use better heuristic for maxcharwidth

use better heuristic for maxcharwidth

BUG=420901
related skia change
https://codereview.chromium.org/692083002/

[reed1, 2014-10-30 19:51:42]

reed@google.com changed reviewers:
+ bungeman@chromium.org, eae@chromium.org, reed@google.com

[reed1, 2014-10-30 19:51:43]

until we land something like this, I cannot fix fXMin, fXMax, since they are
being used here with the assumption that their values are near 1 (instead of
reflecting the point-size).

This approximation matches what we do on Mac, and can be further improved once
we can safely land fixes to fonthost_freetype.

[jbroman, 2014-10-30 19:55:21]

https://codereview.chromium.org/691133002/diff/1/Source/platform/fonts/Simple...
File Source/platform/fonts/SimpleFontData.cpp (right):

https://codereview.chromium.org/691133002/diff/1/Source/platform/fonts/Simple...
Source/platform/fonts/SimpleFontData.cpp:205: m_maxCharWidth =
std::max(m_avgCharWidth, m_fontMetrics.floatAscent());
drive-by: this seems to be the same as the previous case; should they be merged?

[reed1, 2014-10-30 19:56:49]

On 2014/10/30 19:55:21, jbroman wrote:
>
https://codereview.chromium.org/691133002/diff/1/Source/platform/fonts/Simple...
> File Source/platform/fonts/SimpleFontData.cpp (right):
> 
>
https://codereview.chromium.org/691133002/diff/1/Source/platform/fonts/Simple...
> Source/platform/fonts/SimpleFontData.cpp:205: m_maxCharWidth =
> std::max(m_avgCharWidth, m_fontMetrics.floatAscent());
> drive-by: this seems to be the same as the previous case; should they be
merged?

possibly, but I think that is a little misleading, as the fix to remove/change
the given calculation will not be shared (as the fixes are likely in different
places in skia for different platforms). I thought keeping them separate would
make that more obvious.

[reed2, 2014-10-31 00:58:43]

The CQ bit was checked by reed@chromium.org

[commit-bot: I haz the power, 2014-10-31 00:59:26]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/691133002/1

[reed2, 2014-10-31 01:00:39]

The CQ bit was unchecked by reed@chromium.org

[reed2, 2014-10-31 01:01:24]

ptal -- this (improper) use of fXMin and fXMax stops us from being able to fix
them (for which we have other CLs which need that fix).

[reed1, 2014-10-31 15:58:48]

when the skia CL lands in chrome, this CL should be ready to go.

After it lands...

1. rebaseline as needed (should be minimal)
2. remove guarded code in skia (new behavior is now default)
3. remove flag from gypi

[bungeman-skia, 2014-11-04 18:39:23]

bungeman@google.com changed reviewers:
+ bungeman@google.com

[bungeman-skia, 2014-11-04 18:39:24]

The latest linux_blink_rel actually uploaded the layout test results. I believe
they look pretty reasonable. If you flag the failures (and then see the
generated list at the bottom of the page) you should be able to create a list of
NeedsRebaseline suppressions for Blink's TestExpectations file which can go in
with this CL.

[reed1, 2014-11-04 19:00:21]

glad to hear. unfortunately I am out all this week, so if someone else can start
that process, I would appreciate it.

[bungeman-chromium, 2014-11-11 21:21:33]

Closing in favor of https://codereview.chromium.org/718673004/ with expectations
updates.

[bungeman-chromium, 2014-11-11 21:21:49]

bungeman@chromium.org changed reviewers:
- bungeman@google.com