Zero-length path with round/square lineCap should paint circle on Canvas

Zero-length path with round/square lineCap should paint circle on Canvas

Currently CanvasPathMethods does a premature optimization in
stroking line, curve, arc, rect---it drops the point when the
 point is identical to current point, or in the case or
curve, drops the curves when the two given points are
identical. The results in incorrect rendering of zero-length
paths with caps, based on the specification
(https://www.w3.org/TR/2dcontext/#drawing-paths-to-the-canvas).

The five layout tests in philip/tests/2d.path.stroke.prune.*.html
are outdated canvas tests. Based on the spec change and the test
results here
 (https://philip.html5.org/tests/canvas/suite/tests/results.html),
other browsers report false positive errors on these
tests too. We thus match what other browsers do in this
aspect and rebaseline the expected outputs.

BUG=322807, 577655
Committed: https://crrev.com/245b756e89823ae1e668bbc52cf57493c87fb0bd
Cr-Commit-Position: refs/heads/master@{#388551}

[xlai (Olivia), 2016-04-13 00:59:55]

xlai@chromium.org changed reviewers:
+ ajuma@chromium.org, junov@chromium.org

[xlai (Olivia), 2016-04-13 00:59:56]

Spec here (https://www.w3.org/TR/2dcontext/#drawing-paths-to-the-canvas) says:
Points with no lines coming out of them must have two caps placed back-to-back
as if it was really two points connected to each other by an infinitesimally
short straight line in the direction of the point's directionality.

Found the problem lies in the optimization in CanvasPathMethods. If we drop the
point in lineTo when the point is identifical to current point, then the SkPaint
is unable to recognize the existence of second point and do back-to-back caps.

At the same time, I'm also thinking whether we should drop the pre-optimization
in arcTo, bezierCurveTo, quadraticCurveTo in CanvasPathMethods as well. But by
doing some experiments on other browsers I find that they are also quite
inconsistent on that as well. Personally I feel that for the arc/curve, removing
the pre-optmization is not that necessary--why would users choose such a fancy
way to draw a simple zero-length short path? Therefore, I also fix the lineTo
issue.

[ajuma, 2016-04-13 14:26:48]

lgtm

[xlai (Olivia), 2016-04-13 14:35:07]

The CQ bit was checked by xlai@chromium.org

[commit-bot: I haz the power, 2016-04-13 14:35:34]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1883693002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1883693002/1

[xlai (Olivia), 2016-04-13 14:41:39]

The CQ bit was unchecked by xlai@chromium.org

[Justin Novosad, 2016-04-18 20:39:59]

lgtm

[xlai (Olivia), 2016-04-18 21:59:42]

The CQ bit was checked by xlai@chromium.org

[commit-bot: I haz the power, 2016-04-18 22:00:17]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1883693002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1883693002/1

[xlai (Olivia), 2016-04-18 22:09:23]

The CQ bit was unchecked by xlai@chromium.org

[commit-bot: I haz the power, 2016-04-18 22:09:49]

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

[commit-bot: I haz the power, 2016-04-18 22:09:50]

Try jobs failed on following builders:
  chromium_presubmit on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[xlai (Olivia), 2016-04-18 22:32:53]

https://codereview.chromium.org/1883693002/diff/20001/third_party/WebKit/Layo...
File third_party/WebKit/LayoutTests/TestExpectations (right):

https://codereview.chromium.org/1883693002/diff/20001/third_party/WebKit/Layo...
third_party/WebKit/LayoutTests/TestExpectations:114: crbug.com/577655
fast/canvas/canvas-zero-length-lineCap.html [ NeedsManualRebaseline ]
We can't use NeedsRebaseline at this moment.  crbug.com/603753
I'll do rebaseline manually.

Other changes in this file are just rebase from master branch.

[xlai (Olivia), 2016-04-18 22:33:26]

The CQ bit was checked by xlai@chromium.org

[xlai (Olivia), 2016-04-18 22:33:26]

The patchset sent to the CQ was uploaded after l-g-t-m from ajuma@chromium.org,
junov@chromium.org
Link to the patchset: https://codereview.chromium.org/1883693002/#ps20001
(title: "NeedsManualRebaseline")

[commit-bot: I haz the power, 2016-04-18 22:34:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1883693002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1883693002/20001

[commit-bot: I haz the power, 2016-04-18 23:58:25]

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

[commit-bot: I haz the power, 2016-04-18 23:58:26]

Try jobs failed on following builders:
  win_chromium_rel_ng on tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)

[xlai (Olivia), 2016-04-19 13:53:21]

The CQ bit was checked by xlai@chromium.org

[xlai (Olivia), 2016-04-19 13:53:21]

The patchset sent to the CQ was uploaded after l-g-t-m from ajuma@chromium.org,
junov@chromium.org
Link to the patchset: https://codereview.chromium.org/1883693002/#ps40001
(title: "Corrected outdated philip tests")

[commit-bot: I haz the power, 2016-04-19 13:53:37]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1883693002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1883693002/40001

[commit-bot: I haz the power, 2016-04-19 14:47:24]

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

[commit-bot: I haz the power, 2016-04-19 14:47:25]

Try jobs failed on following builders:
  mac_chromium_rel_ng on tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/mac_chromium_rel_...)

