Fix fast/exclusions/shape-outside-floats/shape-outside-floats-non-zero-y.html so that it passes on …

Fix fast/exclusions/shape-outside-floats/shape-outside-floats-non-zero-y.html so that it passes on Windows

This test failed on Windows because in order to create the -expected result, the shape is approximated using floats. Since the javascript code that computes the width of these floats isn't doing exactly the same thing as the C++ code in WebKit that computes the line offsets due to the shape, differences in rounding between platforms can cause small differences in the rendering. It is especially hard to get large curves like the ellipse in this example to work across platforms.

Despite the title of the WebKit bug, this doesn't require an ellipse to test it. So I just replaced it with another rounded rectangle in order to work around the horrible rounding issues that happen when trying to create a ref test for curves.

BUG=227088
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=148683

[commit-bot: I haz the power, 2013-04-09 20:47:03]

No comments yet.

[Bem Jones-Bey (gmail), 2013-04-10 17:37:43]

Julien, can you review this for me?

[Julien - ping for review, 2013-04-10 21:36:51]

On 2013/04/10 17:37:43, bemjb wrote:
> Julien, can you review this for me?

Could you explain the issue, you don't explain why this is better nor what
issues it solves. This makes it hard for me to understand why this is correct.

Also please fix the description (the fact that you are a non-committer bears
little relation to the test change).

[Bem Jones-Bey (gmail), 2013-04-10 23:43:53]

On 2013/04/10 21:36:51, jchaffraix wrote:
> On 2013/04/10 17:37:43, bemjb wrote:
> > Julien, can you review this for me?
> 
> Could you explain the issue, you don't explain why this is better nor what
> issues it solves. This makes it hard for me to understand why this is correct.

Added an explanation.

> 
> Also please fix the description (the fact that you are a non-committer bears
> little relation to the test change).

Sorry about that, that's me not understanding the new tool. I wanted to put a
reply, but couldn't figure out how to do it. I've removed those comments.

[Bem Jones-Bey (gmail), 2013-04-15 16:27:56]

Hey Julien, do you have time to look at my updates?

[Bem Jones-Bey (gmail), 2013-04-16 17:13:54]

Julien is gardening, can anyone else take a look? Thanks.

[eae, 2013-04-16 17:24:21]

LGTM

Seems reasonable, at some point we should probably try to figure out a way get
consistent (pixel perfect) rendering for shapes across platforms though.

[Bem Jones-Bey (gmail), 2013-04-16 17:43:48]

On 2013/04/16 17:24:21, eae wrote:
> LGTM

Thanks!

> Seems reasonable, at some point we should probably try to figure out a way get
> consistent (pixel perfect) rendering for shapes across platforms though.

It's writing the ref test that's the problem: in order to get it right always,
we'd have to implement all of the rounding/etc that happens with subpixel layout
in JavaScript, which seems not worth it.

[Bem Jones-Bey (gmail), 2013-04-16 18:27:49]

I should have asked this earlier, but anyone have time to commit this for me?
Thanks.

[commit-bot: I haz the power, 2013-04-18 00:37:02]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/bemajaniman@gmail.com/13839020/1

[commit-bot: I haz the power, 2013-04-18 00:37:04]

Failed to apply patch for LayoutTests/platform/chromium/TestExpectations:
While running patch -p1 --forward --force --no-backup-if-mismatch;
  can't find file to patch at input line 6
  Perhaps you used the wrong -p or --strip option?
  The text leading up to this was:
  --------------------------
  |Index: LayoutTests/platform/chromium/TestExpectations
  |diff --git a/LayoutTests/platform/chromium/TestExpectations
b/LayoutTests/platform/chromium/TestExpectations
  |index
eeab1b679ced5f7521e49e4f5e3166e2057e5ee3..35942aab6d9b079e6124569f169d8da72f2ec08a
100644
  |--- a/LayoutTests/platform/chromium/TestExpectations
  |+++ b/LayoutTests/platform/chromium/TestExpectations
  --------------------------
  No file to patch.  Skipping patch.
  1 out of 1 hunk ignored

Patch:       LayoutTests/platform/chromium/TestExpectations
Index: LayoutTests/platform/chromium/TestExpectations
diff --git a/LayoutTests/platform/chromium/TestExpectations
b/LayoutTests/platform/chromium/TestExpectations
index
eeab1b679ced5f7521e49e4f5e3166e2057e5ee3..35942aab6d9b079e6124569f169d8da72f2ec08a
100644
--- a/LayoutTests/platform/chromium/TestExpectations
+++ b/LayoutTests/platform/chromium/TestExpectations
@@ -3802,8 +3802,6 @@ Bug(ojan)
fast/text/whitespace/nbsp-mode-and-linewraps.html [ Failure ImageOnlyF
 
 webkit.org/b/113555 [ Win7 Debug ] fast/forms/state-restore-per-form.html [
Pass Timeout ]
 
-webkit.org/b/113673 [ Win ]
fast/exclusions/shape-outside-floats/shape-outside-floats-non-zero-y.html [
ImageOnlyFailure ]
-
 # harfbuzz-ng roll to 0.9.14 needs rebaseline.
 crbug.com/224428 [ Linux ] fast/text/international/khmer-selection.html [
Failure ]
 crbug.com/224428 webkit.org/b/113169 [ Linux ]
fast/text/international/danda-space.html [ Failure ]

[commit-bot: I haz the power, 2013-04-18 20:23:55]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/bemajaniman@gmail.com/13839020/16001

[commit-bot: I haz the power, 2013-04-18 20:46:38]

Change committed as 148683