Update from gnulib.
diff --git a/libcharset/ChangeLog b/libcharset/ChangeLog
index e71236b..638f43d 100644
--- a/libcharset/ChangeLog
+++ b/libcharset/ChangeLog
@@ -1,3 +1,14 @@
+2019-12-22 Bruno Haible <bruno@clisp.org>
+
+ Update from gnulib.
+ * include/localcharset.h.in: Update from gnulib.
+ 2019-12-20 Bruno Haible <bruno@clisp.org>
+ * localcharset.h: Mention which encodings are used as locale
+ encodings on z/OS.
+ 2019-12-17 Bruno Haible <bruno@clisp.org>
+ * localcharset.h (locale_charset): Clarify when the result
+ becomes invalid.
+
2019-12-02 Bruno Haible <bruno@clisp.org>
build: Fix build error with --enable-relocatable on macOS
diff --git a/libcharset/include/localcharset.h.in b/libcharset/include/localcharset.h.in
index e6f2627..f6f3938 100644
--- a/libcharset/include/localcharset.h.in
+++ b/libcharset/include/localcharset.h.in
@@ -1,5 +1,5 @@
/* Determine a canonical name for the current locale's character encoding.
- Copyright (C) 2000-2003, 2009-2018 Free Software Foundation, Inc.
+ Copyright (C) 2000-2003, 2009-2019 Free Software Foundation, Inc.
This file is part of the GNU CHARSET Library.
This program is free software; you can redistribute it and/or modify it
@@ -26,7 +26,10 @@
/* Determine the current locale's character encoding, and canonicalize it
into one of the canonical names listed below.
- The result must not be freed; it is statically allocated.
+ The result must not be freed; it is statically allocated. The result
+ becomes invalid when setlocale() is used to change the global locale, or
+ when the value of one of the environment variables LC_ALL, LC_CTYPE, LANG
+ is changed; threads in multithreaded programs should not do this.
If the canonical name cannot be determined, the result is a non-canonical
name. */
extern const char * locale_charset (void);
@@ -45,15 +48,15 @@
(darwin = Mac OS X, windows = native Windows)
ASCII, ANSI_X3.4-1968 glibc solaris freebsd netbsd darwin minix cygwin
- ISO-8859-1 Y glibc aix hpux irix osf solaris freebsd netbsd openbsd darwin cygwin
- ISO-8859-2 Y glibc aix hpux irix osf solaris freebsd netbsd openbsd darwin cygwin
+ ISO-8859-1 Y glibc aix hpux irix osf solaris freebsd netbsd openbsd darwin cygwin zos
+ ISO-8859-2 Y glibc aix hpux irix osf solaris freebsd netbsd openbsd darwin cygwin zos
ISO-8859-3 Y glibc solaris cygwin
ISO-8859-4 Y hpux osf solaris freebsd netbsd openbsd darwin
- ISO-8859-5 Y glibc aix hpux irix osf solaris freebsd netbsd openbsd darwin cygwin
+ ISO-8859-5 Y glibc aix hpux irix osf solaris freebsd netbsd openbsd darwin cygwin zos
ISO-8859-6 Y glibc aix hpux solaris cygwin
- ISO-8859-7 Y glibc aix hpux irix osf solaris freebsd netbsd openbsd darwin cygwin
- ISO-8859-8 Y glibc aix hpux osf solaris cygwin
- ISO-8859-9 Y glibc aix hpux irix osf solaris freebsd darwin cygwin
+ ISO-8859-7 Y glibc aix hpux irix osf solaris freebsd netbsd openbsd darwin cygwin zos
+ ISO-8859-8 Y glibc aix hpux osf solaris cygwin zos
+ ISO-8859-9 Y glibc aix hpux irix osf solaris freebsd darwin cygwin zos
ISO-8859-13 glibc hpux solaris freebsd netbsd openbsd darwin cygwin
ISO-8859-14 glibc cygwin
ISO-8859-15 glibc aix irix osf solaris freebsd netbsd openbsd darwin cygwin
@@ -76,7 +79,7 @@
CP874 windows dos
CP922 aix
CP932 aix cygwin windows dos
- CP943 aix
+ CP943 aix zos
CP949 osf darwin windows dos
CP950 windows dos
CP1046 aix
@@ -92,17 +95,17 @@
CP1255 glibc windows
CP1256 windows
CP1257 windows
- GB2312 Y glibc aix hpux irix solaris freebsd netbsd darwin cygwin
+ GB2312 Y glibc aix hpux irix solaris freebsd netbsd darwin cygwin zos
EUC-JP Y glibc aix hpux irix osf solaris freebsd netbsd darwin cygwin
- EUC-KR Y glibc aix hpux irix osf solaris freebsd netbsd darwin cygwin
+ EUC-KR Y glibc aix hpux irix osf solaris freebsd netbsd darwin cygwin zos
EUC-TW glibc aix hpux irix osf solaris netbsd
- BIG5 Y glibc aix hpux osf solaris freebsd netbsd darwin cygwin
+ BIG5 Y glibc aix hpux osf solaris freebsd netbsd darwin cygwin zos
BIG5-HKSCS glibc hpux solaris netbsd darwin
GBK glibc aix osf solaris freebsd darwin cygwin windows dos
GB18030 glibc hpux solaris freebsd netbsd darwin
SHIFT_JIS Y hpux osf solaris freebsd netbsd darwin
JOHAB glibc solaris windows
- TIS-620 glibc aix hpux osf solaris cygwin
+ TIS-620 glibc aix hpux osf solaris cygwin zos
VISCII Y glibc
TCVN5712-1 glibc
ARMSCII-8 glibc freebsd netbsd darwin
@@ -116,7 +119,7 @@
HP-KANA8 hpux
DEC-KANJI osf
DEC-HANYU osf
- UTF-8 Y glibc aix hpux osf solaris netbsd darwin cygwin
+ UTF-8 Y glibc aix hpux osf solaris netbsd darwin cygwin zos
Note: Names which are not marked as being a MIME name should not be used in
Internet protocols for information interchange (mail, news, etc.).
