[perl5.git] / locale.c

/*    locale.c
 *
 *    Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999,
 *    2000, 2001, 2002, 2003, 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 = (strEQ(newnum, "C") || 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 = (strEQ(newcoll, "C") || 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 JH	3	* Copyright (C) 1993, 1994, 1995, 1996, 1997, 1998, 1999,
4bb101f2 JH	4	* 2000, 2001, 2002, 2003, 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));
130	PL_numeric_standard = (strEQ(newnum, "C") \|\| strEQ(newnum, "POSIX"));
131	PL_numeric_local = TRUE;
132	set_numeric_radix();
133	}
134
135	#endif /* USE_LOCALE_NUMERIC */
136	}
137
138	void
139	Perl_set_numeric_standard(pTHX)
140	{
141	#ifdef USE_LOCALE_NUMERIC
142
143	if (! PL_numeric_standard) {
144	setlocale(LC_NUMERIC, "C");
145	PL_numeric_standard = TRUE;
146	PL_numeric_local = FALSE;
147	set_numeric_radix();
148	}
149
150	#endif /* USE_LOCALE_NUMERIC */
151	}
152
153	void
154	Perl_set_numeric_local(pTHX)
155	{
156	#ifdef USE_LOCALE_NUMERIC
157
158	if (! PL_numeric_local) {
159	setlocale(LC_NUMERIC, PL_numeric_name);
160	PL_numeric_standard = FALSE;
161	PL_numeric_local = TRUE;
162	set_numeric_radix();
163	}
164
165	#endif /* USE_LOCALE_NUMERIC */
166	}
167
168	/*
169	* Set up for a new ctype locale.
170	*/
171	void
172	Perl_new_ctype(pTHX_ char *newctype)
173	{
174	#ifdef USE_LOCALE_CTYPE
175
176	int i;
177
178	for (i = 0; i < 256; i++) {
179	if (isUPPER_LC(i))
180	PL_fold_locale[i] = toLOWER_LC(i);
181	else if (isLOWER_LC(i))
182	PL_fold_locale[i] = toUPPER_LC(i);
183	else
184	PL_fold_locale[i] = i;
185	}
186
187	#endif /* USE_LOCALE_CTYPE */
188	}
189
190	/*
191	* Set up for a new collation locale.
192	*/
193	void
194	Perl_new_collate(pTHX_ char *newcoll)
195	{
196	#ifdef USE_LOCALE_COLLATE
197
198	if (! newcoll) {
199	if (PL_collation_name) {
200	++PL_collation_ix;
201	Safefree(PL_collation_name);
202	PL_collation_name = NULL;
203	}
204	PL_collation_standard = TRUE;
205	PL_collxfrm_base = 0;
206	PL_collxfrm_mult = 2;
207	return;
208	}
209
210	if (! PL_collation_name \|\| strNE(PL_collation_name, newcoll)) {
211	++PL_collation_ix;
212	Safefree(PL_collation_name);
213	PL_collation_name = stdize_locale(savepv(newcoll));
214	PL_collation_standard = (strEQ(newcoll, "C") \|\| strEQ(newcoll, "POSIX"));
215
216	{
217	/* 2: at most so many chars ('a', 'b'). */
218	/* 50: surely no system expands a char more. */
219	#define XFRMBUFSIZE (2 * 50)
220	char xbuf[XFRMBUFSIZE];
221	Size_t fa = strxfrm(xbuf, "a", XFRMBUFSIZE);
222	Size_t fb = strxfrm(xbuf, "ab", XFRMBUFSIZE);
223	SSize_t mult = fb - fa;
224	if (mult < 1)
225	Perl_croak(aTHX_ "strxfrm() gets absurd");
eb160463	226	PL_collxfrm_base = (fa > (Size_t)mult) ? (fa - mult) : 0;
98994639 HS	227	PL_collxfrm_mult = mult;
	228	}
	229	}
	230
	231	#endif /* USE_LOCALE_COLLATE */
	232	}
	233
	234	/*
	235	* Initialize locale awareness.
