Disable multicol when printing.

Disable multicol when printing.

Until crbug.com/99358 is fixed, we cannot really print multicol properly.

The old multicol implementation has code (in C++) to prevent multicol when
printing, but since the new implementation creates a different layout tree
structure for multicol, it's going to be quite some work to remove multicol
containers when switching the document to printing. So let's do it in CSS
instead.

R=rune@opera.com,dsinclair@chromium.org,jchaffraix@chromium.org

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=195531

[rune, 2015-05-15 13:36:16]

lgtm

[mstensho (USE GERRIT), 2015-05-15 13:55:36]

On 2015/05/15 13:36:16, rune wrote:
> lgtm

Thanks. Need to figure out what to do with the new failures before landing,
though.

Looking at compositing/squashing/squashing-print.html right now.

Seems that "-webkit-columns:auto" has a weird effect, triggering some machinery
to set a width of 1000 instead of 800. If I do "-webkit-columns:0" instead, the
problem goes away, but I'd rather like to figure out why this happens, since I
think it's rather unreasonable. Could be that some part of the code thinks that
"-webkit-columns:auto" is different from the initial value (which is bad) + some
bug in the printing or print testing framework (which is bad too).

Luckily, it doesn't matter whether I set this in the UA stylesheet or in the
test itself. That's good news at least. :) Needs to be inside @media print to
cause trouble, though.

[rune, 2015-05-15 14:45:18]

On 2015/05/15 13:55:36, mstensho wrote:
> On 2015/05/15 13:36:16, rune wrote:
> > lgtm
> 
> Thanks. Need to figure out what to do with the new failures before landing,
> though.
> 
> Looking at compositing/squashing/squashing-print.html right now.
> 
> Seems that "-webkit-columns:auto" has a weird effect, triggering some
machinery
> to set a width of 1000 instead of 800. If I do "-webkit-columns:0" instead,
the
> problem goes away, but I'd rather like to figure out why this happens, since I

-webkit-columns:0 is an invalid value and should be dropped, so that's the same
as not having a rule at all, right?

[mstensho (USE GERRIT), 2015-05-15 20:02:22]

On 2015/05/15 14:45:18, rune wrote:
> -webkit-columns:0 is an invalid value and should be dropped, so that's the
same
> as not having a rule at all, right?

D'oh, right! Thanks for pointing that out.

[mstensho (USE GERRIT), 2015-05-15 21:25:38]

inspector/elements/styles/media-emulation.html now fails. Like this:

https://storage.googleapis.com/chromium-layout-test-archives/linux_blink_rel/...

I could just update the expectation file, but would that be the right thing to
do? CCed a couple of Source/devtools owners.

[dgozman, 2015-05-16 01:33:24]

On 2015/05/15 21:25:38, mstensho wrote:
> inspector/elements/styles/media-emulation.html now fails. Like this:
> 
>
https://storage.googleapis.com/chromium-layout-test-archives/linux_blink_rel/...
> 
> I could just update the expectation file, but would that be the right thing to
> do? CCed a couple of Source/devtools owners.

That's correct. We examine default style rules for different css media values
applied. Looks like you are adding new rule, so rebaseline is necessary.

[mstensho (USE GERRIT), 2015-05-19 11:36:08]

On 2015/05/16 01:33:24, dgozman_ooo wrote:
> On 2015/05/15 21:25:38, mstensho wrote:
> > inspector/elements/styles/media-emulation.html now fails. Like this:
> > 
> >
>
https://storage.googleapis.com/chromium-layout-test-archives/linux_blink_rel/...
> > 
> > I could just update the expectation file, but would that be the right thing
to
> > do? CCed a couple of Source/devtools owners.
> 
> That's correct. We examine default style rules for different css media values
> applied. Looks like you are adding new rule, so rebaseline is necessary.

That's correct. Thanks.

[mstensho (USE GERRIT), 2015-05-19 11:44:42]

Found out what caused this. Setting -webkit-columns:auto on an element wasn't
exactly the same as not setting it at all (so we detected a style change and
marked for relayout). It should be. This was fixed with
https://codereview.chromium.org/1145773002/

The underlying bug here is
https://code.google.com/p/chromium/issues/detail?id=489615 , which causes layout
tests for printing to sometimes not relayout completely (some of the expectation
files are essentially wrong). Not yet sure what's the idiot-proof way of fixing
that one is, but it's not urgent to fix it at the moment, especially since it
doesn't break anything in the actual Chromium-derived browsers, only in the
testing framework.

[mstensho (USE GERRIT), 2015-05-19 11:49:58]

The CQ bit was checked by mstensho@opera.com

[mstensho (USE GERRIT), 2015-05-19 11:49:59]

The patchset sent to the CQ was uploaded after l-g-t-m from rune@opera.com
Link to the patchset: https://codereview.chromium.org/1141943002/#ps40001
(title: "Update
LayoutTests/inspector/elements/styles/media-emulation-expected.txt")

[commit-bot: I haz the power, 2015-05-19 11:50:12]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1141943002/40001

[commit-bot: I haz the power, 2015-05-19 13:26:32]

Committed patchset #3 (id:40001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=195531