Rework BoxBorderPainter's mitre clipping logic

Rework BoxBorderPainter's mitre clipping logic

Change the border clip mitre logic to operate with tri-states (no mitre,
anti-aliased mitre, non-antialiased mitre). Consolidate the mitre and clip
computation logic.

Functional side effect: mitre anti-aliasing is now handled similarly for 
opaque and non-opaque edges (AA when drawing AA edges, non-AA when drawing
non-AA edges, in order to minimize the use of clipping).


BUG=

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=196698

[f(malita), 2015-06-08 19:16:23]

fmalita@chromium.org changed reviewers:
+ chrishtr@chromium.org, fs@opera.com

[f(malita), 2015-06-08 19:25:42]

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

A couple of them are seam bleed progressions:
css2.1/t0805-c5517-brdr-s-00-c.html
css3/blending/background-blend-mode-data-uri-svg-image.html

The rest reflect the change in seam antialiasing for translucent edges.  Always
using AA for translucent seams (legacy behavior) does look nicer under a
magnifier, but I don't think the difference is noticeable otherwise.  It
certainly doesn't make sense to treat translucent edges differently from opaque
edges in this regard - especially since using a different AA mode for mitres vs.
shape introduces extra clips/overhead.  Besides, we'll hopefully get the AA
mitres back when switching to full AA borders.

No perf difference expected or observed locally.  I've been trying to run CT
tests but Blink hasn't rolled for the last few days.  Maybe we'll get a chance
to do that before landing (possibly backport the CL).

[fs, 2015-06-08 21:06:31]

lgtm

https://codereview.chromium.org/1171583003/diff/80001/Source/core/paint/BoxBo...
File Source/core/paint/BoxBorderPainter.h (right):

https://codereview.chromium.org/1171583003/diff/80001/Source/core/paint/BoxBo...
Source/core/paint/BoxBorderPainter.h:31: enum MitreType {
Maybe MiterType (etc.) so them Americans understand too ;-)

[chrishtr, 2015-06-08 21:40:39]

The CQ bit was checked by chrishtr@chromium.org

[chrishtr, 2015-06-08 21:40:40]

lgtm

[commit-bot: I haz the power, 2015-06-08 21:41:06]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1171583003/80001

[commit-bot: I haz the power, 2015-06-08 21:44:21]

Committed patchset #5 (id:80001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=196698