Convert fast/multicol/float-avoidance to reftest.

Convert fast/multicol/float-avoidance to reftest.

Changed column-width such that we're actually using more than one
column. Makes it easier to immediately visually verify that multicol is
being used.

Also changed text contents to be less sensitive to computed font size
across platforms.

R=mstensho@opera.com
BUG=386567
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=184281

[mstensho (USE GERRIT), 2014-10-20 20:35:02]

lgtm, with some suggestions.

Didn't bother to raise issues in vertical-{lr,rl}, but whatever I raised in the
root directory applies there as well.

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
File LayoutTests/fast/multicol/float-avoidance-expected.html (right):

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
LayoutTests/fast/multicol/float-avoidance-expected.html:16: display:
inline-block;
Can use float:left instead of display:inline-block, to save yourself some pain
with white spaces and vertical alignment.

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
File LayoutTests/fast/multicol/float-avoidance.html (left):

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
LayoutTests/fast/multicol/float-avoidance.html:3: text-align:justify;
border:10px solid black">
This test is so simple that I have a hard time anticipating what could possibly
go wrong (unless someone kills me and specs and implements support for multicol
containers that don't establish a BFC). :)

So how about keeping text-align:justify and the top margin on the P element,
just to have some minor evilness?

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
File LayoutTests/fast/multicol/float-avoidance.html (right):

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
LayoutTests/fast/multicol/float-avoidance.html:12: -moz-column-width: 200px;
You can just drop the -moz- stuff.

[mstensho (USE GERRIT), 2014-10-20 20:38:07]

Sorry, forgot one thing: how about nuking the copies in newmulticol/ (pretty
useless test anyway, since the test and the ref are identical)?

[andersr, 2014-10-21 08:58:03]

> Sorry, forgot one thing: how about nuking the copies in newmulticol/ (pretty
useless test anyway, since the test and the ref are identical)?

"Copies", plural? I found (and removed) just one:
newmulticol/float-avoidance.html.

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
File LayoutTests/fast/multicol/float-avoidance-expected.html (right):

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
LayoutTests/fast/multicol/float-avoidance-expected.html:16: display:
inline-block;
On 2014/10/20 20:35:02, mstensho wrote:
> Can use float:left instead of display:inline-block, to save yourself some pain
> with white spaces and vertical alignment.

Done.

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
File LayoutTests/fast/multicol/float-avoidance.html (left):

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
LayoutTests/fast/multicol/float-avoidance.html:3: text-align:justify;
border:10px solid black">
On 2014/10/20 20:35:02, mstensho wrote:
> This test is so simple that I have a hard time anticipating what could
possibly
> go wrong (unless someone kills me and specs and implements support for
multicol
> containers that don't establish a BFC). :)
> 
> So how about keeping text-align:justify and the top margin on the P element,
> just to have some minor evilness?

Unless this affects the multicol container's ability to avoid floats, I think we
should have as little "evilness" as possible.

If you think this case is covered by fast/block/float/float-avoidance (for
example), we should remove it instead.

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
File LayoutTests/fast/multicol/float-avoidance.html (right):

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
LayoutTests/fast/multicol/float-avoidance.html:12: -moz-column-width: 200px;
On 2014/10/20 20:35:02, mstensho wrote:
> You can just drop the -moz- stuff.

Done.

[mstensho (USE GERRIT), 2014-10-21 09:11:43]

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
File LayoutTests/fast/multicol/float-avoidance.html (left):

https://codereview.chromium.org/643923003/diff/1/LayoutTests/fast/multicol/fl...
LayoutTests/fast/multicol/float-avoidance.html:3: text-align:justify;
border:10px solid black">
On 2014/10/21 08:58:03, andersr wrote:
> On 2014/10/20 20:35:02, mstensho wrote:
> > This test is so simple that I have a hard time anticipating what could
> possibly
> > go wrong (unless someone kills me and specs and implements support for
> multicol
> > containers that don't establish a BFC). :)
> > 
> > So how about keeping text-align:justify and the top margin on the P element,
> > just to have some minor evilness?
> 
> Unless this affects the multicol container's ability to avoid floats, I think
we
> should have as little "evilness" as possible.
> 
> If you think this case is covered by fast/block/float/float-avoidance (for
> example), we should remove it instead.

There's no test for multicol there, but that could have been useful. Anyway,
let's keep it here for now. We can live without the extra evilness.

[andersr, 2014-10-23 14:53:36]

The CQ bit was checked by andersr@opera.com

[commit-bot: I haz the power, 2014-10-23 14:54:56]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/643923003/20001

[commit-bot: I haz the power, 2014-10-23 16:01:34]

Committed patchset #2 (id:20001) as 184281