[xlai (Olivia), 2016-04-20 17:56:25]

Description was changed from

==========
Zero-length path with round/square lineCap should paint circle on Canvas

Currently CanvasPathMethods do a premature optimization in addLineTo--it drops
the point when the point is identical to currently point. The results in
incorrect rendering of zero-length paths with caps, based on the specification
(https://www.w3.org/TR/2dcontext/#drawing-paths-to-the-canvas).

BUG=322807,577655
==========

to

==========
Zero-length path with round/square lineCap should paint circle on Canvas

Currently CanvasPathMethods does a premature optimization in 
stroking line, curve, arc, rect---it drops the point when the
 point is identical to current point, or in the case or 
curve, drops the curves when the two given points are 
identical. The results in incorrect rendering of zero-length 
paths with caps, based on the specification
(https://www.w3.org/TR/2dcontext/#drawing-paths-to-the-canvas).

The five layout tests in philip/tests/2d.path.stroke.prune.*.html
are outdated canvas tests. Based on the spec change and the test 
results here
 (https://philip.html5.org/tests/canvas/suite/tests/results.html), 
other browsers report false positive errors on these 
tests too. We thus match what other browsers do in this 
aspect and rebaseline the expected outputs.

BUG=322807,577655
==========

[xlai (Olivia), 2016-04-20 17:58:45]

The CQ bit was checked by xlai@chromium.org

[xlai (Olivia), 2016-04-20 17:58:46]

The patchset sent to the CQ was uploaded after l-g-t-m from ajuma@chromium.org,
junov@chromium.org
Link to the patchset: https://codereview.chromium.org/1883693002/#ps60001
(title: "Rebaseline outdated philip tests to match IE9 and Firefox")

[commit-bot: I haz the power, 2016-04-20 17:59:07]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1883693002/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1883693002/60001

[commit-bot: I haz the power, 2016-04-20 19:34:59]

Description was changed from

==========
Zero-length path with round/square lineCap should paint circle on Canvas

Currently CanvasPathMethods does a premature optimization in 
stroking line, curve, arc, rect---it drops the point when the
 point is identical to current point, or in the case or 
curve, drops the curves when the two given points are 
identical. The results in incorrect rendering of zero-length 
paths with caps, based on the specification
(https://www.w3.org/TR/2dcontext/#drawing-paths-to-the-canvas).

The five layout tests in philip/tests/2d.path.stroke.prune.*.html
are outdated canvas tests. Based on the spec change and the test 
results here
 (https://philip.html5.org/tests/canvas/suite/tests/results.html), 
other browsers report false positive errors on these 
tests too. We thus match what other browsers do in this 
aspect and rebaseline the expected outputs.

BUG=322807,577655
==========

to

==========
Zero-length path with round/square lineCap should paint circle on Canvas

Currently CanvasPathMethods does a premature optimization in 
stroking line, curve, arc, rect---it drops the point when the
 point is identical to current point, or in the case or 
curve, drops the curves when the two given points are 
identical. The results in incorrect rendering of zero-length 
paths with caps, based on the specification
(https://www.w3.org/TR/2dcontext/#drawing-paths-to-the-canvas).

The five layout tests in philip/tests/2d.path.stroke.prune.*.html
are outdated canvas tests. Based on the spec change and the test 
results here
 (https://philip.html5.org/tests/canvas/suite/tests/results.html), 
other browsers report false positive errors on these 
tests too. We thus match what other browsers do in this 
aspect and rebaseline the expected outputs.

BUG=322807,577655
==========

[commit-bot: I haz the power, 2016-04-20 19:35:00]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2016-04-22 19:25:29]

Description was changed from

==========
Zero-length path with round/square lineCap should paint circle on Canvas

Currently CanvasPathMethods does a premature optimization in 
stroking line, curve, arc, rect---it drops the point when the
 point is identical to current point, or in the case or 
curve, drops the curves when the two given points are 
identical. The results in incorrect rendering of zero-length 
paths with caps, based on the specification
(https://www.w3.org/TR/2dcontext/#drawing-paths-to-the-canvas).

The five layout tests in philip/tests/2d.path.stroke.prune.*.html
are outdated canvas tests. Based on the spec change and the test 
results here
 (https://philip.html5.org/tests/canvas/suite/tests/results.html), 
other browsers report false positive errors on these 
tests too. We thus match what other browsers do in this 
aspect and rebaseline the expected outputs.

BUG=322807,577655
==========

to

==========
Zero-length path with round/square lineCap should paint circle on Canvas

Currently CanvasPathMethods does a premature optimization in
stroking line, curve, arc, rect---it drops the point when the
 point is identical to current point, or in the case or
curve, drops the curves when the two given points are
identical. The results in incorrect rendering of zero-length
paths with caps, based on the specification
(https://www.w3.org/TR/2dcontext/#drawing-paths-to-the-canvas).

The five layout tests in philip/tests/2d.path.stroke.prune.*.html
are outdated canvas tests. Based on the spec change and the test
results here
 (https://philip.html5.org/tests/canvas/suite/tests/results.html),
other browsers report false positive errors on these
tests too. We thus match what other browsers do in this
aspect and rebaseline the expected outputs.

BUG=322807,577655

Committed: https://crrev.com/245b756e89823ae1e668bbc52cf57493c87fb0bd
Cr-Commit-Position: refs/heads/master@{#388551}
==========

[commit-bot: I haz the power, 2016-04-22 19:25:31]

Patchset 4 (id:??) landed as
https://crrev.com/245b756e89823ae1e668bbc52cf57493c87fb0bd
Cr-Commit-Position: refs/heads/master@{#388551}