Bug 48179 – Convert SVGAngle to the new SVGPropertyTearOff concept, reimplement it properly

• Component: SVG
• Status: RESOLVED FIXED

[Nikolas Zimmermann]

Convert SVGAngle to the new SVGPropertyTearOff concept.

[Nikolas Zimmermann]

Created attachment 71658 [details]
Patch

Tested local mac debug build on leopard and chromium build on snow leopard.
SVG pixel tests pass w/o regressions. Still have to do a full tests run, going to report soon.
The new SVGAngle tests which requires ScriptTypeSetting for attribute setters works perfectly in V8 & JSC, verified that twice :-)

[WebKit Review Bot]

Attachment 71658 [details] did not pass style-queue:

Failed to run "['WebKitTools/Scripts/check-webkit-style']" exit_code: 1
WebCore/svg/SVGAnimatedAngle.h:25:  Alphabetical sorting problem.  [build/include_order] [4]
Total errors found: 1 in 27 files


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

[Nikolas Zimmermann]

Created attachment 71659 [details]
Patch v2

Fixed style issue.

[Nikolas Zimmermann]

No layout test regression, did a full run-webkit-tests run.

[Early Warning System Bot]

Attachment 71659 [details] did not build on qt:
Build output: http://queues.webkit.org/results/4607060

[Eric Seidel (no email)]

Attachment 71659 [details] did not build on mac:
Build output: http://queues.webkit.org/results/4670061

[Nikolas Zimmermann]

Hm, I just tried a clean build on leopard, worked fine. The chrome build, from scratch, also worked fine. I guess I have to try another snow leopard (non-chrome) build.

[Nikolas Zimmermann]

Created attachment 71677 [details]
Patch v3

*grml* Diffed my working tree against the "Patch v2", turns out I forgot to include a CodeGenerator.pm change. With that one it builds fine...

[Dirk Schulze]

Comment on attachment 71677 [details]
Patch v3

View in context: https://bugs.webkit.org/attachment.cgi?id=71677&action=review

> WebCore/bindings/scripts/CodeGeneratorJS.pm:1848
> +                                           push(@implContent, "    if (!ec)\n"); 
> +                                        push(@implContent, "        imp->commitChange();\n");

Indention looks wrong

Looks great, please take a look at the bots, since EWS will take some while.

r=me

[Nikolas Zimmermann]

Committed r70410: <http://trac.webkit.org/changeset/70410>

[Nikolas Zimmermann]

Ouch, I was not aware the webgl tests don't run on Leopard. I broke them, investigating.

[Nikolas Zimmermann]

Created attachment 71678 [details]
Fix WebGL follow-up patch

[Dirk Schulze]

Comment on attachment 71678 [details]
Fix WebGL follow-up patch

r=me

[Nikolas Zimmermann]

Comment on attachment 71678 [details]
Fix WebGL follow-up patch

Landed speculative WebGL fix in r70411.
Hope these tests work again, otherwhise I have to roll it out.

[Nikolas Zimmermann]

(In reply to comment #14)
> (From update of attachment 71678 [details])
> Landed speculative WebGL fix in r70411.
> Hope these tests work again, otherwhise I have to roll it out.

That did not fix it, running a local snow leopard build now, to figure out whats happening.
Will take some time...

[Nikolas Zimmermann]

Created attachment 71688 [details]
Fix WebGL follow-up patch - v2

[Dimitri Glazkov (Google)]

Comment on attachment 71688 [details]
Fix WebGL follow-up patch - v2

ok

[Nikolas Zimmermann]

Committed r70414: <http://trac.webkit.org/changeset/70414>