[perl5.git] / locale.c

/*    locale.c
 *
 *    Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001,
 *    2002, 2003, 2005, 2006, 2007, 2008 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!
 *
 *     [p.238 of _The Lord of the Rings_, II/i: "Many Meetings"]
 */

/* 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 unspecified 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;

    PERL_ARGS_ASSERT_STDIZE_LOCALE;

    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;

    PERL_ARGS_ASSERT_NEW_CTYPE;

    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_ARGS_ASSERT_NEW_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 && !(fa == 0 && fb == 0))
	      Perl_croak(aTHX_ "panic: strxfrm() gets absurd - a => %"UVuf", ab => %"UVuf,
			 (UV) fa, (UV) fb);
	  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 = (foldEQ(codeset, STR_WITH_LEN("UTF-8"))
			    || foldEQ(codeset, STR_WITH_LEN("UTF8") ));
#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 = (foldEQ(lang, STR_WITH_LEN("UTF-8"))
				 || foldEQ(lang, STR_WITH_LEN("UTF8") ));
#ifdef USE_LOCALE_CTYPE
	      if (curctype)
		   utf8locale = (foldEQ(curctype, STR_WITH_LEN("UTF-8"))
				 || foldEQ(curctype, STR_WITH_LEN("UTF8") ));
#endif
	      if (lc_all)
		   utf8locale = (foldEQ(lc_all, STR_WITH_LEN("UTF-8"))
				 || foldEQ(lc_all, STR_WITH_LEN("UTF8") ));
	 }
#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;
	 if (PL_unicode & PERL_UNICODE_UTF8CACHEASSERT_FLAG)
	     PL_utf8cache = -1;
    }
#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 */

    PERL_ARGS_ASSERT_MEM_COLLXFRM;

    /* 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; ) {
	Size_t xused;

	for (;;) {
	    xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
	    if (xused >= PERL_INT_MAX)
		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	*
1129b882 NC	3	* Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999, 2000, 2001,
1129b882 NC	4	* 2002, 2003, 2005, 2006, 2007, 2008 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	/*
4ac71550	12	* A Elbereth Gilthoniel,
cdad3b53	13	* silivren penna míriel
4ac71550	14	* o menel aglar elenath!
cdad3b53	15	* Na-chaered palan-díriel
4ac71550 TC	16	* o galadhremmin ennorath,
	17	* Fanuilos, le linnathon
	18	* nef aear, si nef aearon!
	19	*
	20	* [p.238 of _The Lord of the Rings_, II/i: "Many Meetings"]
98994639 HS	21	*/
98994639 HS	22
166f8a29 DM	23	/* utility functions for handling locale-specific stuff like what
	24	* character represents the decimal point.
	25	*/
	26
98994639 HS	27	#include "EXTERN.h"
	28	#define PERL_IN_LOCALE_C
	29	#include "perl.h"
	30
	31	#ifdef I_LOCALE
	32	# include <locale.h>
	33	#endif
	34
b310b053 JH	35	#ifdef I_LANGINFO
	36	# include <langinfo.h>
	37	#endif
	38
a4af207c JH	39	#include "reentr.h"
a4af207c JH	40
27da23d5	41	#if defined(USE_LOCALE_NUMERIC) \|\| defined(USE_LOCALE_COLLATE)
98994639 HS	42	/*
	43	* Standardize the locale name from a string returned by 'setlocale'.
	44	*
	45	* The standard return value of setlocale() is either
	46	* (1) "xx_YY" if the first argument of setlocale() is not LC_ALL
	47	* (2) "xa_YY xb_YY ..." if the first argument of setlocale() is LC_ALL
	48	* (the space-separated values represent the various sublocales,
486ec47a	49	* in some unspecified order)
98994639 HS	50	*
	51	* In some platforms it has a form like "LC_SOMETHING=Lang_Country.866\n",
	52	* which is harmful for further use of the string in setlocale().
