Bug 76138 – Provide access to user's list of preferred languages

• Status: RESOLVED
• Resolution: FIXED
• Component: WebCore Misc.

[Eric Carlson]

On some platforms, OS X at least, the user's language preference is an ordered list of languages. As noted in https://bugs.webkit.org/show_bug.cgi?id=74121, the <track> needs to consider the user's language preferences.

[Radar WebKit Bug Importer]

<rdar://problem/10681343>

[Eric Carlson]

Created attachment 122535 [details]
Proposed patch

A patch for the EWS bots to chew on because I have *no* idea if the custom V8 code is correct.

[WebKit Review Bot]

Attachment 122535 [details] did not pass style-queue:

Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'LayoutTests/ChangeLog', u'LayoutTests/fast..." exit_code: 1

Source/WebCore/platform/win/LanguageWin.cpp:31:  Alphabetical sorting problem.  [build/include_order] [4]
Source/WebCore/platform/gtk/LanguageGtk.cpp:26:  Alphabetical sorting problem.  [build/include_order] [4]
Source/WebCore/ChangeLog:42:  Need whitespace between colon and description  [changelog/filechangedescriptionwhitespace] [5]
Source/WebCore/ChangeLog:46:  Need whitespace between colon and description  [changelog/filechangedescriptionwhitespace] [5]
Source/WebCore/testing/js/JSInternalsCustom.cpp:27:  You should add a blank line after implementation file's own header.  [build/include_order] [4]
Total errors found: 5 in 28 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Early Warning System Bot]

Comment on attachment 122535 [details]
Proposed patch

Attachment 122535 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/11149843

[Gustavo Noronha (kov)]

Comment on attachment 122535 [details]
Proposed patch

Attachment 122535 [details] did not pass gtk-ews (gtk):
Output: http://queues.webkit.org/results/11209818

[WebKit Review Bot]

Comment on attachment 122535 [details]
Proposed patch

Attachment 122535 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/11228824

[Alexey Proskuryakov]

Comment on attachment 122535 [details]
Proposed patch

Can we remove platformDefaultLanguage() now, making defaultLanguage() take the first one from the vector?

[Eric Carlson]

(In reply to comment #7)
> (From update of attachment 122535 [details])
> Can we remove platformDefaultLanguage() now, making defaultLanguage() take the first one from the vector?

Great idea, I should have thought of that!

[Eric Carlson]

Created attachment 122711 [details]
Updated patch

[Gyuyoung Kim]

Comment on attachment 122711 [details]
Updated patch

Attachment 122711 [details] did not pass efl-ews (efl):
Output: http://queues.webkit.org/results/11117079

[Early Warning System Bot]

Comment on attachment 122711 [details]
Updated patch

Attachment 122711 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/11117082

[Eric Carlson]

Created attachment 122712 [details]
Updated the poorly chosen function name.

[Early Warning System Bot]

Comment on attachment 122712 [details]
Updated the poorly chosen function name.

Attachment 122712 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/11111071

[Collabora GTK+ EWS bot]

Comment on attachment 122712 [details]
Updated the poorly chosen function name.

Attachment 122712 [details] did not pass gtk-ews (gtk):
Output: http://queues.webkit.org/results/11111085

[WebKit Review Bot]

Comment on attachment 122712 [details]
Updated the poorly chosen function name.

Attachment 122712 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/11269099

[Eric Carlson]

Created attachment 122770 [details]
A few code fixes, a few stab-in-the-dark build file changes. Maybe this one will fly.

[WebKit Review Bot]

Comment on attachment 122770 [details]
A few code fixes, a few stab-in-the-dark build file changes. Maybe this one will fly.

Attachment 122770 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/11267210

[Early Warning System Bot]

Comment on attachment 122770 [details]
A few code fixes, a few stab-in-the-dark build file changes. Maybe this one will fly.

Attachment 122770 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/11220018

[Eric Carlson]

Created attachment 122775 [details]
+symbols.filter

[Early Warning System Bot]

Comment on attachment 122775 [details]
+symbols.filter

Attachment 122775 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/11268227

[WebKit Review Bot]

Comment on attachment 122775 [details]
+symbols.filter

Attachment 122775 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/11262224

[Eric Carlson]

Created attachment 122777 [details]
Add new files to Target.pri (thanks Ossy!), fix typo in custom V8 code.

[Eric Carlson]

Created attachment 122783 [details]
+ WebCoreTestSupport.vcproj

[Alexey Proskuryakov]

Can we get rid of overrideDefaultLanguage() now, as well? It's confusing to have two overrides for basically the same thing.

[Adam Roben (:aroben)]

From WinEWS:

11>WebKit.exp : error LNK2001: unresolved external symbol "public: void __thiscall WebCore::JSInternals::setUserPreferredLanguages(class JSC::ExecState *,class JSC::JSValue)" (?setUserPreferredLanguages@JSInternals@WebCore@@QAEXPAVExecState@JSC@@VJSValue@4@@Z)
11>WebKit.exp : error LNK2001: unresolved external symbol "public: class JSC::JSValue __thiscall WebCore::JSInternals::userPreferredLanguages(class JSC::ExecState *)const " (?userPreferredLanguages@JSInternals@WebCore@@QBE?AVJSValue@JSC@@PAVExecState@4@@Z)

Those mangled symbols need to be added to WebKit2.def. There's a section at the bottom of that file that contains all the Internals symbols we export.