	236	*/
	237	int
	238	Perl_init_i18nl10n(pTHX_ int printwarn)
	239	{
	240	int ok = 1;
	241	/* returns
	242	* 1 = set ok or not applicable,
	243	* 0 = fallback to C locale,
	244	* -1 = fallback to C locale failed
	245	*/
	246
	247	#if defined(USE_LOCALE)
	248
	249	#ifdef USE_LOCALE_CTYPE
	250	char *curctype = NULL;
	251	#endif /* USE_LOCALE_CTYPE */
	252	#ifdef USE_LOCALE_COLLATE
	253	char *curcoll = NULL;
	254	#endif /* USE_LOCALE_COLLATE */
	255	#ifdef USE_LOCALE_NUMERIC
	256	char *curnum = NULL;
	257	#endif /* USE_LOCALE_NUMERIC */
	258	#ifdef __GLIBC__
	259	char *language = PerlEnv_getenv("LANGUAGE");
	260	#endif
	261	char *lc_all = PerlEnv_getenv("LC_ALL");
	262	char *lang = PerlEnv_getenv("LANG");
	263	bool setlocale_failure = FALSE;
	264
	265	#ifdef LOCALE_ENVIRON_REQUIRED
	266
	267	/*
	268	* Ultrix setlocale(..., "") fails if there are no environment
	269	* variables from which to get a locale name.
	270	*/
	271
	272	bool done = FALSE;
	273
	274	#ifdef LC_ALL
	275	if (lang) {
	276	if (setlocale(LC_ALL, ""))
	277	done = TRUE;
	278	else
	279	setlocale_failure = TRUE;
	280	}
	281	if (!setlocale_failure) {
	282	#ifdef USE_LOCALE_CTYPE
	283	if (! (curctype =
	284	setlocale(LC_CTYPE,
	285	(!done && (lang \|\| PerlEnv_getenv("LC_CTYPE")))
	286	? "" : Nullch)))
	287	setlocale_failure = TRUE;
	288	else
	289	curctype = savepv(curctype);
	290	#endif /* USE_LOCALE_CTYPE */
291	#ifdef USE_LOCALE_COLLATE
292	if (! (curcoll =
293	setlocale(LC_COLLATE,
294	(!done && (lang \|\| PerlEnv_getenv("LC_COLLATE")))
295	? "" : Nullch)))
296	setlocale_failure = TRUE;
297	else
298	curcoll = savepv(curcoll);
299	#endif /* USE_LOCALE_COLLATE */
300	#ifdef USE_LOCALE_NUMERIC
301	if (! (curnum =
302	setlocale(LC_NUMERIC,
303	(!done && (lang \|\| PerlEnv_getenv("LC_NUMERIC")))
304	? "" : Nullch)))
305	setlocale_failure = TRUE;
306	else
307	curnum = savepv(curnum);
308	#endif /* USE_LOCALE_NUMERIC */
309	}
310
311	#endif /* LC_ALL */
312
313	#endif /* !LOCALE_ENVIRON_REQUIRED */
314
315	#ifdef LC_ALL
316	if (! setlocale(LC_ALL, ""))
317	setlocale_failure = TRUE;
318	#endif /* LC_ALL */
319
320	if (!setlocale_failure) {
321	#ifdef USE_LOCALE_CTYPE
322	if (! (curctype = setlocale(LC_CTYPE, "")))
323	setlocale_failure = TRUE;
324	else
325	curctype = savepv(curctype);
326	#endif /* USE_LOCALE_CTYPE */
327	#ifdef USE_LOCALE_COLLATE
328	if (! (curcoll = setlocale(LC_COLLATE, "")))
329	setlocale_failure = TRUE;
330	else
331	curcoll = savepv(curcoll);
332	#endif /* USE_LOCALE_COLLATE */
333	#ifdef USE_LOCALE_NUMERIC
334	if (! (curnum = setlocale(LC_NUMERIC, "")))
335	setlocale_failure = TRUE;
336	else
337	curnum = savepv(curnum);
338	#endif /* USE_LOCALE_NUMERIC */
339	}
340
341	if (setlocale_failure) {
342	char *p;
343	bool locwarn = (printwarn > 1 \|\|
344	(printwarn &&