	53	*
	54	*/
	55	STATIC char *
	56	S_stdize_locale(pTHX_ char *locs)
	57	{
7452cf6a	58	const char * const s = strchr(locs, '=');
98994639 HS	59	bool okay = TRUE;
98994639 HS	60
7918f24d NC	61	PERL_ARGS_ASSERT_STDIZE_LOCALE;
7918f24d NC	62
8772537c AL	63	if (s) {
8772537c AL	64	const char * const t = strchr(s, '.');
98994639	65	okay = FALSE;
8772537c AL	66	if (t) {
	67	const char * const u = strchr(t, '\n');
	68	if (u && (u[1] == 0)) {
	69	const STRLEN len = u - s;
	70	Move(s + 1, locs, len, char);
	71	locs[len] = 0;
	72	okay = TRUE;
98994639 HS	73	}
	74	}
	75	}
	76
	77	if (!okay)
	78	Perl_croak(aTHX_ "Can't fix broken locale name \"%s\"", locs);
	79
	80	return locs;
	81	}
27da23d5	82	#endif
98994639 HS	83
	84	void
	85	Perl_set_numeric_radix(pTHX)
	86	{
	87	#ifdef USE_LOCALE_NUMERIC
97aff369	88	dVAR;
98994639	89	# ifdef HAS_LOCALECONV
7452cf6a	90	const struct lconv* const lc = localeconv();
98994639	91
98994639 HS	92	if (lc && lc->decimal_point) {
	93	if (lc->decimal_point[0] == '.' && lc->decimal_point[1] == 0) {
	94	SvREFCNT_dec(PL_numeric_radix_sv);
a0714e2c	95	PL_numeric_radix_sv = NULL;
98994639 HS	96	}
	97	else {
	98	if (PL_numeric_radix_sv)
	99	sv_setpv(PL_numeric_radix_sv, lc->decimal_point);
	100	else
	101	PL_numeric_radix_sv = newSVpv(lc->decimal_point, 0);
	102	}
	103	}
	104	else
a0714e2c	105	PL_numeric_radix_sv = NULL;
98994639 HS	106	# endif /* HAS_LOCALECONV */
	107	#endif /* USE_LOCALE_NUMERIC */
	108	}
	109
	110	/*
	111	* Set up for a new numeric locale.
	112	*/
	113	void
8772537c	114	Perl_new_numeric(pTHX_ const char *newnum)
98994639 HS	115	{
98994639 HS	116	#ifdef USE_LOCALE_NUMERIC
97aff369	117	dVAR;
98994639 HS	118
98994639 HS	119	if (! newnum) {
43c5f42d NC	120	Safefree(PL_numeric_name);
43c5f42d NC	121	PL_numeric_name = NULL;
98994639 HS	122	PL_numeric_standard = TRUE;
	123	PL_numeric_local = TRUE;
	124	return;
	125	}
	126
	127	if (! PL_numeric_name \|\| strNE(PL_numeric_name, newnum)) {
	128	Safefree(PL_numeric_name);
	129	PL_numeric_name = stdize_locale(savepv(newnum));
770526c1 NC	130	PL_numeric_standard = ((*newnum == 'C' && newnum[1] == '\0')
770526c1 NC	131	\|\| strEQ(newnum, "POSIX"));
98994639 HS	132	PL_numeric_local = TRUE;
	133	set_numeric_radix();
	134	}
	135
	136	#endif /* USE_LOCALE_NUMERIC */
	137	}
	138
	139	void
	140	Perl_set_numeric_standard(pTHX)
	141	{
	142	#ifdef USE_LOCALE_NUMERIC
97aff369	143	dVAR;
98994639 HS	144
	145	if (! PL_numeric_standard) {
	146	setlocale(LC_NUMERIC, "C");
	147	PL_numeric_standard = TRUE;
	148	PL_numeric_local = FALSE;
	149	set_numeric_radix();
	150	}
	151
	152	#endif /* USE_LOCALE_NUMERIC */
	153	}
	154
	155	void
	156	Perl_set_numeric_local(pTHX)
	157	{
	158	#ifdef USE_LOCALE_NUMERIC
97aff369	159	dVAR;
98994639 HS	160
	161	if (! PL_numeric_local) {
	162	setlocale(LC_NUMERIC, PL_numeric_name);
	163	PL_numeric_standard = FALSE;
	164	PL_numeric_local = TRUE;
	165	set_numeric_radix();
	166	}
	167
	168	#endif /* USE_LOCALE_NUMERIC */
	169	}
	170
	171	/*
	172	* Set up for a new ctype locale.
