[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"

/*
 * 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)
{
    char *s;
    bool okay = TRUE;

    if ((s = strchr(locs, '='))) {
	char *t;

	okay = FALSE;
	if ((t = strchr(s, '.'))) {
	    char *u;

	    if ((u = strchr(t, '\n'))) {

		if (u[1] == 0) {
		    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;
}

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_ char *newnum)
{
#ifdef USE_LOCALE_NUMERIC

    if (! newnum) {
	if (PL_numeric_name) {
	    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_ char *newctype)
{
#ifdef USE_LOCALE_CTYPE

    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 */
}

/*
 * Set up for a new collation locale.
 */
void
Perl_new_collate(pTHX_ 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];
	  Size_t fa = strxfrm(xbuf, "a",  XFRMBUFSIZE);
	  Size_t fb = strxfrm(xbuf, "ab", XFRMBUFSIZE);
	  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). */
    {
	 char *p = PerlEnv_getenv("PERL_UNICODE");
	 PL_unicode = p ? parse_unicode_opts(&p) : 0;
    }
#endif

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

/* Backwards compatibility. */
int
Perl_init_i18nl14n(pTHX_ int printwarn)
{
    return init_i18nl10n(printwarn);
}

#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;
    New(171, 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 */
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
98994639 HS	39	/*
	40	* Standardize the locale name from a string returned by 'setlocale'.
	41	*
	42	* The standard return value of setlocale() is either
	43	* (1) "xx_YY" if the first argument of setlocale() is not LC_ALL
	44	* (2) "xa_YY xb_YY ..." if the first argument of setlocale() is LC_ALL
	45	* (the space-separated values represent the various sublocales,
	46	* in some unspecificed order)
	47	*
	48	* In some platforms it has a form like "LC_SOMETHING=Lang_Country.866\n",
	49	* which is harmful for further use of the string in setlocale().
	50	*
	51	*/
	52	STATIC char *
	53	S_stdize_locale(pTHX_ char *locs)
	54	{
	55	char *s;
	56	bool okay = TRUE;
	57
	58	if ((s = strchr(locs, '='))) {
	59	char *t;
	60
	61	okay = FALSE;
	62	if ((t = strchr(s, '.'))) {
	63	char *u;
	64
	65	if ((u = strchr(t, '\n'))) {
	66
	67	if (u[1] == 0) {
	68	STRLEN len = u - s;
	69	Move(s + 1, locs, len, char);
	70	locs[len] = 0;
	71	okay = TRUE;
	72	}
	73	}
	74	}
	75	}
	76
	77	if (!okay)
	78	Perl_croak(aTHX_ "Can't fix broken locale name \"%s\"", locs);
	79
	80	return locs;
	81	}
	82
	83	void
	84	Perl_set_numeric_radix(pTHX)
	85	{
	86	#ifdef USE_LOCALE_NUMERIC
	87	# ifdef HAS_LOCALECONV
	88	struct lconv* lc;
	89
	90	lc = localeconv();
	91	if (lc && lc->decimal_point) {
	92	if (lc->decimal_point[0] == '.' && lc->decimal_point[1] == 0) {