345	(!(p = PerlEnv_getenv("PERL_BADLANG")) \|\| atoi(p))));
346
347	if (locwarn) {
348	#ifdef LC_ALL
349
350	PerlIO_printf(Perl_error_log,
351	"perl: warning: Setting locale failed.\n");
352
353	#else /* !LC_ALL */
354
355	PerlIO_printf(Perl_error_log,
356	"perl: warning: Setting locale failed for the categories:\n\t");
357	#ifdef USE_LOCALE_CTYPE
358	if (! curctype)
359	PerlIO_printf(Perl_error_log, "LC_CTYPE ");
360	#endif /* USE_LOCALE_CTYPE */
361	#ifdef USE_LOCALE_COLLATE
362	if (! curcoll)
363	PerlIO_printf(Perl_error_log, "LC_COLLATE ");
364	#endif /* USE_LOCALE_COLLATE */
365	#ifdef USE_LOCALE_NUMERIC
366	if (! curnum)
367	PerlIO_printf(Perl_error_log, "LC_NUMERIC ");
368	#endif /* USE_LOCALE_NUMERIC */
369	PerlIO_printf(Perl_error_log, "\n");
370
371	#endif /* LC_ALL */
372
373	PerlIO_printf(Perl_error_log,
374	"perl: warning: Please check that your locale settings:\n");
375
376	#ifdef __GLIBC__
377	PerlIO_printf(Perl_error_log,
378	"\tLANGUAGE = %c%s%c,\n",
379	language ? '"' : '(',
380	language ? language : "unset",
381	language ? '"' : ')');
382	#endif
383
384	PerlIO_printf(Perl_error_log,
385	"\tLC_ALL = %c%s%c,\n",
386	lc_all ? '"' : '(',
387	lc_all ? lc_all : "unset",
388	lc_all ? '"' : ')');
389
390	#if defined(USE_ENVIRON_ARRAY)
391	{
392	char **e;
393	for (e = environ; *e; e++) {
394	if (strnEQ(*e, "LC_", 3)
395	&& strnNE(*e, "LC_ALL=", 7)
396	&& (p = strchr(*e, '=')))
397	PerlIO_printf(Perl_error_log, "\t%.*s = \"%s\",\n",
398	(int)(p - e), e, p + 1);
399	}
400	}
401	#else
402	PerlIO_printf(Perl_error_log,
403	"\t(possibly more locale environment variables)\n");
404	#endif
405
406	PerlIO_printf(Perl_error_log,
407	"\tLANG = %c%s%c\n",
408	lang ? '"' : '(',
409	lang ? lang : "unset",
410	lang ? '"' : ')');
411
412	PerlIO_printf(Perl_error_log,
413	" are supported and installed on your system.\n");
414	}
415
416	#ifdef LC_ALL
417
418	if (setlocale(LC_ALL, "C")) {
419	if (locwarn)
420	PerlIO_printf(Perl_error_log,
421	"perl: warning: Falling back to the standard locale (\"C\").\n");
422	ok = 0;
423	}
424	else {
425	if (locwarn)
426	PerlIO_printf(Perl_error_log,
427	"perl: warning: Failed to fall back to the standard locale (\"C\").\n");
428	ok = -1;
429	}
430
431	#else /* ! LC_ALL */
432
433	if (0
434	#ifdef USE_LOCALE_CTYPE
435	\|\| !(curctype \|\| setlocale(LC_CTYPE, "C"))
436	#endif /* USE_LOCALE_CTYPE */
437	#ifdef USE_LOCALE_COLLATE
438	\|\| !(curcoll \|\| setlocale(LC_COLLATE, "C"))
439	#endif /* USE_LOCALE_COLLATE */
440	#ifdef USE_LOCALE_NUMERIC
441	\|\| !(curnum \|\| setlocale(LC_NUMERIC, "C"))
442	#endif /* USE_LOCALE_NUMERIC */
443	)
444	{
445	if (locwarn)
446	PerlIO_printf(Perl_error_log,
447	"perl: warning: Cannot fall back to the standard locale (\"C\").\n");
448	ok = -1;
449	}
450
451	#endif /* ! LC_ALL */
452
453	#ifdef USE_LOCALE_CTYPE