diff --git a/libcharset/lib/ChangeLog b/libcharset/lib/ChangeLog
index 2e08f78..5d57f1d 100644
--- a/libcharset/lib/ChangeLog
+++ b/libcharset/lib/ChangeLog
@@ -1,3 +1,21 @@
+2019-12-22 Bruno Haible <bruno@clisp.org>
+
+ Update from gnulib.
+ 2019-12-22 Bruno Haible <bruno@clisp.org>
+ localcharset: Update support for OpenBSD.
+ * localcharset.c (alias_table): Map "US-ASCII" to "ASCII".
+ 2019-12-17 Bruno Haible <bruno@clisp.org>
+ localcharset: Fix multithread-safety bug on Windows and OS/2.
+ * localcharset.c (locale_charset): Use a stack-allocated buffer
+ to assemble the result.
+ 2019-12-17 Bruno Haible <bruno@clisp.org>
+ localcharset: Optimize code for native Windows.
+ * localcharset.c (locale_charset): Don't bother calling
+ setlocale (LC_ALL, NULL) since we're not interested in its
+ result.
+ 2019-09-23 Paul Eggert <eggert@cs.ucla.edu>
+ Update URLs and associated text
+
2019-08-11 Bruno Haible <bruno@clisp.org>
Update from gnulib.
diff --git a/libcharset/lib/localcharset.c b/libcharset/lib/localcharset.c
index 8b2750f..fdef2a3 100644
--- a/libcharset/lib/localcharset.c
+++ b/libcharset/lib/localcharset.c
@@ -150,7 +150,8 @@
{ "ISO8859-2", "ISO-8859-2" },
{ "ISO8859-4", "ISO-8859-4" },
{ "ISO8859-5", "ISO-8859-5" },
- { "ISO8859-7", "ISO-8859-7" }
+ { "ISO8859-7", "ISO-8859-7" },
+ { "US-ASCII", "ASCII" }
# define alias_table_defined
# endif
# if defined __APPLE__ && defined __MACH__ /* Mac OS X */
@@ -378,8 +379,6 @@
by Alex Taylor:
<http://altsan.org/os2/toolkits/uls/index.html#codepages>.
See also "__convcp() of kLIBC":
- <http://trac.netlabs.org/libc/browser/branches/libc-0.6/src/emx/src/lib/locale/__convcp.c>,
- or:
<https://github.com/bitwiseworks/libc/blob/master/src/emx/src/lib/locale/__convcp.c>. */
{ "CP1004", "CP1252" },
/*{ "CP1041", "CP943" },*/
@@ -814,8 +813,11 @@
/* Determine the current locale's character encoding, and canonicalize it
- into one of the canonical names listed in localcharset.h.
- The result must not be freed; it is statically allocated.
+ into one of the canonical names listed below.
+ The result must not be freed; it is statically allocated. The result
+ becomes invalid when setlocale() is used to change the global locale, or
+ when the value of one of the environment variables LC_ALL, LC_CTYPE, LANG
+ is changed; threads in multithreaded programs should not do this.