	173	*/
	174	void
8772537c	175	Perl_new_ctype(pTHX_ const char *newctype)
98994639 HS	176	{
98994639 HS	177	#ifdef USE_LOCALE_CTYPE
27da23d5	178	dVAR;
98994639 HS	179	int i;
98994639 HS	180
7918f24d NC	181	PERL_ARGS_ASSERT_NEW_CTYPE;
7918f24d NC	182
98994639 HS	183	for (i = 0; i < 256; i++) {
	184	if (isUPPER_LC(i))
	185	PL_fold_locale[i] = toLOWER_LC(i);
	186	else if (isLOWER_LC(i))
	187	PL_fold_locale[i] = toUPPER_LC(i);
	188	else
	189	PL_fold_locale[i] = i;
	190	}
	191
	192	#endif /* USE_LOCALE_CTYPE */
7918f24d	193	PERL_ARGS_ASSERT_NEW_CTYPE;
8772537c	194	PERL_UNUSED_ARG(newctype);
96a5add6	195	PERL_UNUSED_CONTEXT;
98994639 HS	196	}
	197
	198	/*
	199	* Set up for a new collation locale.
	200	*/
	201	void
8772537c	202	Perl_new_collate(pTHX_ const char *newcoll)
98994639 HS	203	{
98994639 HS	204	#ifdef USE_LOCALE_COLLATE
97aff369	205	dVAR;
98994639 HS	206
	207	if (! newcoll) {
	208	if (PL_collation_name) {
	209	++PL_collation_ix;
	210	Safefree(PL_collation_name);
	211	PL_collation_name = NULL;
	212	}
	213	PL_collation_standard = TRUE;
	214	PL_collxfrm_base = 0;
	215	PL_collxfrm_mult = 2;
	216	return;
	217	}
	218
	219	if (! PL_collation_name \|\| strNE(PL_collation_name, newcoll)) {
	220	++PL_collation_ix;
	221	Safefree(PL_collation_name);
	222	PL_collation_name = stdize_locale(savepv(newcoll));
770526c1 NC	223	PL_collation_standard = ((*newcoll == 'C' && newcoll[1] == '\0')
770526c1 NC	224	\|\| strEQ(newcoll, "POSIX"));
98994639 HS	225
	226	{
	227	/* 2: at most so many chars ('a', 'b'). */
	228	/* 50: surely no system expands a char more. */
	229	#define XFRMBUFSIZE (2 * 50)
	230	char xbuf[XFRMBUFSIZE];
8772537c AL	231	const Size_t fa = strxfrm(xbuf, "a", XFRMBUFSIZE);
	232	const Size_t fb = strxfrm(xbuf, "ab", XFRMBUFSIZE);
	233	const SSize_t mult = fb - fa;
e0601d3c	234	if (mult < 1 && !(fa == 0 && fb == 0))
ad49ad39 NC	235	Perl_croak(aTHX_ "panic: strxfrm() gets absurd - a => %"UVuf", ab => %"UVuf,
ad49ad39 NC	236	(UV) fa, (UV) fb);
eb160463	237	PL_collxfrm_base = (fa > (Size_t)mult) ? (fa - mult) : 0;
98994639 HS	238	PL_collxfrm_mult = mult;
	239	}
	240	}
	241
	242	#endif /* USE_LOCALE_COLLATE */
	243	}
	244
	245	/*
	246	* Initialize locale awareness.
	247	*/
	248	int
	249	Perl_init_i18nl10n(pTHX_ int printwarn)
	250	{
	251	int ok = 1;
	252	/* returns
	253	* 1 = set ok or not applicable,
	254	* 0 = fallback to C locale,
	255	* -1 = fallback to C locale failed
	256	*/
	257
	258	#if defined(USE_LOCALE)
97aff369	259	dVAR;
98994639 HS	260
	261	#ifdef USE_LOCALE_CTYPE
	262	char *curctype = NULL;
	263	#endif /* USE_LOCALE_CTYPE */
	264	#ifdef USE_LOCALE_COLLATE
	265	char *curcoll = NULL;
	266	#endif /* USE_LOCALE_COLLATE */
	267	#ifdef USE_LOCALE_NUMERIC
	268	char *curnum = NULL;
	269	#endif /* USE_LOCALE_NUMERIC */
	270	#ifdef __GLIBC__
7452cf6a	271	char * const language = PerlEnv_getenv("LANGUAGE");
98994639	272	#endif
7452cf6a AL	273	char * const lc_all = PerlEnv_getenv("LC_ALL");
7452cf6a AL	274	char * const lang = PerlEnv_getenv("LANG");
98994639 HS	275	bool setlocale_failure = FALSE;
	276
	277	#ifdef LOCALE_ENVIRON_REQUIRED
	278
	279	/*
	280	* Ultrix setlocale(..., "") fails if there are no environment
	281	* variables from which to get a locale name.