	93	SvREFCNT_dec(PL_numeric_radix_sv);
	94	PL_numeric_radix_sv = Nullsv;
	95	}
	96	else {
	97	if (PL_numeric_radix_sv)
	98	sv_setpv(PL_numeric_radix_sv, lc->decimal_point);
	99	else
	100	PL_numeric_radix_sv = newSVpv(lc->decimal_point, 0);
	101	}
	102	}
103	else
104	PL_numeric_radix_sv = Nullsv;
105	# endif /* HAS_LOCALECONV */
106	#endif /* USE_LOCALE_NUMERIC */
107	}
108
109	/*
110	* Set up for a new numeric locale.
111	*/
112	void
113	Perl_new_numeric(pTHX_ char *newnum)
114	{
115	#ifdef USE_LOCALE_NUMERIC
116
117	if (! newnum) {
118	if (PL_numeric_name) {
119	Safefree(PL_numeric_name);
120	PL_numeric_name = NULL;
121	}
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
	143
	144	if (! PL_numeric_standard) {
	145	setlocale(LC_NUMERIC, "C");
	146	PL_numeric_standard = TRUE;
	147	PL_numeric_local = FALSE;
	148	set_numeric_radix();
	149	}
	150
	151	#endif /* USE_LOCALE_NUMERIC */
	152	}
	153
	154	void
	155	Perl_set_numeric_local(pTHX)
	156	{
	157	#ifdef USE_LOCALE_NUMERIC
	158
	159	if (! PL_numeric_local) {
	160	setlocale(LC_NUMERIC, PL_numeric_name);
	161	PL_numeric_standard = FALSE;
	162	PL_numeric_local = TRUE;
	163	set_numeric_radix();
	164	}
	165
	166	#endif /* USE_LOCALE_NUMERIC */
	167	}
	168
	169	/*
	170	* Set up for a new ctype locale.
	171	*/
	172	void
	173	Perl_new_ctype(pTHX_ char *newctype)
	174	{
	175	#ifdef USE_LOCALE_CTYPE
	176
	177	int i;
	178
	179	for (i = 0; i < 256; i++) {
	180	if (isUPPER_LC(i))
	181	PL_fold_locale[i] = toLOWER_LC(i);
	182	else if (isLOWER_LC(i))
	183	PL_fold_locale[i] = toUPPER_LC(i);
	184	else
	185	PL_fold_locale[i] = i;
	186	}
	187
	188	#endif /* USE_LOCALE_CTYPE */
	189	}
	190
	191	/*
	192	* Set up for a new collation locale.
	193	*/
	194	void
	195	Perl_new_collate(pTHX_ char *newcoll)
196	{
197	#ifdef USE_LOCALE_COLLATE
198
199	if (! newcoll) {
200	if (PL_collation_name) {
201	++PL_collation_ix;
202	Safefree(PL_collation_name);
203	PL_collation_name = NULL;
204	}
205	PL_collation_standard = TRUE;
206	PL_collxfrm_base = 0;
207	PL_collxfrm_mult = 2;
208	return;
209	}
210
211	if (! PL_collation_name \|\| strNE(PL_collation_name, newcoll)) {
212	++PL_collation_ix;
213	Safefree(PL_collation_name);
214	PL_collation_name = stdize_locale(savepv(newcoll));
770526c1 NC	215	PL_collation_standard = ((*newcoll == 'C' && newcoll[1] == '\0')
770526c1 NC	216	\|\| strEQ(newcoll, "POSIX"));
98994639 HS	217
	218	{
	219	/* 2: at most so many chars ('a', 'b'). */
	220	/* 50: surely no system expands a char more. */
	221	#define XFRMBUFSIZE (2 * 50)
	222	char xbuf[XFRMBUFSIZE];
	223	Size_t fa = strxfrm(xbuf, "a", XFRMBUFSIZE);
	224	Size_t fb = strxfrm(xbuf, "ab", XFRMBUFSIZE);
	225	SSize_t mult = fb - fa;
	226	if (mult < 1)
	227	Perl_croak(aTHX_ "strxfrm() gets absurd");
eb160463	228	PL_collxfrm_base = (fa > (Size_t)mult) ? (fa - mult) : 0;
