Bug 62469 – Fetching a Worker with url that isn't allowed from a file based test causes DRT to crash.

• Status: RESOLVED
• Resolution: FIXED
• Component: DOM

[David Levin]

The problem is that this results in a synchronous failure (in ResourceLoaderDelegate.mm, willSendRequest) which results in deleting the WorkerScriptLoader and calling ActiveDOMObject::unsetPendingActivity (before setPendingActivity has been called).

[David Levin]

Created attachment 96779 [details]
Patch

[WebKit Review Bot]

Attachment 96779 [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/workers/WorkerScriptLoader.h:44:  Code inside a namespace should not be indented.  [whitespace/indent] [4]
Total errors found: 1 in 12 files


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

[Dmitry Titov]

Comment on attachment 96779 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=96779&action=review

r=me. Nice.

>> Source/WebCore/workers/WorkerScriptLoader.h:44
>> +    class ResourceRequest;
> 
> Code inside a namespace should not be indented.  [whitespace/indent] [4]

Could be fixed as a separate patch...

[David Levin]

Created attachment 96818 [details]
Patch for landing

[David Levin]

Attempt to fix Windows build error.

[WebKit Review Bot]

Attachment 96818 [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/workers/WorkerScriptLoader.h:44:  Code inside a namespace should not be indented.  [whitespace/indent] [4]
Total errors found: 1 in 13 files


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

[Dmitry Titov]

Comment on attachment 96818 [details]
Patch for landing

View in context: https://bugs.webkit.org/attachment.cgi?id=96818&action=review

> Source/WebCore/WebCore.xcodeproj/project.pbxproj:7437
>  		417253A91354BBBC00360F2A /* MediaControlElements.h */ = {isa = PBXFileReference; fileEncoding = 4; lastKnownFileType = sourcecode.c.h; path = MediaControlElements.h; sourceTree = "<group>"; };

The cahnge to xcodeproj file seems unrelated.

[David Levin]

Committed as http://trac.webkit.org/changeset/88578 (without the xcodeproj change).

[Ryosuke Niwa]

The test added by this patch is failing on GTK.  Maybe we just need to add a platform-specific result?

http://build.webkit.org/results/GTK%20Linux%2064-bit%20Debug/r88593%20(23265)/fast/workers/worker-crash-with-invalid-location-pretty-diff.html

[Ryosuke Niwa]

http://build.webkit.org/results/GTK%20Linux%2032-bit%20Release/r88596%20(14594)/fast/workers/worker-crash-with-invalid-location-pretty-diff.html
http://build.webkit.org/results/GTK%20Linux%2064-bit%20Debug/r88593%20(23265)/fast/workers/worker-crash-with-invalid-location-pretty-diff.html

[Adam Barth]

(In reply to comment #9)
> The test added by this patch is failing on GTK.  Maybe we just need to add a platform-specific result?

Looks like the GTK DumpRenderTree isn't very good at blocking external network requests!  It's probably hitting the example.com redirect to iana.org on every test run!

[Ryosuke Niwa]

(In reply to comment #11)
> (In reply to comment #9)
> > The test added by this patch is failing on GTK.  Maybe we just need to add a platform-specific result?
> 
> Looks like the GTK DumpRenderTree isn't very good at blocking external network requests!  It's probably hitting the example.com redirect to iana.org on every test run!

That sounds like a really annoying bug for iana.org :(

[Adam Barth]

> That sounds like a really annoying bug for iana.org :(

I suspect they're pretty used to it...  (but we should fix it, of course).

[Ryosuke Niwa]

This test is also failing on Windows:
http://build.webkit.org/results/Windows%207%20Release%20(Tests)/r88594%20(13663)/fast/workers/worker-crash-with-invalid-location-pretty-diff.html

[David Levin]

(In reply to comment #14)
> This test is also failing on Windows:
> http://build.webkit.org/results/Windows%207%20Release%20(Tests)/r88594%20(13663)/fast/workers/worker-crash-with-invalid-location-pretty-diff.html

Windows is also failing fast/workers/worker-event-listener.html. I think there is some other issue going on there that a Window WebKit developer could look at.

(In reply to comment #11)
> (In reply to comment #9)
> > The test added by this patch is failing on GTK.  Maybe we just need to add a platform-specific result?
> 
> Looks like the GTK DumpRenderTree isn't very good at blocking external network requests!  It's probably hitting the example.com redirect to iana.org on every test run!

Addressed http://trac.webkit.org/changeset/88675 (Made it skipped and added a bug for gtk).

[David Levin]

(In reply to comment #15)
> (In reply to comment #14)
> > This test is also failing on Windows:
> > http://build.webkit.org/results/Windows%207%20Release%20(Tests)/r88594%20(13663)/fast/workers/worker-crash-with-invalid-location-pretty-diff.html
> 
> Windows is also failing fast/workers/worker-event-listener.html. I think there is some other issue going on there that a Window WebKit developer could look at.

After further review, it appears that the other Web Worker failure only started occurring after my new test, so I filed a bug and put it in the Skipped list for Windows (http://trac.webkit.org/changeset/88688).

[Ryosuke Niwa]

GTK bug is tracked by the bug 62585.