	282	*/
	283
	284	bool done = FALSE;
	285
	286	#ifdef LC_ALL
	287	if (lang) {
	288	if (setlocale(LC_ALL, ""))
	289	done = TRUE;
	290	else
	291	setlocale_failure = TRUE;
	292	}
	293	if (!setlocale_failure) {
	294	#ifdef USE_LOCALE_CTYPE
56279a21	295	Safefree(curctype);
98994639 HS	296	if (! (curctype =
	297	setlocale(LC_CTYPE,
	298	(!done && (lang \|\| PerlEnv_getenv("LC_CTYPE")))
bd61b366	299	? "" : NULL)))
98994639 HS	300	setlocale_failure = TRUE;
	301	else
	302	curctype = savepv(curctype);
	303	#endif /* USE_LOCALE_CTYPE */
	304	#ifdef USE_LOCALE_COLLATE
56279a21	305	Safefree(curcoll);
98994639 HS	306	if (! (curcoll =
	307	setlocale(LC_COLLATE,
	308	(!done && (lang \|\| PerlEnv_getenv("LC_COLLATE")))
bd61b366	309	? "" : NULL)))
98994639 HS	310	setlocale_failure = TRUE;
	311	else
	312	curcoll = savepv(curcoll);
	313	#endif /* USE_LOCALE_COLLATE */
	314	#ifdef USE_LOCALE_NUMERIC
56279a21	315	Safefree(curnum);
98994639 HS	316	if (! (curnum =
	317	setlocale(LC_NUMERIC,
	318	(!done && (lang \|\| PerlEnv_getenv("LC_NUMERIC")))
bd61b366	319	? "" : NULL)))
98994639 HS	320	setlocale_failure = TRUE;
	321	else
	322	curnum = savepv(curnum);
	323	#endif /* USE_LOCALE_NUMERIC */
	324	}
	325
	326	#endif /* LC_ALL */
	327
	328	#endif /* !LOCALE_ENVIRON_REQUIRED */
	329
	330	#ifdef LC_ALL
	331	if (! setlocale(LC_ALL, ""))
	332	setlocale_failure = TRUE;
	333	#endif /* LC_ALL */
	334
	335	if (!setlocale_failure) {
	336	#ifdef USE_LOCALE_CTYPE
6cb43dbf	337	Safefree(curctype);
98994639 HS	338	if (! (curctype = setlocale(LC_CTYPE, "")))
	339	setlocale_failure = TRUE;
	340	else
	341	curctype = savepv(curctype);
	342	#endif /* USE_LOCALE_CTYPE */
	343	#ifdef USE_LOCALE_COLLATE
6cb43dbf	344	Safefree(curcoll);
98994639 HS	345	if (! (curcoll = setlocale(LC_COLLATE, "")))
	346	setlocale_failure = TRUE;
	347	else
	348	curcoll = savepv(curcoll);
	349	#endif /* USE_LOCALE_COLLATE */
	350	#ifdef USE_LOCALE_NUMERIC
6cb43dbf	351	Safefree(curnum);
98994639 HS	352	if (! (curnum = setlocale(LC_NUMERIC, "")))
	353	setlocale_failure = TRUE;
	354	else
	355	curnum = savepv(curnum);
	356	#endif /* USE_LOCALE_NUMERIC */
	357	}
	358
	359	if (setlocale_failure) {
	360	char *p;
0bd48802	361	const bool locwarn = (printwarn > 1 \|\|
98994639 HS	362	(printwarn &&
	363	(!(p = PerlEnv_getenv("PERL_BADLANG")) \|\| atoi(p))));
	364
	365	if (locwarn) {
	366	#ifdef LC_ALL
	367
	368	PerlIO_printf(Perl_error_log,
	369	"perl: warning: Setting locale failed.\n");
	370
	371	#else /* !LC_ALL */
	372
	373	PerlIO_printf(Perl_error_log,
	374	"perl: warning: Setting locale failed for the categories:\n\t");
	375	#ifdef USE_LOCALE_CTYPE
	376	if (! curctype)
	377	PerlIO_printf(Perl_error_log, "LC_CTYPE ");
	378	#endif /* USE_LOCALE_CTYPE */
