getComputedStyle returns wrong value for CSS3 2D transformations

[Alexander Shalamov]

Created attachment 158763 [details]
Test

When 2D transformation is applied, "borderBoxRect" is used to calculate transformation values.
However, when style values are computed using getComputedStyle, content box size is used to calculate transformation value.

In attached test, div element is moved down (y) by 60 pixels, while getComputedStyle reports that it was moved by 50px.

[Alexander Shalamov]

Created attachment 158782 [details]
Patch

Fix for style calculation and for layout test.

[WebKit Review Bot]

Comment on attachment 158782 [details]
Patch

Attachment 158782 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/13513558

New failing tests:
fast/css/getComputedStyle/getComputedStyle-transform.html

[WebKit Review Bot]

Created attachment 158789 [details]
Archive of layout-test-results from gce-cr-linux-03

The attached test failures were seen while running run-webkit-tests on the chromium-ews.
Bot: gce-cr-linux-03  Port: <class 'webkitpy.common.config.ports.ChromiumXVFBPort'>  Platform: Linux-2.6.39-gcg-201203291735-x86_64-with-Ubuntu-10.04-lucid

[Alexander Shalamov]

Created attachment 158793 [details]
Patch

Fixed typos in changelog.

[WebKit Review Bot]

Attachment 158793 [details] did not pass style-queue:

Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files']" exit_code: 1
Total errors found: 0 in 0 files


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

[Alexander Shalamov]

looks like upload failed. Will re-upload patch.

[Alexander Shalamov]

Created attachment 158798 [details]
Patch

Re-uploading patch

[Simon Fraser (smfr)]

Comment on attachment 158798 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=158798&action=review

> LayoutTests/fast/css/getComputedStyle/getComputedStyle-transform.html:47
> +      for(i=1; i < 7; i++) {

Spaces around the = please. Also need 'var i'.

> LayoutTests/fast/css/getComputedStyle/getComputedStyle-transform.html:49
> +         if ( !(Math.abs(parsedComputed[i] - parsedExpected[i]) < 0.05) )

No space before the !.

[Simon Fraser (smfr)]

Comment on attachment 158798 [details]
Patch

Alexander is going to do a bit more work here.

[Alexander Shalamov]

Created attachment 159658 [details]
Patch

- Applied fixes according to comments
- Added check if renderer has transform
- Added test for non-box element
- Rebased to master

[WebKit Review Bot]

Comment on attachment 159658 [details]
Patch

Attachment 159658 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/13549320

[Early Warning System Bot]

Comment on attachment 159658 [details]
Patch

Attachment 159658 [details] did not pass qt-ews (qt):
Output: http://queues.webkit.org/results/13555040

[Early Warning System Bot]

Comment on attachment 159658 [details]
Patch

Attachment 159658 [details] did not pass qt-wk2-ews (qt):
Output: http://queues.webkit.org/results/13543742

[Alexander Shalamov]

Created attachment 159667 [details]
Patch

newly added pixelSnappedSizingBox is not required - removed

[WebKit Review Bot]

Comment on attachment 159667 [details]
Patch

Attachment 159667 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/13544837

New failing tests:
fast/css/getComputedStyle/getComputedStyle-transform.html

[WebKit Review Bot]

Created attachment 159714 [details]
Archive of layout-test-results from gce-cr-linux-01

The attached test failures were seen while running run-webkit-tests on the chromium-ews.
Bot: gce-cr-linux-01  Port: <class 'webkitpy.common.config.ports.ChromiumXVFBPort'>  Platform: Linux-2.6.39-gcg-201203291735-x86_64-with-Ubuntu-10.04-lucid

[Simon Fraser (smfr)]

Comment on attachment 159667 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=159667&action=review

I thought you were going to fix transform-origin too?

> LayoutTests/ChangeLog:8
> +        This change will fix getComputedStyle-transform.html test.

s/will fix/fixes.

> LayoutTests/ChangeLog:12
> +        Currently, the test doesn't do anything. When new transformation
> +        value is set, it immediately read back as an old value, thus
> +        computed style is equal to an old style. Moreover, expected
> +        result is never checked.

Using 'currently' in a changelog is ambiguous. Does it refer to the state before or after the change? I'd say "before this change, the test didn't do anything".

[Alexander Shalamov]

Created attachment 159960 [details]
Patch

- Fixed transform-origin computation
- Added two cases to transforms/2d/computed-style-origin.html

[Alexander Shalamov]

(In reply to comment #17)
> I thought you were going to fix transform-origin too?

Fixed, and added two test cases to transforms/2d/computed-style-origin.html

> > LayoutTests/ChangeLog:8
> > +        This change will fix getComputedStyle-transform.html test.
> 
> s/will fix/fixes.

Fixed

> 
> > LayoutTests/ChangeLog:12
> Using 'currently' in a changelog is ambiguous. Does it refer to the state before or after the change? I'd say "before this change, the test didn't do anything".

Fixed

[WebKit Review Bot]

Comment on attachment 159960 [details]
Patch

Attachment 159960 [details] did not pass chromium-ews (chromium-xvfb):
Output: http://queues.webkit.org/results/13564349

New failing tests:
fast/css/getComputedStyle/getComputedStyle-zoom-and-background-size.html
svg/transforms/transform-origin-presentation-attribute.xhtml
transitions/shorthand-transitions.html
svg/css/getComputedStyle-basic.xhtml
fast/css/getComputedStyle/getComputedStyle-length-unit.html
fast/css/getComputedStyle/computed-style.html
fast/css/getComputedStyle/computed-style-with-zoom.html
transforms/2d/computed-style-origin.html

[WebKit Review Bot]

Created attachment 159968 [details]
Archive of layout-test-results from gce-cr-linux-06

The attached test failures were seen while running run-webkit-tests on the chromium-ews.
Bot: gce-cr-linux-06  Port: <class 'webkitpy.common.config.ports.ChromiumXVFBPort'>  Platform: Linux-2.6.39-gcg-201203291735-x86_64-with-Ubuntu-10.04-lucid

[Alexander Shalamov]

Created attachment 160102 [details]
Patch

Fixed chromium bot failures

[WebKit Review Bot]

Comment on attachment 160102 [details]
Patch

Clearing flags on attachment: 160102

Committed r126443: <http://trac.webkit.org/changeset/126443>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.