533814 - Make posix locale claim to be nb-NO if it hits no_NO

Status: VERIFIED FIXED

(Core :: Internationalization, defect)

Description

[Axel Hecht [:Pike]]

The maemo device is sending the deprecated no_NO locale code for Norwegian, and we should map that to the best match we have, which is nb-NO.

Comment 1

[Simon Montagu :smontagu]

In what context exactly? Do you mean that GetSystemLocale should return nb-NO when the locale it gets from the OS is no-NO?

Comment 2

[Axel Hecht [:Pike]]

AFAICT, it boils down to GetXPLocale, http://mxr.mozilla.org/mozilla/source/intl/locale/src/unix/nsPosixLocale.cpp#100. It starts in the chrome registry, http://mxr.mozilla.org/mozilla-central/source/chrome/src/nsChromeRegistry.cpp#444.

But yes, if we see no-NO (actually, no_NO), report that as nb-NO instead of no-NO.

Comment 3

[Simon Montagu :smontagu]

Attachment: Patch

Not asking review yet because I have no way to test it this week, but attaching in case someone else wants to test.

Comment 4

[Axel Hecht [:Pike]]

Comment on attachment 417022 [details] [diff] [review]
Patch

AFAICT, this should happen on lang_code instead of country_code.

Comment 5

[Simon Montagu :smontagu]

Attachment: Patch v.2

Thanks, Axel! (the previous version wouldn't even have compiled because I left out the "_" in country_code)

Comment 6

[Simon Montagu :smontagu]

Attachment: Patch v.3

Third time's the charm...

Comment 7

[Axel Hecht [:Pike]]

Comment on attachment 417034 [details] [diff] [review]
Patch v.3

Looks good to me, any chance to test this yet?

Comment 8

[Simon Montagu :smontagu]

I've run into a catch-22 testing this: the only system that I have that uses posix locale is Ubuntu, where I can install nb_NO or nn_NO but no no_NO.

Comment 9

[Staś Małolepszy :stas]

Simon: I can test on n900 and n810 if you have a test build that I can run.

Comment 10

[Axel Hecht [:Pike]]

Comment on attachment 417034 [details] [diff] [review]
Patch v.3

stricmp doesn't compile on arm, trying with strcasecmp now. Not sure if that is more portable.

Comment 11

[Axel Hecht [:Pike]]

With strcasecmp, I got a working build that detects norks OK.

Do we have something in nspr or so to abstract the platform-dependent extensions?

Comment 12

[Simon Montagu :smontagu]

Attachment: Use nsDependentCString.LowerCaseEqualsLiteral

This should be portable enough

Comment 13

[Axel Hecht [:Pike]]

Comment on attachment 418717 [details] [diff] [review]
Use nsDependentCString.LowerCaseEqualsLiteral

Looks good to me. Can you run this through the try server?

Comment 14

[Simon Montagu :smontagu]

Pushed to try server in http://hg.mozilla.org/try/rev/9e4982f218f0

Comment 15

[Simon Montagu :smontagu]

Can someone test the build at https://build.mozilla.org/tryserver-builds/[e-mail]-try-9e4982f218f0/maemo/ ?

Comment 16

[Staś Małolepszy :stas]

I downloaded the build from comment 15, but it was a single-locale build (en-US). 

I tried to build myself yesterday and was able to get a working build with the patch from attachment 418717 [details] [diff] [review]: <http://l10n.mozilla.org/~stas/builds/fennec-1.0b6pre-20091221.multi.linux-gnueabi-arm.tar> (You may still need the xulrunner package from the url in comment 15.)

I ran make chrome-ab-CD for every locale in maemo-locales before running make package. With matchOS set to true, it correctly detected nb-NO when the OS was set to no-NO.

Looks like it's working :)

Comment 17

[Simon Montagu :smontagu]

http://hg.mozilla.org/mozilla-central/rev/195db4224a60

Comment 18

[Mike Beltzner [:beltzner, not reading bugmail]]

Comment on attachment 418717 [details] [diff] [review]
Use nsDependentCString.LowerCaseEqualsLiteral

a192=beltzner

Comment 19

[Simon Montagu :smontagu]

http://hg.mozilla.org/releases/mozilla-1.9.2/rev/7c5ea4f7b7ed

Comment 20

[Staś Małolepszy :stas]

Verified on Mozilla/5.0 (X11; U; Linux armv7l; en-US; rv:1.9.2b6pre) Gecko/20091223 Firefox/3.6b6pre Fennec/1.0b6pre running on N900 on Maemo 5 (1.2009.42-11.002) in Nowegian ("Norsk (Norge)")

$ locale   
LANG=no_NO
LC_CTYPE="no_NO"
LC_NUMERIC=no_NO
LC_TIME=no_NO
LC_COLLATE="no_NO"
LC_MONETARY=no_NO
LC_MESSAGES=no_NO
LC_PAPER=no_NO
LC_NAME=no_NO
LC_ADDRESS=no_NO
LC_TELEPHONE=no_NO
LC_MEASUREMENT=no_NO
LC_IDENTIFICATION=no_NO
LC_ALL=
$