ASSERTION FAILED: y2 >= y1 in WebCore::RenderObject::drawSolidBoxSide

ASSERTION FAILED: y2 >= y1 in WebCore::RenderObject::drawSolidBoxSide

Don't draw the outline if the rectangle to draw is empty.

Backported from WebKit: http://trac.webkit.org/changeset/154064 (Patch by Rob Buis)

R=leviw@chromium.org
BUG=351043
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=172917

[leviw_travelin_and_unemployed, 2014-03-11 16:35:35]

lgtm

Okay.

[commit-bot: I haz the power, 2014-03-11 16:35:59]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rhodovan.u-szeged@partner.samsung.com/...

[commit-bot: I haz the power, 2014-03-11 17:05:51]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2014-03-11 17:05:56]

Try jobs failed on following builders: mac_blink_rel

[leviw_travelin_and_unemployed, 2014-03-13 18:20:39]

Looks like this is causing a test failure. I'm guessing it just needs to be
rebaselined, but please take a look.

[reni, 2014-03-27 17:30:42]

On 2014/03/13 18:20:39, Levi wrote:
> Looks like this is causing a test failure. I'm guessing it just needs to be
> rebaselined, but please take a look.

Could you (or anybody) restart the trybots, pls? I cannot reproduce any
regressions locally and the previous results are not available anymore.

[pdr., 2014-03-27 20:27:58]

On 2014/03/27 17:30:42, renata.hodovan wrote:
> On 2014/03/13 18:20:39, Levi wrote:
> > Looks like this is causing a test failure. I'm guessing it just needs to be
> > rebaselined, but please take a look.
> 
> Could you (or anybody) restart the trybots, pls? I cannot reproduce any
> regressions locally and the previous results are not available anymore.

Done

[reni, 2014-03-28 14:10:04]

On 2014/03/27 20:27:58, pdr wrote:
> On 2014/03/27 17:30:42, renata.hodovan wrote:
> > On 2014/03/13 18:20:39, Levi wrote:
> > > Looks like this is causing a test failure. I'm guessing it just needs to
be
> > > rebaselined, but please take a look.
> > 
> > Could you (or anybody) restart the trybots, pls? I cannot reproduce any
> > regressions locally and the previous results are not available anymore.
> 
> Done

@pdr: thanks.

@Levi: fast/css/outline-narrowLine.html seems to be the only valid regression
after the patch. The problem here is that because of the early return on empty
boxes (m_width/m_height <= 0) the <br> tags with the outline properties are
dropped (since their size is 0). This is equivalent behavior with Opera and
WebKit, but not with Firefox.

Another solution would be to return only if any of the dimensions of the box is
negative (m_width/m_height < 0). This version would not cause any regressions.

The question is which one is the preferred/correct version.

[reni, 2014-04-02 16:39:11]

In short, the question is what is the expected result here? :

<span style="outline: solid green 50px"><br></span>

Do we want to display a green rectangle here or not? (Firefox does but Opera and
WebKit don't.)

[reni, 2014-04-10 14:55:01]

I've updated the code to only make the early return only if the size of the
outline box is a negative value. I preferred this solution, since according to
the W3C standard, the outline property applies to every inline elements and <br>
is also an inline element. Furthermore, this way we don't break the current
behavior of outlines in chrome.

[reni, 2014-04-15 11:31:38]

On 2014/04/10 14:55:01, renata.hodovan wrote:
> I've updated the code to only make the early return only if the size of the
> outline box is a negative value. I preferred this solution, since according to
> the W3C standard, the outline property applies to every inline elements and
<br>
> is also an inline element. Furthermore, this way we don't break the current
> behavior of outlines in chrome.

Could someone take a look at this and/or put a try bot flag on this, pls?

[reni, 2014-04-29 17:28:26]

On 2014/04/15 11:31:38, renata.hodovan wrote:
> On 2014/04/10 14:55:01, renata.hodovan wrote:
> > I've updated the code to only make the early return only if the size of the
> > outline box is a negative value. I preferred this solution, since according
to
> > the W3C standard, the outline property applies to every inline elements and
> <br>
> > is also an inline element. Furthermore, this way we don't break the current
> > behavior of outlines in chrome.
> 
> Could someone take a look at this and/or put a try bot flag on this, pls?

I've reseted the failing try bots and the two win and one linux bot is still red
but I'm pretty sure that those problems are not related to my change.

[leviw_travelin_and_unemployed, 2014-04-29 17:42:58]

The CQ bit was checked by leviw@chromium.org

[commit-bot: I haz the power, 2014-04-29 17:43:22]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rhodovan.u-szeged@partner.samsung.com/...

[commit-bot: I haz the power, 2014-04-30 04:22:34]

Change committed as 172917