Bug 167197 - [Win] Test output results should not be written twice.
Summary: [Win] Test output results should not be written twice.
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Tools / Tests (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Per Arne Vollan
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2017-01-19 04:35 PST by Per Arne Vollan
Modified: 2017-01-20 10:17 PST (History)
4 users (show)

See Also:

Attachments
Patch (1.91 KB, patch)
2017-01-19 04:42 PST, Per Arne Vollan 		no flags Details | Formatted Diff | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Per Arne Vollan 2017-01-19 04:35:31 PST 
Test results indicate that test output is sometimes written twice for the same test, causing flakiness.
Comment 1 Per Arne Vollan 2017-01-19 04:42:05 PST 
Created attachment 299243 [details]
Patch
Comment 2 Brent Fulgham 2017-01-20 09:26:46 PST 
Comment on attachment 299243 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=299243&action=review

r=me

> Tools/DumpRenderTree/win/DumpRenderTree.cpp:685
> +

It would be good to understand why this happens. Do we need a similar protection in macOS and other platforms? Or does Windows call dump() in more places?
Comment 3 WebKit Commit Bot 2017-01-20 09:51:35 PST 
Comment on attachment 299243 [details]
Patch

Clearing flags on attachment: 299243

Committed r210965: <http://trac.webkit.org/changeset/210965>
Comment 4 WebKit Commit Bot 2017-01-20 09:51:38 PST 
All reviewed patches have been landed.  Closing bug.
Comment 5 Per Arne Vollan 2017-01-20 10:17:11 PST 
(In reply to comment #2)
> Comment on attachment 299243 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=299243&action=review
> 
> r=me
> 
> > Tools/DumpRenderTree/win/DumpRenderTree.cpp:685
> > +
> 
> It would be good to understand why this happens. Do we need a similar
> protection in macOS and other platforms? Or does Windows call dump() in more
> places?

macOS will not process messages after dump() is called, while Windows will process a few messages before receiving the WM_QUIT message. This opens up for dump() to be called again on Windows. This might happen if the test performs a reload, and calls notifyDone() a second time.

Thanks for reviewing!