[perl5.git] / locale.c

/*    locale.c
 *
 *    Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999,
 *    2000, 2001, 2002, 2003, 2005, 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 *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
# ifdef HAS_LOCALECONV
    struct lconv* lc;

    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 = Nullsv;
	}
	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 = Nullsv;
# 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

    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

    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

    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

    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)

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