Bug 63169

Summary:	REGRESSION (r89305-r89312): webarchive/loading/mainresource-null-mimetype-crash.html failing on Leopard/SnowLeopard release
Product:	WebKit	Reporter:	Jessie Berlin <jberlin>
Component:	Page Loading	Assignee:	Nobody <webkit-unassigned>
Status:	RESOLVED WORKSFORME
Severity:	Normal	CC:	abarth, eric, jberlin, rniwa, zimmermann
Priority:	P2	Keywords:	InRadar, LayoutTestFailure, Regression
Version:	528+ (Nightly build)
Hardware:	Mac
OS:	OS X 10.6
URL:	http://build.webkit.org/results/SnowLeopard Intel Release (Tests)/r89312 (30567)/results.html

Description Jessie Berlin 2011-06-22 12:50:59 PDT

webarchive/loading/mainresource-null-mimetype-crash.html has been failing on SnowLeopard Intel Release (Tests) since r89312.

http://build.webkit.org/results/SnowLeopard%20Intel%20Release%20(Tests)/r89305%20(30566)/results.html passed
http://build.webkit.org/results/SnowLeopard%20Intel%20Release%20(Tests)/r89312%20(30567)/results.html failed

As far as I can tell, the most suspicious change-set in that range is http://trac.webkit.org/changeset/89312.

I will land the failing expected results in order to get the bots green.

Comment 1 Jessie Berlin 2011-06-22 13:08:13 PDT

Added failing expected results in http://trac.webkit.org/changeset/89464

Comment 2 Jessie Berlin 2011-06-22 13:11:59 PDT

<rdar://problem/9657597>

Comment 3 Adam Barth 2011-06-22 13:55:15 PDT

Seems likely to be my fault.

Comment 4 Adam Barth 2011-06-22 15:06:43 PDT

This is actually a progression.  The old results had the wrong URL in the callback.

Comment 5 Adam Barth 2011-06-22 15:09:15 PDT

/me is confused.  Maybe it's different in a release build?

Comment 6 Adam Barth 2011-06-22 15:41:32 PDT

The issue doesn't reproduce for me.

Comment 7 Adam Barth 2011-06-22 15:43:06 PDT

I'm not sure how to continue here.  I suspect there's some sort of memory issue that's corrupting the string somehow non-deterministically.  (I also suspect the underlying bug has nothing to do with my patch, but that's just because I'm pig headed.)

Comment 8 Nikolas Zimmermann 2011-06-25 00:17:57 PDT

It also fails on  Leoaprd: http://build.webkit.org/results/Leopard%20Intel%20Release%20(Tests)/r89732%20(33501)/results.html

I'm going to skip this test for now, to get the bots green.

Comment 9 Nikolas Zimmermann 2011-06-25 00:35:09 PDT

I'm on a gardening mission to turn bots green again: skipped test on mac in general in r89736.

Comment 10 Jessie Berlin 2011-06-27 13:41:11 PDT

Skipping the test on Mac caused another test to start failing on Snow Leopard Release:

http://build.webkit.org/builders/SnowLeopard%20Intel%20Release%20(Tests)/builds/30755
http://build.webkit.org/builders/SnowLeopard%20Intel%20Release%20(Tests)/builds/30756
http://build.webkit.org/results/SnowLeopard%20Intel%20Release%20(Tests)/r89736%20(30756)/webarchive/loading/missing-data-pretty-diff.html

Comment 11 Adam Barth 2011-06-27 15:01:10 PDT

Sounds like there's some sort of inter-test dependency.