	379	#ifdef USE_LOCALE_COLLATE
	380	if (! curcoll)
	381	PerlIO_printf(Perl_error_log, "LC_COLLATE ");
	382	#endif /* USE_LOCALE_COLLATE */
	383	#ifdef USE_LOCALE_NUMERIC
	384	if (! curnum)
	385	PerlIO_printf(Perl_error_log, "LC_NUMERIC ");
	386	#endif /* USE_LOCALE_NUMERIC */
	387	PerlIO_printf(Perl_error_log, "\n");
	388
	389	#endif /* LC_ALL */
	390
	391	PerlIO_printf(Perl_error_log,
	392	"perl: warning: Please check that your locale settings:\n");
	393
	394	#ifdef __GLIBC__
	395	PerlIO_printf(Perl_error_log,
	396	"\tLANGUAGE = %c%s%c,\n",
	397	language ? '"' : '(',
	398	language ? language : "unset",
	399	language ? '"' : ')');
	400	#endif
	401
	402	PerlIO_printf(Perl_error_log,
	403	"\tLC_ALL = %c%s%c,\n",
	404	lc_all ? '"' : '(',
	405	lc_all ? lc_all : "unset",
	406	lc_all ? '"' : ')');
	407
	408	#if defined(USE_ENVIRON_ARRAY)
	409	{
	410	char **e;
	411	for (e = environ; *e; e++) {
	412	if (strnEQ(*e, "LC_", 3)
	413	&& strnNE(*e, "LC_ALL=", 7)
	414	&& (p = strchr(*e, '=')))
	415	PerlIO_printf(Perl_error_log, "\t%.*s = \"%s\",\n",
	416	(int)(p - e), e, p + 1);
	417	}
	418	}
	419	#else
	420	PerlIO_printf(Perl_error_log,
	421	"\t(possibly more locale environment variables)\n");
	422	#endif
	423
	424	PerlIO_printf(Perl_error_log,
	425	"\tLANG = %c%s%c\n",
426	lang ? '"' : '(',
427	lang ? lang : "unset",
428	lang ? '"' : ')');
429
430	PerlIO_printf(Perl_error_log,
431	" are supported and installed on your system.\n");
432	}
433
434	#ifdef LC_ALL
435
436	if (setlocale(LC_ALL, "C")) {
437	if (locwarn)
438	PerlIO_printf(Perl_error_log,
439	"perl: warning: Falling back to the standard locale (\"C\").\n");
440	ok = 0;
441	}
442	else {
443	if (locwarn)
444	PerlIO_printf(Perl_error_log,
445	"perl: warning: Failed to fall back to the standard locale (\"C\").\n");
446	ok = -1;
447	}
448
449	#else /* ! LC_ALL */
450
451	if (0
452	#ifdef USE_LOCALE_CTYPE
453	\|\| !(curctype \|\| setlocale(LC_CTYPE, "C"))
454	#endif /* USE_LOCALE_CTYPE */
455	#ifdef USE_LOCALE_COLLATE
456	\|\| !(curcoll \|\| setlocale(LC_COLLATE, "C"))
457	#endif /* USE_LOCALE_COLLATE */
458	#ifdef USE_LOCALE_NUMERIC
459	\|\| !(curnum \|\| setlocale(LC_NUMERIC, "C"))
460	#endif /* USE_LOCALE_NUMERIC */
461	)
462	{
463	if (locwarn)
464	PerlIO_printf(Perl_error_log,
465	"perl: warning: Cannot fall back to the standard locale (\"C\").\n");
466	ok = -1;
467	}
468
469	#endif /* ! LC_ALL */
470
471	#ifdef USE_LOCALE_CTYPE
56279a21	472	Safefree(curctype);
bd61b366	473	curctype = savepv(setlocale(LC_CTYPE, NULL));
98994639 HS	474	#endif /* USE_LOCALE_CTYPE */
98994639 HS	475	#ifdef USE_LOCALE_COLLATE
56279a21	476	Safefree(curcoll);
bd61b366	477	curcoll = savepv(setlocale(LC_COLLATE, NULL));
98994639 HS	478	#endif /* USE_LOCALE_COLLATE */
