skip zero-length linear gradient intervals to make new code more consistent with older table-based …

skip zero-length linear gradient intervals to make new code more consistent with older table-based implementation

BUG=skia:4841
GOLD_TRYBOT_URL= https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&issue=1619413002

Committed: https://skia.googlesource.com/skia/+/f2b8662b5c73e03648ed1a154b717e354753a0e1

[lsalzman1, 2016-01-22 16:07:51]

Description was changed from

==========
skip zero-length linear gradient intervals to make new code more consistent with
older table-based implementation

BUG=skia:4841
==========

to

==========
skip zero-length linear gradient intervals to make new code more consistent with
older table-based implementation

BUG=skia:4841
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[lsalzman1, 2016-01-22 16:09:25]

Description was changed from

==========
skip zero-length linear gradient intervals to make new code more consistent with
older table-based implementation

BUG=skia:4841
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
skip zero-length linear gradient intervals to make new code more consistent with
older table-based implementation

BUG=skia:4841
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

[lsalzman1, 2016-01-22 16:09:26]

lsalzman@mozilla.com changed reviewers:
+ reed@google.com

[lsalzman1, 2016-01-22 16:10:41]

This makes the new linear gradient handling consistent with the table-based
implementation in so far as the old table code skipped zero-length intervals.
Our Firefox unit tests expect this behavior to match our other platforms, so
this patch fixes those unit tests.

It also fixes a divide by zero issue where fPosScale was getting set to inf and
messing up later interpolation math.

[mtklein, 2016-01-22 16:34:04]

The CQ bit was checked by mtklein@google.com to run a CQ dry run

[commit-bot: I haz the power, 2016-01-22 16:34:15]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1619413002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1619413002/1

[mtklein, 2016-01-22 16:34:33]

(don't mind me, just kicking off the trybots to see the diffs out of curiosity)

[commit-bot: I haz the power, 2016-01-22 16:51:16]

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

[commit-bot: I haz the power, 2016-01-22 16:51:17]

Dry run: This issue passed the CQ dry run.

[reed1, 2016-01-22 18:32:29]

I'm adding a new GM to help illustrate the bad pixels...

[reed1, 2016-01-22 21:59:23]

Inspired me to consider a larger fix for the future

https://bugs.chromium.org/p/skia/issues/detail?id=4844

lgtm

[lsalzman1, 2016-01-22 22:01:47]

The CQ bit was checked by lsalzman@mozilla.com

[commit-bot: I haz the power, 2016-01-22 22:02:24]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1619413002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1619413002/1

[commit-bot: I haz the power, 2016-01-22 22:03:05]

Description was changed from

==========
skip zero-length linear gradient intervals to make new code more consistent with
older table-based implementation

BUG=skia:4841
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...
==========

to

==========
skip zero-length linear gradient intervals to make new code more consistent with
older table-based implementation

BUG=skia:4841
GOLD_TRYBOT_URL=
https://gold.skia.org/search2?unt=true&query=source_type%3Dgm&master=false&is...

Committed:
https://skia.googlesource.com/skia/+/f2b8662b5c73e03648ed1a154b717e354753a0e1
==========

[commit-bot: I haz the power, 2016-01-22 22:03:05]

Committed patchset #1 (id:1) as
https://skia.googlesource.com/skia/+/f2b8662b5c73e03648ed1a154b717e354753a0e1