98994639 HS	229	PL_collxfrm_mult = mult;
	230	}
	231	}
	232
	233	#endif /* USE_LOCALE_COLLATE */
	234	}
	235
	236	/*
	237	* Initialize locale awareness.
	238	*/
	239	int
	240	Perl_init_i18nl10n(pTHX_ int printwarn)
	241	{
	242	int ok = 1;
	243	/* returns
	244	* 1 = set ok or not applicable,
	245	* 0 = fallback to C locale,
	246	* -1 = fallback to C locale failed
	247	*/
	248
	249	#if defined(USE_LOCALE)
	250
	251	#ifdef USE_LOCALE_CTYPE
	252	char *curctype = NULL;
	253	#endif /* USE_LOCALE_CTYPE */
	254	#ifdef USE_LOCALE_COLLATE
	255	char *curcoll = NULL;
	256	#endif /* USE_LOCALE_COLLATE */
	257	#ifdef USE_LOCALE_NUMERIC
	258	char *curnum = NULL;
	259	#endif /* USE_LOCALE_NUMERIC */
	260	#ifdef __GLIBC__
	261	char *language = PerlEnv_getenv("LANGUAGE");
	262	#endif
	263	char *lc_all = PerlEnv_getenv("LC_ALL");
	264	char *lang = PerlEnv_getenv("LANG");
	265	bool setlocale_failure = FALSE;
	266
	267	#ifdef LOCALE_ENVIRON_REQUIRED
	268
	269	/*
	270	* Ultrix setlocale(..., "") fails if there are no environment
	271	* variables from which to get a locale name.
	272	*/
	273
	274	bool done = FALSE;
	275
	276	#ifdef LC_ALL
	277	if (lang) {
	278	if (setlocale(LC_ALL, ""))
	279	done = TRUE;
	280	else
	281	setlocale_failure = TRUE;
	282	}
	283	if (!setlocale_failure) {
	284	#ifdef USE_LOCALE_CTYPE
	285	if (! (curctype =
	286	setlocale(LC_CTYPE,
	287	(!done && (lang \|\| PerlEnv_getenv("LC_CTYPE")))
	288	? "" : Nullch)))
	289	setlocale_failure = TRUE;
	290	else
	291	curctype = savepv(curctype);
	292	#endif /* USE_LOCALE_CTYPE */
293	#ifdef USE_LOCALE_COLLATE
294	if (! (curcoll =
295	setlocale(LC_COLLATE,
296	(!done && (lang \|\| PerlEnv_getenv("LC_COLLATE")))
297	? "" : Nullch)))
298	setlocale_failure = TRUE;
299	else
300	curcoll = savepv(curcoll);
301	#endif /* USE_LOCALE_COLLATE */
302	#ifdef USE_LOCALE_NUMERIC
303	if (! (curnum =
304	setlocale(LC_NUMERIC,
305	(!done && (lang \|\| PerlEnv_getenv("LC_NUMERIC")))
306	? "" : Nullch)))
307	setlocale_failure = TRUE;
308	else
309	curnum = savepv(curnum);
310	#endif /* USE_LOCALE_NUMERIC */
311	}
312
313	#endif /* LC_ALL */
314
315	#endif /* !LOCALE_ENVIRON_REQUIRED */
316
317	#ifdef LC_ALL
318	if (! setlocale(LC_ALL, ""))
319	setlocale_failure = TRUE;
320	#endif /* LC_ALL */
321
322	if (!setlocale_failure) {
323	#ifdef USE_LOCALE_CTYPE
324	if (! (curctype = setlocale(LC_CTYPE, "")))
325	setlocale_failure = TRUE;
326	else
327	curctype = savepv(curctype);
328	#endif /* USE_LOCALE_CTYPE */
329	#ifdef USE_LOCALE_COLLATE
330	if (! (curcoll = setlocale(LC_COLLATE, "")))
331	setlocale_failure = TRUE;
332	else
333	curcoll = savepv(curcoll);
334	#endif /* USE_LOCALE_COLLATE */
335	#ifdef USE_LOCALE_NUMERIC
336	if (! (curnum = setlocale(LC_NUMERIC, "")))
337	setlocale_failure = TRUE;
