848297 - getSelectedLocale in chrome registry should allow overwritten provider

Status: RESOLVED FIXED

(Core :: Internationalization: Localization, defect)

Description

[Axel Hecht [:Pike]]

We want sparse localizations of toolkit for both Android and Firefox OS multilocale builds.

The easiest way to do that is to overload the files that are actually used with a different package, but then getSelectedLocale() is wrong. So is GetLocalesForPackage().

Let's allow to tell the chrome registry which provider to ask a different package by adding a pref branch.

Comment 1

[Axel Hecht [:Pike]]

Let's see what try thinks, https://tbpl.mozilla.org/?tree=Try&rev=cd9a6eb14d45

Comment 2

[Axel Hecht [:Pike]]

Attachment: add a prefbranch to override getSelectedLocale

I didn't get feedback from the Android folks yet, but I expect we'll end up needing this for b2g RSN, so requesting review.

Comment 3

[Benjamin Smedberg]

Comment on attachment 726393 [details] [diff] [review]
add a prefbranch to override getSelectedLocale

I don't want to hold the long-lived branch object, and we don't need the special safe-mode behavior. Please just have nsChromeRegistryChrome::OverrideLocalePackage use mozilla::Preferences directly and skip the branch goop.

Comment 4

[Axel Hecht [:Pike]]

Attachment: now with easier pref logic

Addressing the review comments, using the mozilla::Preferences api, don't store a pref branch.

Comment 5

[Benjamin Smedberg]

Comment on attachment 732328 [details] [diff] [review]
now with easier pref logic

This really sucks and if there isn't a bug filed on removing this hack, please file one. But it's ok for now.

Comment 6

[Axel Hecht [:Pike]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/21cd4d9e679b, and I'll file a follow-up next.

Comment 7

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/21cd4d9e679b

Comment 8

[Jeff Beatty [:gueroJeff]]

(In reply to Axel Hecht [:Pike] from comment #6)
> https://hg.mozilla.org/integration/mozilla-inbound/rev/21cd4d9e679b, and
> I'll file a follow-up next.

What is the follow up bug number?

Comment 9

[Axel Hecht [:Pike]]

I've filed two follow-ups, bug 869385 to not use prefs if we have a better storage for this data, and bug 869387 to rip out this code path completely once we support sparse localizations/l10n fallback at runtime.