Bug 163222 – Add support for languagechange event

• Component: DOM
• Status: RESOLVED FIXED

[Chris Dumez]

Add support for languagechange event:
- https://html.spec.whatwg.org/#dom-navigator-languages

[Chris Dumez]

Created attachment 291184 [details]
Patch

[Chris Dumez]

Created attachment 291185 [details]
Patch

[Ryosuke Niwa]

Comment on attachment 291185 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=291185&action=review

> Source/WebCore/page/DOMWindow.cpp:426
> +
> +    addLanguageChangeObserver(this, &languagesChangedCallback);

This would be racy with regards with the cache being invalidated in FontGenericFamilies.
We need to either force invalidating the cache in there first
or add some other mechanism to make sure this observer is called at the very end or very first
so that there is no observable race conditions.

[Chris Dumez]

Comment on attachment 291185 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=291185&action=review

>> Source/WebCore/page/DOMWindow.cpp:426
>> +    addLanguageChangeObserver(this, &languagesChangedCallback);
> 
> This would be racy with regards with the cache being invalidated in FontGenericFamilies.
> We need to either force invalidating the cache in there first
> or add some other mechanism to make sure this observer is called at the very end or very first
> so that there is no observable race conditions.

Sorry, I do not understand your comment at all. I see that FontGenericFamilies is also registering a language change callback but I do not understand how this impacts Window registering its own callback.
Which cache are you referring to?

[Chris Dumez]

Comment on attachment 291185 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=291185&action=review

>>> Source/WebCore/page/DOMWindow.cpp:426
>>> +    addLanguageChangeObserver(this, &languagesChangedCallback);
>> 
>> This would be racy with regards with the cache being invalidated in FontGenericFamilies.
>> We need to either force invalidating the cache in there first
>> or add some other mechanism to make sure this observer is called at the very end or very first
>> so that there is no observable race conditions.
> 
> Sorry, I do not understand your comment at all. I see that FontGenericFamilies is also registering a language change callback but I do not understand how this impacts Window registering its own callback.
> Which cache are you referring to?

Oh, I think I understand your worry now. How about we simply fire the JS event asynchronously?

[Ryosuke Niwa]

(In reply to comment #5)
> Comment on attachment 291185 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=291185&action=review
> 
> >>> Source/WebCore/page/DOMWindow.cpp:426
> >>> +    addLanguageChangeObserver(this, &languagesChangedCallback);
> >> 
> >> This would be racy with regards with the cache being invalidated in FontGenericFamilies.
> >> We need to either force invalidating the cache in there first
> >> or add some other mechanism to make sure this observer is called at the very end or very first
> >> so that there is no observable race conditions.
> > 
> > Sorry, I do not understand your comment at all. I see that FontGenericFamilies is also registering a language change callback but I do not understand how this impacts Window registering its own callback.
> > Which cache are you referring to?
> 
> Oh, I think I understand your worry now. How about we simply fire the JS
> event asynchronously?

Yeah, that works too.

[Chris Dumez]

Created attachment 291223 [details]
Patch

[WebKit Commit Bot]

Comment on attachment 291223 [details]
Patch

Clearing flags on attachment: 291223

Committed r207040: <http://trac.webkit.org/changeset/207040>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.