If the canonical name cannot be determined, the result is a non-canonical
name. */
@@ -827,6 +829,13 @@
{
const char *codeset;
+ /* This function must be multithread-safe. To achieve this without using
+ thread-local storage, we use a simple strcpy or memcpy to fill this static
+ buffer. Filling it through, for example, strcpy + strcat would not be
+ guaranteed to leave the buffer's contents intact if another thread is
+ currently accessing it. If necessary, the contents is first assembled in
+ a stack-allocated buffer. */
+
#if HAVE_LANGINFO_CODESET || defined WINDOWS_NATIVE || defined OS2
# if HAVE_LANGINFO_CODESET
@@ -841,7 +850,7 @@
if (codeset != NULL && strcmp (codeset, "US-ASCII") == 0)
{
const char *locale;
- static char buf[2 + 10 + 1];
+ static char resultbuf[2 + 10 + 1];
locale = getenv ("LC_ALL");
if (locale == NULL || locale[0] == '\0')
@@ -865,11 +874,12 @@
modifier = strchr (dot, '@');
if (modifier == NULL)
return dot;
- if (modifier - dot < sizeof (buf))
+ if (modifier - dot < sizeof (resultbuf))
{
- memcpy (buf, dot, modifier - dot);
- buf [modifier - dot] = '\0';
- return buf;
+ /* This way of filling resultbuf is multithread-safe. */
+ memcpy (resultbuf, dot, modifier - dot);
+ resultbuf [modifier - dot] = '\0';
+ return resultbuf;
}
}
}
@@ -885,8 +895,13 @@
converting to GetConsoleOutputCP(). This leads to correct results,
except when SetConsoleOutputCP has been called and a raster font is
in use. */
- sprintf (buf, "CP%u", GetACP ());
- codeset = buf;
+ {
+ char buf[2 + 10 + 1];
+
+ sprintf (buf, "CP%u", GetACP ());
+ strcpy (resultbuf, buf);
+ codeset = resultbuf;
+ }
}
# endif
@@ -896,34 +911,28 @@
# elif defined WINDOWS_NATIVE
- static char buf[2 + 10 + 1];
+ char buf[2 + 10 + 1];
+ static char resultbuf[2 + 10 + 1];
/* The Windows API has a function returning the locale's codepage as
a number, but the value doesn't change according to what the
'setlocale' call specified. So we use it as a last resort, in
case the string returned by 'setlocale' doesn't specify the
codepage. */
- char *current_locale = setlocale (LC_ALL, NULL);
- char *pdot;
+ char *current_locale = setlocale (LC_CTYPE, NULL);
+ char *pdot = strrchr (current_locale, '.');
- /* If they set different locales for different categories,
- 'setlocale' will return a semi-colon separated list of locale
- values. To make sure we use the correct one, we choose LC_CTYPE. */
- if (strchr (current_locale, ';'))
- current_locale = setlocale (LC_CTYPE, NULL);
-
- pdot = strrchr (current_locale, '.');
if (pdot && 2 + strlen (pdot + 1) + 1 <= sizeof (buf))
sprintf (buf, "CP%s", pdot + 1);
else
{
/* The Windows API has a function returning the locale's codepage as a
- number: GetACP().
- When the output goes to a console window, it needs to be provided in
- GetOEMCP() encoding if the console is using a raster font, or in
- GetConsoleOutputCP() encoding if it is using a TrueType font.
- But in GUI programs and for output sent to files and pipes, GetACP()
- encoding is the best bet. */
+ number: GetACP().
+ When the output goes to a console window, it needs to be provided in
+ GetOEMCP() encoding if the console is using a raster font, or in
+ GetConsoleOutputCP() encoding if it is using a TrueType font.
+ But in GUI programs and for output sent to files and pipes, GetACP()
+ encoding is the best bet. */
sprintf (buf, "CP%u", GetACP ());
}
/* For a locale name such as "French_France.65001", in Windows 10,
@@ -931,12 +940,15 @@
if (strcmp (buf + 2, "65001") == 0 || strcmp (buf + 2, "utf8") == 0)
codeset = "UTF-8";
else
- codeset = buf;
+ {
+ strcpy (resultbuf, buf);
+ codeset = resultbuf;
+ }
# elif defined OS2
const char *locale;
- static char buf[2 + 10 + 1];
+ static char resultbuf[2 + 10 + 1];
ULONG cp[3];
ULONG cplen;
@@ -965,11 +977,12 @@
modifier = strchr (dot, '@');
if (modifier == NULL)
return dot;
- if (modifier - dot < sizeof (buf))
+ if (modifier - dot < sizeof (resultbuf))
{
- memcpy (buf, dot, modifier - dot);
- buf [modifier - dot] = '\0';
- return buf;
+ /* This way of filling resultbuf is multithread-safe. */
+ memcpy (resultbuf, dot, modifier - dot);
+ resultbuf [modifier - dot] = '\0';
+ return resultbuf;
}
}
@@ -985,8 +998,11 @@
codeset = "";
else
{
+ char buf[2 + 10 + 1];
+
sprintf (buf, "CP%u", cp[0]);
- codeset = buf;
+ strcpy (resultbuf, buf);
+ codeset = resultbuf;
}
}
