Bug 23105 – canvas setFillColor function ignores alpha argument

• Component: DOM
• Status: RESOLVED FIXED

[Darin Adler]

I noticed this while looking at unused argument warnings.

[Darin Adler]

Created attachment 26423 [details]
patch

[Oliver Hunt]

Comment on attachment 26423 [details]
patch

r=me, that's hilarious

[Darin Adler]

http://trac.webkit.org/changeset/39595

[Simon Fraser (smfr)]

fast/canvas/set-colors.html is failing now:
<http://build.webkit.org/results/trunk-mac-ppc-pixel/771/results.html>

[Darin Adler]

(In reply to comment #4)
> fast/canvas/set-colors.html is failing now:
> <http://build.webkit.org/results/trunk-mac-ppc-pixel/771/results.html>

There are two kinds of failures there:

    1) The CMYK conversion seems to be coming out slightly differently. I can either remove the CMYK test entirely or choose values for CMYK that give more consistent results.

    2) The very thing I was trying to fix seems to still be broken on Tiger!

I have no idea what to do about (2).

[Darin Adler]

I presume both of these failures are Tiger-only, because the tests are passing for me on Leopard machines.

[Simon Fraser (smfr)]

I saw a failure on my Leopard machine.

Does it depend on the screen color profile?

[Simon Fraser (smfr)]

Related to bug 22300?

[Darin Adler]

(In reply to comment #7)
> I saw a failure on my Leopard machine.

The same 5 failures as on the buildbot?

> Does it depend on the screen color profile?

It shouldn't. Perhaps it does.

I need to do something quick to get the buildbot green again. The big mystery is the setFillColor case, since that's the bug I was actually fixing!

[Darin Adler]

On my computer I am seeing the CMYK test failures both in Safari and in DumpRenderTree.

And I am seeing the setFillColor failure, but only in DumpRenderTree.

[Darin Adler]

All figured out. All fixed now.

http://trac.webkit.org/changeset/39612