-?8/i
085a54d9	500	- $ENV{LC_CTYPE} contains /^utf-?8/i
61de9fb5 JH	501	- $ENV{LANG} contains /^utf-?8/i
	502	The LC_ALL, LC_CTYPE, LANG obey the usual override
	503	hierarchy of locale environment variables. (LANGUAGE
	504	affects only LC_MESSAGES only under glibc.) (If present,
	505	it overrides LC_MESSAGES for GNU gettext, and it also
	506	can have more than one locale, separated by spaces,
	507	in case you need to know.)
8aa8f774	508	If PL_utf8locale and PL_unicode (set by -C or by $ENV{PERL_UNICODE})
a05d7ebb	509	are true, perl.c:S_parse_body() will turn on the PerlIO :utf8 layer
fde18df1	510	on STDIN, STDOUT, STDERR, _and_ the default open discipline.
085a54d9	511	*/
fde18df1	512	bool utf8locale = FALSE;
b310b053 JH	513	char *codeset = NULL;
	514	#if defined(HAS_NL_LANGINFO) && defined(CODESET)
	515	codeset = nl_langinfo(CODESET);
	516	#endif
61de9fb5	517	if (codeset)
89529cee GA	518	utf8locale = (Perl_ibcmp(aTHX_ codeset, STR_WITH_LEN("UTF-8")) == 0 \|\|
89529cee GA	519	Perl_ibcmp(aTHX_ codeset, STR_WITH_LEN("UTF8") ) == 0);
077a72a9	520	#if defined(USE_LOCALE)
61de9fb5 JH	521	else { /* nl_langinfo(CODESET) is supposed to correctly
	522	* interpret the locale environment variables,
	523	* but just in case it fails, let's do this manually. */
	524	if (lang)
89529cee GA	525	utf8locale = (Perl_ibcmp(aTHX_ lang, STR_WITH_LEN("UTF-8")) == 0 \|\|
89529cee GA	526	Perl_ibcmp(aTHX_ lang, STR_WITH_LEN("UTF8") ) == 0);
b310b053	527	#ifdef USE_LOCALE_CTYPE
61de9fb5	528	if (curctype)
89529cee GA	529	utf8locale = (Perl_ibcmp(aTHX_ curctype, STR_WITH_LEN("UTF-8")) == 0 \|\|
89529cee GA	530	Perl_ibcmp(aTHX_ curctype, STR_WITH_LEN("UTF8") ) == 0);
b310b053	531	#endif
61de9fb5	532	if (lc_all)
89529cee GA	533	utf8locale = (Perl_ibcmp(aTHX_ lc_all, STR_WITH_LEN("UTF-8")) == 0 \|\|
89529cee GA	534	Perl_ibcmp(aTHX_ lc_all, STR_WITH_LEN("UTF8") ) == 0);
61de9fb5	535	}
fde18df1 JH	536	#endif /* USE_LOCALE */
	537	if (utf8locale)
	538	PL_utf8locale = TRUE;
	539	}
a05d7ebb	540	/* Set PL_unicode to $ENV{PERL_UNICODE} if using PerlIO.
fde18df1 JH	541	This is an alternative to using the -C command line switch
	542	(the -C if present will override this). */
	543	{
dd374669	544	const char *p = PerlEnv_getenv("PERL_UNICODE");
a05d7ebb	545	PL_unicode = p ? parse_unicode_opts(&p) : 0;
b310b053	546	}
ec71e770	547	#endif
b310b053	548
98994639	549	#ifdef USE_LOCALE_CTYPE
43c5f42d	550	Safefree(curctype);
98994639 HS	551	#endif /* USE_LOCALE_CTYPE */
98994639 HS	552	#ifdef USE_LOCALE_COLLATE
43c5f42d	553	Safefree(curcoll);
98994639 HS	554	#endif /* USE_LOCALE_COLLATE */
98994639 HS	555	#ifdef USE_LOCALE_NUMERIC
43c5f42d	556	Safefree(curnum);
98994639 HS	557	#endif /* USE_LOCALE_NUMERIC */
	558	return ok;
	559	}
	560
98994639 HS	561	#ifdef USE_LOCALE_COLLATE
	562
	563	/*
	564	* mem_collxfrm() is a bit like strxfrm() but with two important
	565	* differences. First, it handles embedded NULs. Second, it allocates
	566	* a bit more memory than needed for the transformed data itself.
	567	* The real transformed data begins at offset sizeof(collationix).
	568	* Please see sv_collxfrm() to see how this is used.
	569	*/
166f8a29	570
98994639 HS	571	char *
	572	Perl_mem_collxfrm(pTHX_ const char s, STRLEN len, STRLEN xlen)
	573	{
97aff369	574	dVAR;
98994639 HS	575	char *xbuf;
	576	STRLEN xAlloc, xin, xout; /* xalloc is a reserved word in VC */
	577
	578	/* the first sizeof(collationix) bytes are used by sv_collxfrm(). */
	579	/* the +1 is for the terminating NUL. */
	580
	581	xAlloc = sizeof(PL_collation_ix) + PL_collxfrm_base + (PL_collxfrm_mult * len) + 1;
a02a5408	582	Newx(xbuf, xAlloc, char);
98994639 HS	583	if (! xbuf)
	584	goto bad;
	585
	586	(U32)xbuf = PL_collation_ix;
	587	xout = sizeof(PL_collation_ix);
	588	for (xin = 0; xin < len; ) {
	589	SSize_t xused;
	590
	591	for (;;) {
	592	xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
	593	if (xused == -1)
	594	goto bad;
eb160463	595	if ((STRLEN)xused < xAlloc - xout)
98994639 HS	596	break;
	597	xAlloc = (2 * xAlloc) + 1;
	598	Renew(xbuf, xAlloc, char);
	599	if (! xbuf)
	600	goto bad;
	601	}
	602
	603	xin += strlen(s + xin) + 1;
	604	xout += xused;
	605
	606	/* Embedded NULs are understood but silently skipped
	607	* because they make no sense in locale collation. */
	608	}
	609
	610	xbuf[xout] = '\0';
	611	*xlen = xout - sizeof(PL_collation_ix);
	612	return xbuf;
	613
	614	bad:
	615	Safefree(xbuf);
	616	*xlen = 0;
	617	return NULL;
	618	}
	619
	620	#endif /* USE_LOCALE_COLLATE */
	621
66610fdd RGS	622	/*
	623	* Local variables:
	624	* c-indentation-style: bsd
	625	* c-basic-offset: 4
	626	* indent-tabs-mode: t
	627	* End:
	628	*
37442d52 RGS	629	* ex: set ts=8 sts=4 sw=4 noet:
37442d52 RGS	630	*/