Le Monday 12 September 2005 à 04:07, Masahito Omote a écrit:
> Hello,
>
> On Sun, 11 Sep 2005 20:15:55 +0200
> Alban Crequy <[EMAIL PROTECTED]> wrote:
>
> > all_locales() returned NULL. I don't know why. The patch add this:
> > > if (name == NULL) name = "";
> > so compose_localenames_from_im_lang() does not return NULL anymore.
>
> Please try without @euro suffix in LC_CTYPE and LANG. For example set
> LC_CTYPE=fr_FR.ISO8859-15
> , LC_CTYPE=fr_FR.UTF-8, and so on. For more details see locale_map[] global
> variables in
> locale.cpp.
Whatever I set, it does not work (I tried fr, fr_FR, fr_FR.ISO8859-1,
fr_FR.ISO8859-15, fr_FR.UTF-8, fr_BE.UTF-8, en_US, en_US.ISO8859-1,
en_US.UTF-8, en_US:UTF-8).
I added some debug printf:
----->8--------->8----------->8-----
--- xim/locale.cpp.ORIG 2005-09-11 22:41:35.000000000 +0200
+++ xim/locale.cpp 2005-09-11 22:43:46.000000000 +0200
@@ -297,6 +297,7 @@ get_valid_locales(const char *locales)
// locales is separated with ':'
while ((locale = strsep(&tmpp, ":")) != NULL) {
if (setlocale(LC_CTYPE, locale) != NULL) {
+ printf("DEBUG setlocale(%s) OK\n", locale);
asprintf(&validated, "%s:", locale);
len += strlen(validated);
if (valid_locales) {
@@ -309,6 +310,7 @@ get_valid_locales(const char *locales)
} else {
// retry with supplemental encodings
int i;
+ printf("DEBUG setlocale(%s) FAILURE\n", locale);
for (i = 0; locale_map[i].localename; i++) {
if (is_locale_included(locale_map[i].localename, locale))
break;
----->8--------->8----------->8-----
The result is that setlocale() always failed. Not a single "OK":
DEBUG setlocale(af_ZA) FAILURE
DEBUG setlocale(am_ET) FAILURE
DEBUG setlocale(ar_AA) FAILURE
DEBUG setlocale(ar_BH) FAILURE
DEBUG setlocale(ar_DZ) FAILURE
[...]
DEBUG setlocale(fr_FR) FAILURE
DEBUG setlocale(fr_BE) FAILURE
DEBUG setlocale(fr_CA) FAILURE
DEBUG setlocale(fr_CH) FAILURE
DEBUG setlocale(fr_LU) FAILURE
[...]
DEBUG setlocale(zh_HK) FAILURE
I read "man setlocale" which says:
« For a list of all supported locales, try "locale -a", cf. locale(1). »
So I tried:
$ locale -a
C
[EMAIL PROTECTED]
[EMAIL PROTECTED]
POSIX
The problem is that none of the 4 locales I have on my computer is in
the locale_map global variable.
In order to be sure that this is the problem, I tested the following
patch:
----->8--------->8----------->8-----
--- xim/locale.cpp.ORIG 2005-09-11 22:41:35.000000000 +0200
+++ xim/locale.cpp 2005-09-11 22:49:49.000000000 +0200
@@ -93,7 +93,7 @@ static struct {
{"fa", "fa_IR", "UTF-8:ISIRI-3342"},
{"fi", "fi_FI", "ISO8859-15:ISO8859-1:UTF-8"},
{"fo", "fo_FO", "ISO8859-1:ISO8859-15:UTF-8"},
- {"fr", "fr_FR:fr_BE:fr_CA:fr_CH:fr_LU", "ISO8859-1:ISO8859-15:UTF-8"},
+ {"fr", "fr_FR:[EMAIL PROTECTED]:fr_BE:fr_CA:fr_CH:fr_LU",
"ISO8859-1:ISO8859-15:UTF-8"},
{"ga", "ga_IE", "ISO8859-1:ISO8859-14:ISO8859-15"},
{"gd", "gd_GB", "ISO8859-1:ISO8859-14:ISO8859-15"},
{"gl", "gl_ES", "ISO8859-1:ISO8859-15:UTF-8"},
----->8--------->8----------->8-----
And it works: I get my debug printf:
DEBUG setlocale(fr_FR) FAILURE
DEBUG setlocale([EMAIL PROTECTED]) OK
And I don't have Segmentation fault anymore.
Should "[EMAIL PROTECTED]" be included somewhere like this? Or "C" or "POSIX"
must be included somewhere in the locale_map global variable? I am not
familiar with locales; I don't know what should be done.
Note that "[EMAIL PROTECTED]" exists in /usr/share/i18n/SUPPORTED and it is
distinct from "fr_FR".
$ grep fr_FR /usr/share/i18n/SUPPORTED
fr_FR.UTF-8 UTF-8
[EMAIL PROTECTED] UTF-8
fr_FR ISO-8859-1
[EMAIL PROTECTED] ISO-8859-15
Thanks,
Alban