[Adam Roben (:aroben)]

(In reply to comment #25)
> From WinEWS:
> 
> 11>WebKit.exp : error LNK2001: unresolved external symbol "public: void __thiscall WebCore::JSInternals::setUserPreferredLanguages(class JSC::ExecState *,class JSC::JSValue)" (?setUserPreferredLanguages@JSInternals@WebCore@@QAEXPAVExecState@JSC@@VJSValue@4@@Z)
> 11>WebKit.exp : error LNK2001: unresolved external symbol "public: class JSC::JSValue __thiscall WebCore::JSInternals::userPreferredLanguages(class JSC::ExecState *)const " (?userPreferredLanguages@JSInternals@WebCore@@QBE?AVJSValue@JSC@@PAVExecState@4@@Z)
> 
> Those mangled symbols need to be added to WebKit2.def. There's a section at the bottom of that file that contains all the Internals symbols we export.

Oh, wait, I misread the error. These symbols are failing to be found when linking WebKit.dll. This means that the code that provides those symbols isn't being compiled for some reason.

[WebKit Review Bot]

Comment on attachment 122783 [details]
+ WebCoreTestSupport.vcproj

Attachment 122783 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/11117273

New failing tests:
inspector/protocol/runtime-agent.html

[Adam Roben (:aroben)]

It looks like those symbols correspond to functions that are in WebCoreTestSupport.lib. Those symbols shouldn't be listed in WebKit2[CFLite].def at all.

[Eric Carlson]

Created attachment 122795 [details]
One more time...

[Eric Carlson]

Created attachment 122869 [details]
One last version, got rid of overrideDefaultLanguage() as per request

[Alexey Proskuryakov]

Comment on attachment 122869 [details]
One last version, got rid of overrideDefaultLanguage() as per request

This still has separate override variables (preferredLanguagesOverride() and languageOverride()). Is that unavoidable?

[Eric Carlson]

(In reply to comment #31)
> (From update of attachment 122869 [details])
> This still has separate override variables (preferredLanguagesOverride() and languageOverride()). Is that unavoidable?

No, it is just me attaching an old version of the patch :-(

New version coming up.

[Eric Carlson]

Created attachment 122951 [details]
Updated patch.

[Alexey Proskuryakov]

Comment on attachment 122951 [details]
Updated patch.

View in context: https://bugs.webkit.org/attachment.cgi?id=122951&action=review

Looks great, thank you!

> Source/WebCore/platform/Language.h:30
>  #include <wtf/Forward.h>
> +#include <wtf/Vector.h>

I'm surprised that this was needed - Forward.h declares Vector.

[Alexey Proskuryakov]

It's unfortunate that EWS couldn't apply this patch.

[Eric Carlson]

(In reply to comment #35)
> It's unfortunate that EWS couldn't apply this patch.

I agree. Earlier versions of this patch bounced enough times that I am going to rebase and upload again so I can make sure it works.

[Eric Carlson]

Created attachment 122973 [details]
Rebased patch for the EWS bots.

[Eric Carlson]

http://trac.webkit.org/changeset/105315

[Eric Carlson]

Plus http://trac.webkit.org/changeset/105333 to fix a bug in the original patch.

[Gabor Rapcsanyi]

It breaks these bots:
 http://build.webkit.org/waterfall?show=GTK%20Linux%2032-bit%20Release
 http://build.webkit.sed.hu/waterfall?show=ARMv7%20Linux%20Gtk%20Release%20(Build)


Can somebody fix this error?

CXXLD  Programs/DumpRenderTree
./.libs/libWebCoreInternals.a(libWebCoreInternals_la-Internals.o): In function `WebCore::Internals::setUserPreferredLanguages(WTF::Vector<WTF::String, 0u> const&)':
Internals.cpp:(.text+0x78): undefined reference to `WebCore::overrideUserPreferredLanguages(WTF::Vector<WTF::String, 0u> const&)'
collect2: ld returned 1 exit status
make[1]: *** [Programs/DumpRenderTree] Error 1
make[1]: Leaving directory `/ramdisk/arm-gtk-linux-release/build/WebKitBuild/Release'
make: *** [all] Error 2

[Eric Carlson]

(In reply to comment #40)
> It breaks these bots:
>  http://build.webkit.org/waterfall?show=GTK%20Linux%2032-bit%20Release
>  http://build.webkit.sed.hu/waterfall?show=ARMv7%20Linux%20Gtk%20Release%20(Build)
> 
> 
> Can somebody fix this error?
> 
> CXXLD  Programs/DumpRenderTree
> ./.libs/libWebCoreInternals.a(libWebCoreInternals_la-Internals.o): In function `WebCore::Internals::setUserPreferredLanguages(WTF::Vector<WTF::String, 0u> const&)':
> Internals.cpp:(.text+0x78): undefined reference to `WebCore::overrideUserPreferredLanguages(WTF::Vector<WTF::String, 0u> const&)'
> collect2: ld returned 1 exit status
> make[1]: *** [Programs/DumpRenderTree] Error 1
> make[1]: Leaving directory `/ramdisk/arm-gtk-linux-release/build/WebKitBuild/Release'
> make: *** [all] Error 2

I don't understand this, what would make the new WebCore symbol available in a Debug build but not in a Release build?