Bug 58233 - fast/css/dashboard-regions-attr-crash.html asserts
Summary: fast/css/dashboard-regions-attr-crash.html asserts
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: CSS (show other bugs)
Version: 528+ (Nightly build)
Hardware: PC All
: P2 Normal
Assignee: Nobody
URL:
Keywords:
Depends on:
Blocks: 53488
  Show dependency treegraph
 
Reported: 2011-04-11 07:19 PDT by Csaba Osztrogonác
Modified: 2011-04-12 09:00 PDT (History)
2 users (show)

See Also:


Attachments
proposed fix (2.10 KB, patch)
2011-04-12 08:18 PDT, Csaba Osztrogonác
no flags Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Csaba Osztrogonác 2011-04-11 07:19:05 PDT
fast/css/dashboard-regions-attr-crash.html asserts on the Qt debug bots.
It is is a very very old assertion.

stdout for r83435:
-------------------
ASSERTION FAILED: unimplemented propertyID: 1287
../../../Source/WebCore/css/SVGCSSParser.cpp(294) : bool WebCore::CSSParser::parseSVGValue(int, bool)
Comment 1 Andras Becsi 2011-04-11 08:48:10 PDT
(In reply to comment #0)
> fast/css/dashboard-regions-attr-crash.html asserts on the Qt debug bots.
> It is is a very very old assertion.
> 
> stdout for r83435:
> -------------------
> ASSERTION FAILED: unimplemented propertyID: 1287
> ../../../Source/WebCore/css/SVGCSSParser.cpp(294) : bool WebCore::CSSParser::parseSVGValue(int, bool)

AFAIK dashboard region is Mac-specific, and in CSSParser::parseValue the CSSPropertyWebkitDashboardRegion case is guarded by ENABLE(DASHBOARD_SUPPORT). Since Qt does not support it, it falls though further to parseSVGValue where in the default case there is this ASSERT_WITH_MESSAGE.

I think the test should either be skipped permanently or there should be an elif define with a break in parseValue to catch the CSSPropertyWebkitDashboardRegion property.

I vote for the first approach ot make the debug bot happy.
Comment 2 Csaba Osztrogonác 2011-04-12 08:18:18 PDT
Created attachment 89205 [details]
proposed fix
Comment 3 Andras Becsi 2011-04-12 08:30:20 PDT
(In reply to comment #2)
> Created an attachment (id=89205) [details]
> proposed fix

LGTM!

Even better!

For features.pri having -= lines is simply wrong, maybe even check-webkit-style could shout for it.
Comment 4 Csaba Osztrogonác 2011-04-12 08:42:25 PDT
(In reply to comment #3)
> (In reply to comment #2)
> > Created an attachment (id=89205) [details] [details]
> > proposed fix
> 
> LGTM!
> 
> Even better!
> 
> For features.pri having -= lines is simply wrong, maybe even check-webkit-style could shout for it.

It was good long time ago, because the initialization of DASHBOARDSUPPORTCSSPROPERTIES was before this -= hack.
(When there aren't any WebCore.pri, DerivedSources.pro, 
CodeGenerators.pri, but a monumental and ugly WebCore.pro)
Comment 5 Benjamin Poulain 2011-04-12 08:55:00 PDT
Comment on attachment 89205 [details]
proposed fix

Look sane, trust you
Comment 6 Csaba Osztrogonác 2011-04-12 08:59:53 PDT
Comment on attachment 89205 [details]
proposed fix

Landed in http://trac.webkit.org/changeset/83593