Bug 43104 – recent colorspace changes have fixed svg/custom/grayscale-gradient-mask.svg

• Status: RESOLVED
• Resolution: FIXED
• Component: SVG

[Alex Nicolaou]

In my opinion svg/custom/grayscale-gradient-mask.svg renders close to correctly now that chromium-linux and chromium-mac ports use LinearRGB colour space for the gradient. This particular example is really hard to verify because of the gentle change to the gradient but the mac diffs show only a scattering of bad pixels and the linux diffs (which are worse) are probably mostly due to SKIA roundoff differences in the compositing. Creating this bug so that I can submit new baselines.

[Alex Nicolaou]

This test is passing without new expectations and that is reflected in the new expectations in bug 42428 ... closing as duplicate.

*** This bug has been marked as a duplicate of bug 42428 ***

[Alex Nicolaou]

That was only true for Mac, this test and another were affected by the colorspace change in masks on Linux/Windows and I'll attach rebaselines in a minute.

[Alex Nicolaou]

Created attachment 63413 [details]
Patch

[WebKit Commit Bot]

Comment on attachment 63413 [details]
Patch

Rejecting patch 63413 from commit-queue.

Failed to run "[u'/Users/eseidel/Projects/CommitQueue/WebKitTools/Scripts/svn-apply', u'--reviewer', u'Darin Adler', u'--force']" exit_code: 255
Last 500 characters of output:
th fuzz 3.
patching file LayoutTests/platform/chromium-linux/svg/custom/grayscale-gradient-mask-expected.checksum
patching file LayoutTests/platform/chromium-win/svg/custom/grayscale-gradient-mask-expected.checksum
delta is only avaliable with the XS version at /Library/Perl/5.8.8/Compress/Zlib.pm line 9
BEGIN failed--compilation aborted at /Library/Perl/5.8.8/Compress/Zlib.pm line 9.
Compilation failed in require at /Users/eseidel/Projects/CommitQueue/WebKitTools/Scripts/VCSUtils.pm line 1662.

Full output: http://queues.webkit.org/results/3576953

[Eric Seidel (no email)]

Comment on attachment 63413 [details]
Patch

I've cpan'd up the cq.  We'll see if it works better now.

[WebKit Commit Bot]

Comment on attachment 63413 [details]
Patch

Rejecting patch 63413 from commit-queue.

Failed to run "[u'/Users/eseidel/Projects/CommitQueue/WebKitTools/Scripts/svn-apply', u'--reviewer', u'Darin Adler', u'--force']" exit_code: 255
Last 500 characters of output:
th fuzz 3.
patching file LayoutTests/platform/chromium-linux/svg/custom/grayscale-gradient-mask-expected.checksum
patching file LayoutTests/platform/chromium-win/svg/custom/grayscale-gradient-mask-expected.checksum
delta is only avaliable with the XS version at /Library/Perl/5.8.8/Compress/Zlib.pm line 9
BEGIN failed--compilation aborted at /Library/Perl/5.8.8/Compress/Zlib.pm line 9.
Compilation failed in require at /Users/eseidel/Projects/CommitQueue/WebKitTools/Scripts/VCSUtils.pm line 1662.

Full output: http://queues.webkit.org/results/3568902

[Eric Seidel (no email)]

These are delta patches.  We only support literal diffs for svn-apply.  I'm confused by why the zlib error, but this patch wouldn't work even w/o it, iirc.

[Alex Nicolaou]

I've made all my patches the same way, using the script that creates and uploads the patch. I may have made the patch on different systems ... but I think they come out the same. For example patch 62917 from bug 42428 looks like the same style as this one to me and it was landed OK by the bot.

If the patch format is wrong we should fix or delete the patch script. I'd vote for fixing it but I'm not super familiar with the script.

Still, I'm puzzled why the other patches have applied and this one not. I would have guessed that it has more to do with the cq moving to snow leopard than with my patches?

[Alex Nicolaou]

Created attachment 63802 [details]
Patch

[Adam Barth]

Committed r64896: <http://trac.webkit.org/changeset/64896>