338	else
339	curnum = savepv(curnum);
340	#endif /* USE_LOCALE_NUMERIC */
341	}
342
343	if (setlocale_failure) {
344	char *p;
345	bool locwarn = (printwarn > 1 \|\|
346	(printwarn &&
347	(!(p = PerlEnv_getenv("PERL_BADLANG")) \|\| atoi(p))));
348
349	if (locwarn) {
350	#ifdef LC_ALL
351
352	PerlIO_printf(Perl_error_log,
353	"perl: warning: Setting locale failed.\n");
354
355	#else /* !LC_ALL */
356
357	PerlIO_printf(Perl_error_log,
358	"perl: warning: Setting locale failed for the categories:\n\t");
359	#ifdef USE_LOCALE_CTYPE
360	if (! curctype)
361	PerlIO_printf(Perl_error_log, "LC_CTYPE ");
362	#endif /* USE_LOCALE_CTYPE */
363	#ifdef USE_LOCALE_COLLATE
364	if (! curcoll)
365	PerlIO_printf(Perl_error_log, "LC_COLLATE ");
366	#endif /* USE_LOCALE_COLLATE */
367	#ifdef USE_LOCALE_NUMERIC
368	if (! curnum)
369	PerlIO_printf(Perl_error_log, "LC_NUMERIC ");
370	#endif /* USE_LOCALE_NUMERIC */
371	PerlIO_printf(Perl_error_log, "\n");
372
373	#endif /* LC_ALL */
374
375	PerlIO_printf(Perl_error_log,
376	"perl: warning: Please check that your locale settings:\n");
377
378	#ifdef __GLIBC__
379	PerlIO_printf(Perl_error_log,
380	"\tLANGUAGE = %c%s%c,\n",
381	language ? '"' : '(',
382	language ? language : "unset",
383	language ? '"' : ')');
384	#endif
385
386	PerlIO_printf(Perl_error_log,
387	"\tLC_ALL = %c%s%c,\n",
388	lc_all ? '"' : '(',
389	lc_all ? lc_all : "unset",
390	lc_all ? '"' : ')');
391
392	#if defined(USE_ENVIRON_ARRAY)
393	{
394	char **e;
395	for (e = environ; *e; e++) {
396	if (strnEQ(*e, "LC_", 3)
397	&& strnNE(*e, "LC_ALL=", 7)
398	&& (p = strchr(*e, '=')))
399	PerlIO_printf(Perl_error_log, "\t%.*s = \"%s\",\n",
400	(int)(p - e), e, p + 1);
401	}
402	}
403	#else
404	PerlIO_printf(Perl_error_log,
405	"\t(possibly more locale environment variables)\n");
406	#endif
407
408	PerlIO_printf(Perl_error_log,
409	"\tLANG = %c%s%c\n",
410	lang ? '"' : '(',
411	lang ? lang : "unset",
412	lang ? '"' : ')');
413
414	PerlIO_printf(Perl_error_log,
415	" are supported and installed on your system.\n");
416	}
417
418	#ifdef LC_ALL
419
420	if (setlocale(LC_ALL, "C")) {
421	if (locwarn)
422	PerlIO_printf(Perl_error_log,
423	"perl: warning: Falling back to the standard locale (\"C\").\n");
424	ok = 0;
425	}
426	else {
427	if (locwarn)
428	PerlIO_printf(Perl_error_log,
429	"perl: warning: Failed to fall back to the standard locale (\"C\").\n");
430	ok = -1;
431	}
432
433	#else /* ! LC_ALL */
434
435	if (0
436	#ifdef USE_LOCALE_CTYPE
437	\|\| !(curctype \|\| setlocale(LC_CTYPE, "C"))
438	#endif /* USE_LOCALE_CTYPE */
439	#ifdef USE_LOCALE_COLLATE
440	\|\| !(curcoll \|\| setlocale(LC_COLLATE, "C"))
441	#endif /* USE_LOCALE_COLLATE */
442	#ifdef USE_LOCALE_NUMERIC
443	\|\| !(curnum \|\| setlocale(LC_NUMERIC, "C"))
444	#endif /* USE_LOCALE_NUMERIC */
