Calculate minimum column height after layout.

Calculate minimum column height after layout.

This moves some multicol specific code out from LayoutBlock and
LayoutBlockFlow. More importantly, though, is that once we add optimizations
to skip subtrees during multicol layout, collecting this information during
layout isn't going to work.

R=leviw@chromium.org

Committed: https://crrev.com/479a72d61ee559c651a5a69da1edf5df7b4f3eb5
Cr-Commit-Position: refs/heads/master@{#358171}

[mstensho (USE GERRIT), 2015-11-05 12:55:58]

leviw, can you take a look, please?

I need to improve the minimum height stuff for nested multicol, but I figured
I'd get it out from LayoutBlock & co first.

[leviw_travelin_and_unemployed, 2015-11-05 18:44:26]

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/layout/ColumnBalancer.cpp (right):

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/layout/ColumnBalancer.cpp:113: if
(box.paginationBreakability() != LayoutBox::AllowAnyBreaks) {
This is new code?

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/layout/ColumnBalancer.cpp:130: unsigned lineCount
= std::max<unsigned>(style.hasAutoOrphans() ? 1 : style.orphans(),
style.widows());
minimumLineCount?

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/layout/ColumnBalancer.h (right):

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/layout/ColumnBalancer.h:77: void
examineLine(const LayoutBlockFlow& containingBlock, const RootInlineBox&);
Since we're only looking at RootInlineBoxes, the containing block is always
accessible via the block() method on RootInlineBlock in O(1) time. Splitting the
object off seems more error prone to me.

[mstensho (USE GERRIT), 2015-11-05 19:15:56]

Just answering on your comments. Will make the code changes tomorrow.

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/layout/ColumnBalancer.cpp (right):

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/layout/ColumnBalancer.cpp:113: if
(box.paginationBreakability() != LayoutBox::AllowAnyBreaks) {
On 2015/11/05 18:44:26, leviw wrote:
> This is new code?

This used to be handled in LayoutBlockFlow::adjustForUnsplittableChild(), by
calling updateMinimumHeight().

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/layout/ColumnBalancer.cpp:130: unsigned lineCount
= std::max<unsigned>(style.hasAutoOrphans() ? 1 : style.orphans(),
style.widows());
On 2015/11/05 18:44:26, leviw wrote:
> minimumLineCount?

Good idea. This code was just copied from LayoutBlockFlow.

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/layout/ColumnBalancer.h (right):

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/layout/ColumnBalancer.h:77: void
examineLine(const LayoutBlockFlow& containingBlock, const RootInlineBox&);
On 2015/11/05 18:44:26, leviw wrote:
> Since we're only looking at RootInlineBoxes, the containing block is always
> accessible via the block() method on RootInlineBlock in O(1) time. Splitting
the
> object off seems more error prone to me.

I was unsure, with the ongoing efforts to untangle line layout from LayoutObject
in mind. But if I'm not introducing any unwanted coupling by doing it, I'll do
it. :)

[leviw_travelin_and_unemployed, 2015-11-05 19:24:45]

On 2015/11/05 at 19:15:56, mstensho wrote:
> Just answering on your comments. Will make the code changes tomorrow.
> 
>
https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
> File third_party/WebKit/Source/core/layout/ColumnBalancer.cpp (right):
> 
>
https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
> third_party/WebKit/Source/core/layout/ColumnBalancer.cpp:113: if
(box.paginationBreakability() != LayoutBox::AllowAnyBreaks) {
> On 2015/11/05 18:44:26, leviw wrote:
> > This is new code?
> 
> This used to be handled in LayoutBlockFlow::adjustForUnsplittableChild(), by
calling updateMinimumHeight().
> 
>
https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
> third_party/WebKit/Source/core/layout/ColumnBalancer.cpp:130: unsigned
lineCount = std::max<unsigned>(style.hasAutoOrphans() ? 1 : style.orphans(),
style.widows());
> On 2015/11/05 18:44:26, leviw wrote:
> > minimumLineCount?
> 
> Good idea. This code was just copied from LayoutBlockFlow.
> 
>
https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
> File third_party/WebKit/Source/core/layout/ColumnBalancer.h (right):
> 
>
https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
> third_party/WebKit/Source/core/layout/ColumnBalancer.h:77: void
examineLine(const LayoutBlockFlow& containingBlock, const RootInlineBox&);
> On 2015/11/05 18:44:26, leviw wrote:
> > Since we're only looking at RootInlineBoxes, the containing block is always
> > accessible via the block() method on RootInlineBlock in O(1) time. Splitting
the
> > object off seems more error prone to me.
> 
> I was unsure, with the ongoing efforts to untangle line layout from
LayoutObject in mind. But if I'm not introducing any unwanted coupling by doing
it, I'll do it. :)

Totally understandable. That method should probably be renamed to
"containingBlock" for clarity. Anyhow, it's safe here :)

[mstensho (USE GERRIT), 2015-11-05 22:17:10]

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/layout/ColumnBalancer.cpp (right):

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/layout/ColumnBalancer.cpp:130: unsigned lineCount
= std::max<unsigned>(style.hasAutoOrphans() ? 1 : style.orphans(),
style.widows());
On 2015/11/05 19:15:56, mstensho wrote:
> On 2015/11/05 18:44:26, leviw wrote:
> > minimumLineCount?
> 
> Good idea. This code was just copied from LayoutBlockFlow.

Done.

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/layout/ColumnBalancer.h (right):

https://codereview.chromium.org/1406973008/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/layout/ColumnBalancer.h:77: void
examineLine(const LayoutBlockFlow& containingBlock, const RootInlineBox&);
On 2015/11/05 19:15:56, mstensho wrote:
> On 2015/11/05 18:44:26, leviw wrote:
> > Since we're only looking at RootInlineBoxes, the containing block is always
> > accessible via the block() method on RootInlineBlock in O(1) time. Splitting
> the
> > object off seems more error prone to me.
> 
> I was unsure, with the ongoing efforts to untangle line layout from
LayoutObject
> in mind. But if I'm not introducing any unwanted coupling by doing it, I'll do
> it. :)

Done.

[leviw_travelin_and_unemployed, 2015-11-05 22:18:36]

The CQ bit was checked by leviw@chromium.org

[leviw_travelin_and_unemployed, 2015-11-05 22:18:36]

lgtm

[commit-bot: I haz the power, 2015-11-05 22:20:15]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1406973008/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1406973008/20001

[commit-bot: I haz the power, 2015-11-05 23:27:26]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-11-05 23:28:10]

Patchset 2 (id:??) landed as
https://crrev.com/479a72d61ee559c651a5a69da1edf5df7b4f3eb5
Cr-Commit-Position: refs/heads/master@{#358171}