Optimize RenderTable::colToEffCol() for tables without colspans

Optimize RenderTable::colToEffCol() for tables without colspans

colToEffCol() is very hot in tables with collapsed borders and
it is a very expensive function which iterates through every
column in the table up to the current one. But it only needs to
do this when the table's list of columns contains one whose
colspan exceeds the width of the table, as that is the only
situation in which |ColumnStruct| will still have a span.
(See RenderTableSection:addCell() where the span is
redistributed among other columns in the table during a
second pass of the table.)

Create a fast path for colToEffCol() and effColToCol() that
avoids iterating through every column unless the |m_columns|
vector still has a |ColumnStruct| with a span. The slow
path is extremely rare and probably not
worth optimizing for, as it only applies to tables
where a cell has a colspan that exceeds the total number of
columns provided by other rows in the table.

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=167120

[Julien - ping for review, 2014-02-12 23:05:44]

lgtm, FYI I expect the performance of the collapsed border code to be just
terrible.

https://codereview.chromium.org/154243002/diff/230001/PerformanceTests/Layout...
File
PerformanceTests/Layout/resources/large-table-with-collapsed-borders-and-colspans.js
(right):

https://codereview.chromium.org/154243002/diff/230001/PerformanceTests/Layout...
PerformanceTests/Layout/resources/large-table-with-collapsed-borders-and-colspans.js:29:
}
It really seems like this code could be shared among the 3 tests if you accepted
a parameter for the |colSpan|.

[rhogan, 2014-02-13 20:58:14]

The CQ bit was checked by robhogan@gmail.com

[commit-bot: I haz the power, 2014-02-13 20:58:22]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/robhogan@gmail.com/154243002/300001

[eseidel, 2014-02-13 21:02:14]

The CQ bit was unchecked by eseidel@chromium.org

[eseidel, 2014-02-13 21:02:15]

The CQ bit was checked by eseidel@chromium.org

[eseidel, 2014-02-13 21:03:40]

lgtm

Hmm.  I remember working a bunch on this code when optimizing tests from
https://github.com/dglazkov/performance-tests  I'm a bit surprised there was
still fruit here.

[commit-bot: I haz the power, 2014-02-13 23:22:28]

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

[commit-bot: I haz the power, 2014-02-13 23:22:29]

Commit queue rejected this change because the description was changed
between the time the change entered the commit queue and the time it
was ready to commit. You can safely check the commit box again.

[rhogan, 2014-02-13 23:28:31]

The CQ bit was checked by robhogan@gmail.com

[commit-bot: I haz the power, 2014-02-13 23:28:47]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/robhogan@gmail.com/154243002/300001

[commit-bot: I haz the power, 2014-02-13 23:29:08]

Change committed as 167120