98994639 HS	479	#ifdef USE_LOCALE_NUMERIC
56279a21	480	Safefree(curnum);
bd61b366	481	curnum = savepv(setlocale(LC_NUMERIC, NULL));
98994639 HS	482	#endif /* USE_LOCALE_NUMERIC */
	483	}
	484	else {
	485
	486	#ifdef USE_LOCALE_CTYPE
	487	new_ctype(curctype);
	488	#endif /* USE_LOCALE_CTYPE */
	489
	490	#ifdef USE_LOCALE_COLLATE
	491	new_collate(curcoll);
	492	#endif /* USE_LOCALE_COLLATE */
	493
	494	#ifdef USE_LOCALE_NUMERIC
	495	new_numeric(curnum);
	496	#endif /* USE_LOCALE_NUMERIC */
b310b053	497
98994639 HS	498	}
	499
	500	#endif /* USE_LOCALE */
	501
ec71e770	502	#ifdef USE_PERLIO
b310b053	503	{
fde18df1	504	/* Set PL_utf8locale to TRUE if using PerlIO _and_
ec71e770	505	any of the following are true:
085a54d9	506	- nl_langinfo(CODESET) contains /^utf-?8/i
61de9fb5	507	- $ENV{LC_ALL} contains /^utf-?8/i
085a54d9	508	- $ENV{LC_CTYPE} contains /^utf-?8/i
61de9fb5 JH	509	- $ENV{LANG} contains /^utf-?8/i
	510	The LC_ALL, LC_CTYPE, LANG obey the usual override
	511	hierarchy of locale environment variables. (LANGUAGE
	512	affects only LC_MESSAGES only under glibc.) (If present,
	513	it overrides LC_MESSAGES for GNU gettext, and it also
	514	can have more than one locale, separated by spaces,
	515	in case you need to know.)
8aa8f774	516	If PL_utf8locale and PL_unicode (set by -C or by $ENV{PERL_UNICODE})
a05d7ebb	517	are true, perl.c:S_parse_body() will turn on the PerlIO :utf8 layer
fde18df1	518	on STDIN, STDOUT, STDERR, _and_ the default open discipline.
085a54d9	519	*/
fde18df1	520	bool utf8locale = FALSE;
b310b053 JH	521	char *codeset = NULL;
	522	#if defined(HAS_NL_LANGINFO) && defined(CODESET)
	523	codeset = nl_langinfo(CODESET);
	524	#endif
61de9fb5	525	if (codeset)
e6226b18 KW	526	utf8locale = (foldEQ(codeset, STR_WITH_LEN("UTF-8"))
e6226b18 KW	527	\|\| foldEQ(codeset, STR_WITH_LEN("UTF8") ));
077a72a9	528	#if defined(USE_LOCALE)
61de9fb5 JH	529	else { /* nl_langinfo(CODESET) is supposed to correctly
	530	* interpret the locale environment variables,
	531	* but just in case it fails, let's do this manually. */
	532	if (lang)
e6226b18 KW	533	utf8locale = (foldEQ(lang, STR_WITH_LEN("UTF-8"))
e6226b18 KW	534	\|\| foldEQ(lang, STR_WITH_LEN("UTF8") ));
b310b053	535	#ifdef USE_LOCALE_CTYPE
61de9fb5	536	if (curctype)
e6226b18 KW	537	utf8locale = (foldEQ(curctype, STR_WITH_LEN("UTF-8"))
e6226b18 KW	538	\|\| foldEQ(curctype, STR_WITH_LEN("UTF8") ));
b310b053	539	#endif
61de9fb5	540	if (lc_all)
e6226b18 KW	541	utf8locale = (foldEQ(lc_all, STR_WITH_LEN("UTF-8"))
e6226b18 KW	542	\|\| foldEQ(lc_all, STR_WITH_LEN("UTF8") ));
61de9fb5	543	}
fde18df1 JH	544	#endif /* USE_LOCALE */
	545	if (utf8locale)
	546	PL_utf8locale = TRUE;
	547	}
a05d7ebb	548	/* Set PL_unicode to $ENV{PERL_UNICODE} if using PerlIO.
