1228755 - ASSERTION: language code too short: '(len == 2) || (len == 3)', file intl/locale/unix/nsPosixLocale.cpp, line 126

Status: RESOLVED FIXED

(Core :: Internationalization, defect)

Description

[Mats Palmgren (inactive)]

STEPS TO REPRODUCE
1. (export LANG="C.UTF-8"; firefox )

ACTUAL RESULTS
ASSERTION: language code too short: '(len == 2) || (len == 3)', file intl/locale/unix/nsPosixLocale.cpp, line 126

BUILDS AND PLATFORMS TESTED
Bug occurs in a DEBUG trunk build on Linux.

FYI
LANG="C.UTF-8" is the value Ubuntu gives me with System Settings ->
Regional Settings -> Formats -> Region: Default (C)

Comment 1

[Simon Montagu :smontagu]

nsPosixLocale::GetXPLocale already special-cases the "C" locale. We should probably add special-casing for "C.UTF-8" as well.

Comment 2

[Mats Palmgren (inactive)]

Attachment: fix?

Perhaps like so?

Comment 3

[Simon Montagu :smontagu]

Comment on attachment 8693285 [details] [diff] [review]
fix?

Review of attachment 8693285 [details] [diff] [review]:
-----------------------------------------------------------------

Yes, I think this is the way to go.

Comment 4

[Pulsebot]

https://hg.mozilla.org/integration/mozilla-inbound/rev/fcda0da2ed2a

Comment 5

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/fcda0da2ed2a

Comment 6

[ISHIKAWA, Chiaki]

I think this patch is wrong on two counts ( Bug 981463 - Allow one letter "C" as in C.UTF-8 as a valid language code

1. This patch does not seem to work, AND MORE GRAVELY

2. We should not change language code to en_US when the user specifically sets it
to "C".

    I mean, instead of en-US.UTF-8we can assign "js_JP.UTF-8" or whatever, but
    that does not make sense. The user wanted to use "C" as defined in ISO-C or POSIX locale.
So I think using en-US.UTF-8 is quite improper here.

+    if (strcmp(posixLocale,"C.UTF-8")==0) {
+      locale.AssignLiteral("en-US.UTF-8");
+      return NS_OK;
+    }


As for point 1, did the patch work for the original poster?
Under what platform?

On Debian GNU/Linux, it does not work.
 
I prefer the approach in Bug 981463.

TIA

Comment 7

[ISHIKAWA, Chiaki]

A little more background can be found in Bug 960957

Comment 8

[ISHIKAWA, Chiaki]

Actually comment https://bugzilla.mozilla.org/show_bug.cgi?id=960957#c20
in 960957 explains why the patch in 981463 was created the way it was written.

TIA