Mac language/locale cleanup...

Mac language/locale cleanup
- Effectively revert revision 28193 (http://codereview.chromium.org/258037), this makes Mac match the other platform for what at it's core is used for the chrome concept of locale.
- For the ApplicationLanguage, the browser will end up with what Cocoa picks (same as before)
- All other process types will honor the language they got on the command line when starting up.
- When asked the apps language, have the same side effect as Windows and Linux has of pushing the language through to ICU also (so dates format right, etc.)
- During browser startup, if someone passed a language, bail because Mac can't support that.

TEST=The tips on the NTP and the dates on the history page are in the same language at the UI.
BUG=26856
BUG=22727
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=32515

[jungshik at Google, 2009-11-18 23:42:35]

LGTM

Thanks a lot for sorting this out !!!

http://codereview.chromium.org/399086/diff/16/20
File chrome/browser/browser_main_mac.mm (right):

http://codereview.chromium.org/399086/diff/16/20#newcode53
Line 53: "Prefences to change languages.";
Can we just support that? Being able to control the UI language via a command
line is handy for testing :-)
Well, Linux does not do that, either (although it effectively does with
'LC_ALL=ll_RR.UTF-8 chrome' or 'LANGUAGE=ll chrome'). 

It doesn't really matter much, but just a thought. Perhaps, if we do that, there
may be other conflicts (between Chrome's UI lang and a part of Chrome that
relies on System service). 

So, I'm perfectly ok with this.

[TVL, 2009-11-19 01:42:01]

Ok, hopefully I'm not jinxing things, but this looks like it is passing on Mac,
and I've sent it to win/linux just for the sake of it.

I had to change some of the approach from what I first uploaded because some
tests pass the lang flag to chrome, so we have to properly ignore those (just
like linux does).

[Mark Mentovai, 2009-11-19 01:58:21]

lg but...

http://codereview.chromium.org/399086/diff/33/2007
File app/l10n_util.cc (right):

http://codereview.chromium.org/399086/diff/33/2007#newcode489
Line 489: // Use any set override (Cocoa for the Browser), otherwise use the
command
remove "set"

un-capitalize Browser

http://codereview.chromium.org/399086/diff/33/2007#newcode497
Line 497: // The two above should handle all of the cases Chrome normally hits,
but
What if someday there are more (or less) than two above?  Just "the above" is
fine.

http://codereview.chromium.org/399086/diff/33/2007#newcode498
Line 498: // for some unittests, we need something to fall back too.
unit tests

http://codereview.chromium.org/399086/diff/33/2007#newcode504
Line 504: // on Mac.
Why not just call IsLocaleAvailable then?  Or CheckAndResolveLocale?

http://codereview.chromium.org/399086/diff/33/2006
File app/l10n_util_mac.mm (right):

http://codereview.chromium.org/399086/diff/33/2006#newcode15
Line 15: const std::string& Value() const { return value_; }
Name this value.

http://codereview.chromium.org/399086/diff/33/2006#newcode16
Line 16: void SetValue(const std::string override_value) { value_ =
override_value; }
Name this set_value.

http://codereview.chromium.org/399086/diff/33/2006#newcode18
Line 18: std::string value_;
DISALLOW_COPY_AND_ASSIGN(OverrideLocaleHolder);

http://codereview.chromium.org/399086/diff/33/2006#newcode21
Line 21: static base::LazyInstance<OverrideLocaleHolder>
You're in an anonymous namespace, you don't need to be static.

http://codereview.chromium.org/399086/diff/33/2006#newcode37
Line 37: // and language that can be set independently.  After talking with
Cole, the
Who is Cole?

http://codereview.chromium.org/399086/diff/33/2006#newcode38
Line 38: // best path from an experience pov is to map the Mac OS X language
into the
What is pov?

http://codereview.chromium.org/399086/diff/33/2006#newcode39
Line 39: // Chrome locale.  This way strings like "Yesterday", "Today" are in
the
"Yesterday" and "Today" ("and" instead of comma).

http://codereview.chromium.org/399086/diff/33/2006#newcode40
Line 40: // same language as raw dates "March 20, 1999" (strings vs ICU
formatted
raw dates like "March 20

http://codereview.chromium.org/399086/diff/33/2006#newcode41
Line 41: // data).  This also makes the Mac act like other Chrome platforms.
I don't know what "strings vs ICU formatted data" means.

http://codereview.chromium.org/399086/diff/33/2006#newcode41
Line 41: // data).  This also makes the Mac act like other Chrome platforms.
act -> acts

http://codereview.chromium.org/399086/diff/33/2006#newcode53
Line 53: locale_value = "en-US";
I'm unsure that this is cool.  Why is en-US right for an Aussie?

http://codereview.chromium.org/399086/diff/33/2009
File chrome/browser/browser_main_mac.mm (right):

http://codereview.chromium.org/399086/diff/33/2009#newcode11
Line 11: #include "app/app_switches.h"
Sort

http://codereview.chromium.org/399086/diff/33/2009#newcode44
Line 44: // The browser process only wants to support the language Cocoa says,
so force
Cocoa says?

http://codereview.chromium.org/399086/diff/33/2009#newcode45
Line 45: // the app locale to be overriden with what Cocoa says.
Cocoa says?