454	curctype = savepv(setlocale(LC_CTYPE, Nullch));
455	#endif /* USE_LOCALE_CTYPE */
456	#ifdef USE_LOCALE_COLLATE
457	curcoll = savepv(setlocale(LC_COLLATE, Nullch));
458	#endif /* USE_LOCALE_COLLATE */
459	#ifdef USE_LOCALE_NUMERIC
460	curnum = savepv(setlocale(LC_NUMERIC, Nullch));
461	#endif /* USE_LOCALE_NUMERIC */
462	}
463	else {
464
465	#ifdef USE_LOCALE_CTYPE
466	new_ctype(curctype);
467	#endif /* USE_LOCALE_CTYPE */
468
469	#ifdef USE_LOCALE_COLLATE
470	new_collate(curcoll);
471	#endif /* USE_LOCALE_COLLATE */
472
473	#ifdef USE_LOCALE_NUMERIC
474	new_numeric(curnum);
475	#endif /* USE_LOCALE_NUMERIC */
b310b053	476
98994639 HS	477	}
	478
	479	#endif /* USE_LOCALE */
	480
ec71e770	481	#ifdef USE_PERLIO
b310b053	482	{
fde18df1	483	/* Set PL_utf8locale to TRUE if using PerlIO _and_
ec71e770	484	any of the following are true:
085a54d9	485	- nl_langinfo(CODESET) contains /^utf-?8/i
61de9fb5	486	- $ENV{LC_ALL} contains /^utf-?8/i
085a54d9	487	- $ENV{LC_CTYPE} contains /^utf-?8/i
61de9fb5 JH	488	- $ENV{LANG} contains /^utf-?8/i
	489	The LC_ALL, LC_CTYPE, LANG obey the usual override
	490	hierarchy of locale environment variables. (LANGUAGE
	491	affects only LC_MESSAGES only under glibc.) (If present,
	492	it overrides LC_MESSAGES for GNU gettext, and it also
	493	can have more than one locale, separated by spaces,
	494	in case you need to know.)
8aa8f774	495	If PL_utf8locale and PL_unicode (set by -C or by $ENV{PERL_UNICODE})
a05d7ebb	496	are true, perl.c:S_parse_body() will turn on the PerlIO :utf8 layer
fde18df1	497	on STDIN, STDOUT, STDERR, _and_ the default open discipline.
085a54d9	498	*/
fde18df1	499	bool utf8locale = FALSE;
b310b053 JH	500	char *codeset = NULL;
	501	#if defined(HAS_NL_LANGINFO) && defined(CODESET)
	502	codeset = nl_langinfo(CODESET);
	503	#endif
61de9fb5	504	if (codeset)
fde18df1 JH	505	utf8locale = (ibcmp(codeset, "UTF-8", 5) == 0 \|\|
fde18df1 JH	506	ibcmp(codeset, "UTF8", 4) == 0);
077a72a9	507	#if defined(USE_LOCALE)
61de9fb5 JH	508	else { /* nl_langinfo(CODESET) is supposed to correctly
	509	* interpret the locale environment variables,
	510	* but just in case it fails, let's do this manually. */
	511	if (lang)
fde18df1 JH	512	utf8locale = (ibcmp(lang, "UTF-8", 5) == 0 \|\|
fde18df1 JH	513	ibcmp(lang, "UTF8", 4) == 0);
b310b053	514	#ifdef USE_LOCALE_CTYPE
61de9fb5	515	if (curctype)
fde18df1 JH	516	utf8locale = (ibcmp(curctype, "UTF-8", 5) == 0 \|\|
fde18df1 JH	517	ibcmp(curctype, "UTF8", 4) == 0);
b310b053	518	#endif
61de9fb5	519	if (lc_all)
fde18df1 JH	520	utf8locale = (ibcmp(lc_all, "UTF-8", 5) == 0 \|\|
fde18df1 JH	521	ibcmp(lc_all, "UTF8", 4) == 0);
61de9fb5	522	}
fde18df1 JH	523	#endif /* USE_LOCALE */
	524	if (utf8locale)
	525	PL_utf8locale = TRUE;
	526	}
a05d7ebb	527	/* Set PL_unicode to $ENV{PERL_UNICODE} if using PerlIO.
