Bug 73851 – [MutationObservers] Enable in Chromium trunk

• Component: DOM
• Status: RESOLVED FIXED

[Rafael Weinstein]

Mutation observers is now functionally complete on the chromium port (some work in JSC remains for completeness). We'd like to now parallelize getting developer feedback and completing the development work. 

In particular, we plan to invite chrome extension developers who are currently using DOM Mutation Events to experiment with re-implementing using Mutation Observers.

Note that there is currently no plan to release Mutation Observers to chrome beta or release channel, and doing so is predicated on (at minimum) completing all blocking work on the webkit Mutation Observers meta bug (68729).

[Rafael Weinstein]

Created attachment 117967 [details]
Patch

[WebKit Review Bot]

Comment on attachment 117967 [details]
Patch

Clearing flags on attachment: 117967

Committed r102102: <http://trac.webkit.org/changeset/102102>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.

[Vsevolod Vlasov]

This patch caused a lot of failures.

1)
https://bugs.webkit.org/show_bug.cgi?id=73919
Layout Test inspector/debugger/dom-breakpoints.html is failing on chromium linux debug

2)
This patch caused compile failures on all chromium.webkit win builder bots once another patch triggered recompilation of WebKit.cpp.
Compile failures were fixed by clobbering builder bots.

3)
https://bugs.webkit.org/show_bug.cgi?id=73925
REGRESSION(r102102): Causes segmentation fault on chromium unit_tests, content_unittests, sync_unit_tests (Requested by vsevik on #webkit).

-----

Please run chromium try bots before landing this patch next time. 
http://www.chromium.org/developers/testing/try-server-usage#TOC-Simultaneously-patch-code-in-the-We describes how to do that using --sub_rep.

I think you should also notify chromium/webkit sheriffs before landing this patch because rolling webkit DEPS with this patch can potentially break chromium tree (win builders)

[Adam Klein]

(In reply to comment #4)
> This patch caused a lot of failures.

Sorry the trouble.

> 1)
> https://bugs.webkit.org/show_bug.cgi?id=73919
> Layout Test inspector/debugger/dom-breakpoints.html is failing on chromium linux debug
> 
> 2)
> This patch caused compile failures on all chromium.webkit win builder bots once another patch triggered recompilation of WebKit.cpp.
> Compile failures were fixed by clobbering builder bots.

Here's the failure:

http://build.chromium.org/p/chromium.webkit/builders/Win%20Builder/builds/13012/steps/compile/logs/stdio

It's not clear to me whether there's anything we can do here besides clobbering (haven't seen these errors on Linux or Mac builds).

> 3)
> https://bugs.webkit.org/show_bug.cgi?id=73925
> REGRESSION(r102102): Causes segmentation fault on chromium unit_tests, content_unittests, sync_unit_tests (Requested by vsevik on #webkit).

Do you have a stack trace for these segfaults? I don't see any output from the bots.

> 
> -----
> 
> Please run chromium try bots before landing this patch next time. 
> http://www.chromium.org/developers/testing/try-server-usage#TOC-Simultaneously-patch-code-in-the-We describes how to do that using --sub_rep.
> 
> I think you should also notify chromium/webkit sheriffs before landing this patch because rolling webkit DEPS with this patch can potentially break chromium tree (win builders)

[Vsevolod Vlasov]

> > 3)
> > https://bugs.webkit.org/show_bug.cgi?id=73925
> > REGRESSION(r102102): Causes segmentation fault on chromium unit_tests, content_unittests, sync_unit_tests (Requested by vsevik on #webkit).
> 
> Do you have a stack trace for these segfaults? I don't see any output from the bots.
No. The bots didn't have anything. I've just built everything release configuration locally and bisected to find the failure reason, so I don't have a stack trace.
It reproduces locally easily (I did it on linux).

[Adam Klein]

(In reply to comment #6)
> > > 3)
> > > https://bugs.webkit.org/show_bug.cgi?id=73925
> > > REGRESSION(r102102): Causes segmentation fault on chromium unit_tests, content_unittests, sync_unit_tests (Requested by vsevik on #webkit).
> > 
> > Do you have a stack trace for these segfaults? I don't see any output from the bots.
> No. The bots didn't have anything. I've just built everything release configuration locally and bisected to find the failure reason, so I don't have a stack trace.
> It reproduces locally easily (I did it on linux).

Reproduces locally, the problem is WebKit::initialize() is called with WebKitPlatformSupport::currentThread() NULL.  Patch on the way.

[Adam Klein]

Created attachment 118276 [details]
Patch

[Adam Klein]

Committed r102275: <http://trac.webkit.org/changeset/102275>

[Jarek Foksa]

I'm interested in testing this feature, but window.MutationObserver object is not present in current Chromium snapashot (http://commondatastorage.googleapis.com/chromium-browser-snapshots/index.html?path=Linux_x64/114205/)

Do I need some additional flag to enable it?

[Adam Klein]

(In reply to comment #10)
> I'm interested in testing this feature, but window.MutationObserver object is not present in current Chromium snapashot (http://commondatastorage.googleapis.com/chromium-browser-snapshots/index.html?path=Linux_x64/114205/)
> 
> Do I need some additional flag to enable it?

It's vendor-prefixed: look for window.WebKitMutationObserver.