Bug 123265 – PCRE workaround in Source/WTF/wtf/text/StringStatics.cpp can be removed

• Status: RESOLVED
• Resolution: INVALID
• Component: Web Template Framework

[Peter Molnar]

PCRE workaround in Source/WTF/wtf/text/StringStatics.cpp can be removed, as PCRE was removed 2.5 years ago: http://trac.webkit.org/changeset/78211

[Peter Molnar]

Created attachment 215057 [details]
Remove PCRE workaround from Source/WTF/wtf/text/StringStatics.cpp

[WebKit Commit Bot]

Comment on attachment 215057 [details]
Remove PCRE workaround from Source/WTF/wtf/text/StringStatics.cpp

Clearing flags on attachment: 215057

Committed r157931: <http://trac.webkit.org/changeset/157931>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Alexey Proskuryakov]

There are tons of string related assertions on bots now, this patch looks like the most likely culprit. Both layout tests are API tests are broken:

http://build.webkit.org/builders/Apple%20MountainLion%20Debug%20WK1%20%28Tests%29/builds/10963

[WebKit Commit Bot]

Re-opened since this is blocked by bug 123284

[Peter Molnar]

Created attachment 215299 [details]
Remove PCRE workaround from Source/WTF/wtf/text/StringStatics.cpp - fixed alignment problem

Fixed the alignment problem of previous patch, that caused failing API tests on Mac platforms.

[Build Bot]

Comment on attachment 215299 [details]
Remove PCRE workaround from Source/WTF/wtf/text/StringStatics.cpp - fixed alignment problem

Attachment 215299 [details] did not pass mac-ews (mac):
Output: http://webkit-queues.appspot.com/results/11168175

New failing tests:
js/kde/StringObject.html

[Build Bot]

Created attachment 215305 [details]
Archive of layout-test-results from webkit-ews-05 for mac-mountainlion

The attached test failures were seen while running run-webkit-tests on the mac-ews.
Bot: webkit-ews-05  Port: mac-mountainlion  Platform: Mac OS X 10.8.5

[Peter Molnar]

Created attachment 215379 [details]
Remove PCRE workaround - fix failing layout test.

[Peter Molnar]

With the latest patch, both API and layout test breakages are fixed. I only have access to a MacOSX 10.7 machine, API tests were passing on it as of r157986, the revision before dropping OSX 10.7 support.

[Brent Fulgham]

Comment on attachment 215379 [details]
Remove PCRE workaround - fix failing layout test.

Let's try again! r=me.

[WebKit Commit Bot]

Comment on attachment 215379 [details]
Remove PCRE workaround - fix failing layout test.

Clearing flags on attachment: 215379

Committed r158299: <http://trac.webkit.org/changeset/158299>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Ryosuke Niwa]

This patch caused fast/canvas/canvas-color-serialization.html and fast/forms/input-text-paste-maxlength.html to hit assertions:

http://webkit-test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fast%2Fcanvas%2Fcanvas-color-serialization.html%2Cfast%2Fforms%2Finput-text-paste-maxlength.html

e.g.
http://build.webkit.org/results/Apple%20MountainLion%20Debug%20WK2%20(Tests)/r158341%20(13796)/results.html

[Ryosuke Niwa]

I'm reverting the patch for now since I don't think this change fixes any bug in WebKit.

I can reliably reproduce the crash on my Mountain Lion local machine (it also reproduced on Mavericks and GTK+) so please investigate why assertion failures are occurring before relanding the patch again.

[WebKit Commit Bot]

Re-opened since this is blocked by bug 123558

[Brent Fulgham]

This patch seems to cause large test breaks. Is it really worth landing?

Alternatively, should these tests be re baselines or changed in some way to reduce flakiness?

[Brent Fulgham]

The implementation of this empty string function is totally different now.