WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
REOPENED
25458
Restore alignment to device pixel boundaries.
https://bugs.webkit.org/show_bug.cgi?id=25458
Summary
Restore alignment to device pixel boundaries.
David Carson
Reported
2009-04-28 14:58:58 PDT
It seems that the calls to roundToDevicePixel() were dropped in
http://trac.webkit.org/changeset/34210
This results in blurry images when zoomed out as the images are not aligned to pixel boundaries. I spoke with the author of
r34210
change and he said that it caused some regression of images. After some investigation, the regression seems to be because of the srcRect being aligned as well as the destination rect. I would like to restore this feature.
Attachments
patch
(1.36 KB, patch)
2009-04-28 15:02 PDT
,
David Carson
no flags
Details
Formatted Diff
Diff
View All
Add attachment
proposed patch, testcase, etc.
David Carson
Comment 1
2009-04-28 15:02:43 PDT
Created
attachment 29864
[details]
patch
David Carson
Comment 2
2009-04-28 15:03:40 PDT
Ran DRT and there was no regressions on my platform.
Dave Hyatt
Comment 3
2009-04-28 16:10:55 PDT
Comment on
attachment 29864
[details]
patch r=me
David Carson
Comment 4
2009-04-28 18:00:38 PDT
Didn't run pixel tests. When I run those, these tests have an issue: svg/W3C-SVG-1.1/render-groups-03-t.svg expected image image diffs 0.16% svg/carto.net/scrollbar.svg expected image image diffs 0.15% svg/custom/focus-ring.svg expected image image diffs 0.19% svg/custom/image-small-width-height.svg expected image image diffs 0.33%
David Carson
Comment 5
2009-04-28 18:21:22 PDT
With the test case: svg/custom/focus-ring.svg The focus ring is now in the correct place! With the test case: svg/custom/image-small-width-height.svg There was slight blurriness where the blue bar meets the green checker pattern. The line is now much sharper.
David Kilzer (:ddkilzer)
Comment 6
2009-04-28 21:39:18 PDT
This pixel test seems to have regressed with this patch: fast/repaint/fixed.html
David Kilzer (:ddkilzer)
Comment 7
2009-04-28 21:49:45 PDT
(In reply to
comment #6
)
> This pixel test seems to have regressed with this patch: > fast/repaint/fixed.html
Mitz informs me this is:
https://bugs.webkit.org/show_bug.cgi?id=25269
Since the test is already broken, I will check in new results with this patch.
David Kilzer (:ddkilzer)
Comment 8
2009-04-28 22:52:59 PDT
$ git svn dcommit Committing to
http://svn.webkit.org/repository/webkit/trunk
... M LayoutTests/ChangeLog M LayoutTests/platform/mac/svg/W3C-SVG-1.1/render-groups-03-t-expected.checksum M LayoutTests/platform/mac/svg/W3C-SVG-1.1/render-groups-03-t-expected.png M LayoutTests/platform/mac/svg/custom/focus-ring-expected.checksum M LayoutTests/platform/mac/svg/custom/focus-ring-expected.png M LayoutTests/platform/mac/svg/custom/image-small-width-height-expected.checksum M LayoutTests/platform/mac/svg/custom/image-small-width-height-expected.png M WebCore/ChangeLog M WebCore/platform/graphics/cg/ImageCG.cpp Committed
r42987
http://trac.webkit.org/changeset/42987
mitz
Comment 9
2009-04-29 21:42:00 PDT
This change caused
bug 25476
.
David Kilzer (:ddkilzer)
Comment 10
2009-04-30 11:29:01 PDT
Reopening as this change was backed out for
Bug 25476
.
David Levin
Comment 11
2009-05-06 12:56:45 PDT
Comment on
attachment 29864
[details]
patch Clearing r+ to move out of commit queue (and since the change was backed out due to some failed test).
Ahmad Saleem
Comment 12
2022-09-03 10:23:09 PDT
The patch was trying to add this and now we have it in Webkit Source:
https://github.com/WebKit/WebKit/blob/61e05e09838b2da199e3bf1afe7a6120eee9cb4d/Source/WebCore/platform/graphics/cg/GraphicsContextCG.cpp#L337
But it is for iOS only. Is something else needed or we can mark this as "RESOLVED CONFIGURATION CHANGED"? Thanks!
Note
You need to
log in
before you can comment on or make changes to this bug.
Top of Page
Format For Printing
XML
Clone This Bug