445	)
446	{
447	if (locwarn)
448	PerlIO_printf(Perl_error_log,
449	"perl: warning: Cannot fall back to the standard locale (\"C\").\n");
450	ok = -1;
451	}
452
453	#endif /* ! LC_ALL */
454
455	#ifdef USE_LOCALE_CTYPE
456	curctype = savepv(setlocale(LC_CTYPE, Nullch));
457	#endif /* USE_LOCALE_CTYPE */
458	#ifdef USE_LOCALE_COLLATE
459	curcoll = savepv(setlocale(LC_COLLATE, Nullch));
460	#endif /* USE_LOCALE_COLLATE */
461	#ifdef USE_LOCALE_NUMERIC
462	curnum = savepv(setlocale(LC_NUMERIC, Nullch));
463	#endif /* USE_LOCALE_NUMERIC */
464	}
465	else {
466
467	#ifdef USE_LOCALE_CTYPE
468	new_ctype(curctype);
469	#endif /* USE_LOCALE_CTYPE */
470
471	#ifdef USE_LOCALE_COLLATE
472	new_collate(curcoll);
473	#endif /* USE_LOCALE_COLLATE */
474
475	#ifdef USE_LOCALE_NUMERIC
476	new_numeric(curnum);
477	#endif /* USE_LOCALE_NUMERIC */
b310b053	478
98994639 HS	479	}
	480
	481	#endif /* USE_LOCALE */
	482
ec71e770	483	#ifdef USE_PERLIO
b310b053	484	{
fde18df1	485	/* Set PL_utf8locale to TRUE if using PerlIO _and_
ec71e770	486	any of the following are true:
085a54d9	487	- nl_langinfo(CODESET) contains /^utf-?8/i
61de9fb5	488	- $ENV{LC_ALL} contains /^utf-?8/i
085a54d9	489	- $ENV{LC_CTYPE} contains /^utf-?8/i
61de9fb5 JH	490	- $ENV{LANG} contains /^utf-?8/i
	491	The LC_ALL, LC_CTYPE, LANG obey the usual override
	492	hierarchy of locale environment variables. (LANGUAGE
	493	affects only LC_MESSAGES only under glibc.) (If present,
	494	it overrides LC_MESSAGES for GNU gettext, and it also
	495	can have more than one locale, separated by spaces,
	496	in case you need to know.)
8aa8f774	497	If PL_utf8locale and PL_unicode (set by -C or by $ENV{PERL_UNICODE})
a05d7ebb	498	are true, perl.c:S_parse_body() will turn on the PerlIO :utf8 layer
fde18df1	499	on STDIN, STDOUT, STDERR, _and_ the default open discipline.
085a54d9	500	*/
fde18df1	501	bool utf8locale = FALSE;
b310b053 JH	502	char *codeset = NULL;
	503	#if defined(HAS_NL_LANGINFO) && defined(CODESET)
	504	codeset = nl_langinfo(CODESET);
	505	#endif
61de9fb5	506	if (codeset)
fde18df1 JH	507	utf8locale = (ibcmp(codeset, "UTF-8", 5) == 0 \|\|
fde18df1 JH	508	ibcmp(codeset, "UTF8", 4) == 0);
077a72a9	509	#if defined(USE_LOCALE)
61de9fb5 JH	510	else { /* nl_langinfo(CODESET) is supposed to correctly
	511	* interpret the locale environment variables,
	512	* but just in case it fails, let's do this manually. */
	513	if (lang)
fde18df1 JH	514	utf8locale = (ibcmp(lang, "UTF-8", 5) == 0 \|\|
fde18df1 JH	515	ibcmp(lang, "UTF8", 4) == 0);
b310b053	516	#ifdef USE_LOCALE_CTYPE
61de9fb5	517	if (curctype)
fde18df1 JH	518	utf8locale = (ibcmp(curctype, "UTF-8", 5) == 0 \|\|
fde18df1 JH	519	ibcmp(curctype, "UTF8", 4) == 0);
b310b053	520	#endif
61de9fb5	521	if (lc_all)
fde18df1 JH	522	utf8locale = (ibcmp(lc_all, "UTF-8", 5) == 0 \|\|