fde18df1 JH	528	This is an alternative to using the -C command line switch
	529	(the -C if present will override this). */
	530	{
a05d7ebb JH	531	char *p = PerlEnv_getenv("PERL_UNICODE");
a05d7ebb JH	532	PL_unicode = p ? parse_unicode_opts(&p) : 0;
b310b053	533	}
ec71e770	534	#endif
b310b053	535
98994639 HS	536	#ifdef USE_LOCALE_CTYPE
	537	if (curctype != NULL)
	538	Safefree(curctype);
	539	#endif /* USE_LOCALE_CTYPE */
	540	#ifdef USE_LOCALE_COLLATE
	541	if (curcoll != NULL)
	542	Safefree(curcoll);
	543	#endif /* USE_LOCALE_COLLATE */
	544	#ifdef USE_LOCALE_NUMERIC
	545	if (curnum != NULL)
	546	Safefree(curnum);
	547	#endif /* USE_LOCALE_NUMERIC */
	548	return ok;
	549	}
	550
	551	/* Backwards compatibility. */
	552	int
	553	Perl_init_i18nl14n(pTHX_ int printwarn)
	554	{
	555	return init_i18nl10n(printwarn);
	556	}
	557
	558	#ifdef USE_LOCALE_COLLATE
	559
	560	/*
	561	* mem_collxfrm() is a bit like strxfrm() but with two important
	562	* differences. First, it handles embedded NULs. Second, it allocates
	563	* a bit more memory than needed for the transformed data itself.
	564	* The real transformed data begins at offset sizeof(collationix).
	565	* Please see sv_collxfrm() to see how this is used.
	566	*/
166f8a29	567
98994639 HS	568	char *
	569	Perl_mem_collxfrm(pTHX_ const char s, STRLEN len, STRLEN xlen)
	570	{
	571	char *xbuf;
	572	STRLEN xAlloc, xin, xout; /* xalloc is a reserved word in VC */
	573
	574	/* the first sizeof(collationix) bytes are used by sv_collxfrm(). */
	575	/* the +1 is for the terminating NUL. */
	576
	577	xAlloc = sizeof(PL_collation_ix) + PL_collxfrm_base + (PL_collxfrm_mult * len) + 1;
	578	New(171, xbuf, xAlloc, char);
	579	if (! xbuf)
	580	goto bad;
	581
	582	(U32)xbuf = PL_collation_ix;
	583	xout = sizeof(PL_collation_ix);
	584	for (xin = 0; xin < len; ) {
	585	SSize_t xused;
	586
	587	for (;;) {
	588	xused = strxfrm(xbuf + xout, s + xin, xAlloc - xout);
	589	if (xused == -1)
	590	goto bad;
eb160463	591	if ((STRLEN)xused < xAlloc - xout)
98994639 HS	592	break;
	593	xAlloc = (2 * xAlloc) + 1;
	594	Renew(xbuf, xAlloc, char);
	595	if (! xbuf)
	596	goto bad;
	597	}
	598
	599	xin += strlen(s + xin) + 1;
	600	xout += xused;
	601
	602	/* Embedded NULs are understood but silently skipped
	603	* because they make no sense in locale collation. */
	604	}
	605
	606	xbuf[xout] = '\0';
	607	*xlen = xout - sizeof(PL_collation_ix);
	608	return xbuf;
	609
	610	bad:
	611	Safefree(xbuf);
	612	*xlen = 0;
	613	return NULL;
	614	}
	615
	616	#endif /* USE_LOCALE_COLLATE */
	617