fde18df1 JH	549	This is an alternative to using the -C command line switch
	550	(the -C if present will override this). */
	551	{
dd374669	552	const char *p = PerlEnv_getenv("PERL_UNICODE");
a05d7ebb	553	PL_unicode = p ? parse_unicode_opts(&p) : 0;
5a22a2bb NC	554	if (PL_unicode & PERL_UNICODE_UTF8CACHEASSERT_FLAG)
5a22a2bb NC	555	PL_utf8cache = -1;
b310b053	556	}
ec71e770	557	#endif
b310b053	558
98994639	559	#ifdef USE_LOCALE_CTYPE
43c5f42d	560	Safefree(curctype);
98994639 HS	561	#endif /* USE_LOCALE_CTYPE */
98994639 HS	562	#ifdef USE_LOCALE_COLLATE
43c5f42d	563	Safefree(curcoll);
98994639 HS	564	#endif /* USE_LOCALE_COLLATE */
98994639 HS	565	#ifdef USE_LOCALE_NUMERIC
43c5f42d	566	Safefree(curnum);
98994639 HS	567	#endif /* USE_LOCALE_NUMERIC */
	568	return ok;
	569	}
	570
98994639 HS	571	#ifdef USE_LOCALE_COLLATE
	572
	573	/*
	574	* mem_collxfrm() is a bit like strxfrm() but with two important
	575	* differences. First, it handles embedded NULs. Second, it allocates
	576	* a bit more memory than needed for the transformed data itself.
	577	* The real transformed data begins at offset sizeof(collationix).
	578	* Please see sv_collxfrm() to see how this is used.
	579	*/
166f8a29	580
98994639 HS	581	char *
	582	Perl_mem_collxfrm(pTHX_ const char s, STRLEN len, STRLEN xlen)
	583	{
97aff369	584	dVAR;
98994639 HS	585	char *xbuf;
	586	STRLEN xAlloc, xin, xout; /* xalloc is a reserved word in VC */
	587
7918f24d NC	588	PERL_ARGS_ASSERT_MEM_COLLXFRM;
7918f24d NC	589
98994639 HS	590	/* the first sizeof(collationix) bytes are used by sv_collxfrm(). */
	591	/* the +1 is for the terminating NUL. */
	592
	593	xAlloc = sizeof(PL_collation_ix) + PL_collxfrm_base + (PL_collxfrm_mult * len) + 1;
a02a5408	594	Newx(xbuf, xAlloc, char);
98994639 HS	595	if (! xbuf)
	596	goto bad;
	597
	598	(U32)xbuf = PL_collation_ix;
	599	xout = sizeof(PL_collation_ix);
	600	for (xin = 0; xin < len; ) {
224e8ef5	601	Size_t xused;
98994639 HS	602
	603	for (;;) {
	604	xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
224e8ef5	605	if (xused >= PERL_INT_MAX)
98994639	606	goto bad;
eb160463	607	if ((STRLEN)xused < xAlloc - xout)
98994639 HS	608	break;
	609	xAlloc = (2 * xAlloc) + 1;
	610	Renew(xbuf, xAlloc, char);
	611	if (! xbuf)
	612	goto bad;
	613	}
	614
	615	xin += strlen(s + xin) + 1;
	616	xout += xused;
	617
	618	/* Embedded NULs are understood but silently skipped
	619	* because they make no sense in locale collation. */
	620	}
	621
	622	xbuf[xout] = '\0';
	623	*xlen = xout - sizeof(PL_collation_ix);
	624	return xbuf;
	625
	626	bad:
	627	Safefree(xbuf);
	628	*xlen = 0;
	629	return NULL;
	630	}
	631
	632	#endif /* USE_LOCALE_COLLATE */
	633
66610fdd RGS	634	/*
	635	* Local variables:
	636	* c-indentation-style: bsd
	637	* c-basic-offset: 4
	638	* indent-tabs-mode: t
	639	* End:
	640	*
37442d52 RGS	641	* ex: set ts=8 sts=4 sw=4 noet:
37442d52 RGS	642	*/