fde18df1 JH	523	ibcmp(lc_all, "UTF8", 4) == 0);
61de9fb5	524	}
fde18df1 JH	525	#endif /* USE_LOCALE */
	526	if (utf8locale)
	527	PL_utf8locale = TRUE;
	528	}
a05d7ebb	529	/* Set PL_unicode to $ENV{PERL_UNICODE} if using PerlIO.
fde18df1 JH	530	This is an alternative to using the -C command line switch
	531	(the -C if present will override this). */
	532	{
a05d7ebb JH	533	char *p = PerlEnv_getenv("PERL_UNICODE");
a05d7ebb JH	534	PL_unicode = p ? parse_unicode_opts(&p) : 0;
b310b053	535	}
ec71e770	536	#endif
b310b053	537
98994639 HS	538	#ifdef USE_LOCALE_CTYPE
	539	if (curctype != NULL)
	540	Safefree(curctype);
	541	#endif /* USE_LOCALE_CTYPE */
	542	#ifdef USE_LOCALE_COLLATE
	543	if (curcoll != NULL)
	544	Safefree(curcoll);
	545	#endif /* USE_LOCALE_COLLATE */
	546	#ifdef USE_LOCALE_NUMERIC
	547	if (curnum != NULL)
	548	Safefree(curnum);
	549	#endif /* USE_LOCALE_NUMERIC */
	550	return ok;
	551	}
	552
	553	/* Backwards compatibility. */
	554	int
	555	Perl_init_i18nl14n(pTHX_ int printwarn)
	556	{
	557	return init_i18nl10n(printwarn);
	558	}
	559
	560	#ifdef USE_LOCALE_COLLATE
	561
	562	/*
	563	* mem_collxfrm() is a bit like strxfrm() but with two important
	564	* differences. First, it handles embedded NULs. Second, it allocates
	565	* a bit more memory than needed for the transformed data itself.
	566	* The real transformed data begins at offset sizeof(collationix).
	567	* Please see sv_collxfrm() to see how this is used.
	568	*/
166f8a29	569
98994639 HS	570	char *
	571	Perl_mem_collxfrm(pTHX_ const char s, STRLEN len, STRLEN xlen)
	572	{
	573	char *xbuf;
	574	STRLEN xAlloc, xin, xout; /* xalloc is a reserved word in VC */
	575
	576	/* the first sizeof(collationix) bytes are used by sv_collxfrm(). */
	577	/* the +1 is for the terminating NUL. */
	578
	579	xAlloc = sizeof(PL_collation_ix) + PL_collxfrm_base + (PL_collxfrm_mult * len) + 1;
	580	New(171, xbuf, xAlloc, char);
	581	if (! xbuf)
	582	goto bad;
	583
	584	(U32)xbuf = PL_collation_ix;
	585	xout = sizeof(PL_collation_ix);
	586	for (xin = 0; xin < len; ) {
	587	SSize_t xused;
	588
	589	for (;;) {
	590	xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
	591	if (xused == -1)
	592	goto bad;
eb160463	593	if ((STRLEN)xused < xAlloc - xout)
98994639 HS	594	break;
	595	xAlloc = (2 * xAlloc) + 1;
	596	Renew(xbuf, xAlloc, char);
	597	if (! xbuf)
	598	goto bad;
	599	}
	600
	601	xin += strlen(s + xin) + 1;
	602	xout += xused;
	603
	604	/* Embedded NULs are understood but silently skipped
	605	* because they make no sense in locale collation. */
	606	}
	607
	608	xbuf[xout] = '\0';
	609	*xlen = xout - sizeof(PL_collation_ix);
	610	return xbuf;
	611
	612	bad:
	613	Safefree(xbuf);
	614	*xlen = 0;
	615	return NULL;
	616	}
	617
	618	#endif /* USE_LOCALE_COLLATE */
	619