SP: Always anti-alias borders in slimming paint mode.

SP: Always anti-alias borders in slimming paint mode.

BUG=382491, 502531
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=197528

[chrishtr, 2015-06-19 21:52:24]

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

[chrishtr, 2015-06-19 21:52:25]

See regressions in painting diagonals for inline element outlines in
fast/css/outline-narrowLine-diffs.html and
fast/borders/outline-alpha-inline-diffs.html. Maybe this is uncommon enough to
be acceptable?

[fs, 2015-06-19 23:52:59]

On 2015/06/19 21:52:25, chrishtr wrote:
> See regressions in painting diagonals for inline element outlines in
> fast/css/outline-narrowLine-diffs.html and
> fast/borders/outline-alpha-inline-diffs.html. Maybe this is uncommon enough to
> be acceptable?

Well, all of them have diagonal seaming, it's just more apparent on the wider
outlines I guess. I have a hard time imagining that wide outlines are that
common, so it's probably ok to choose this path for the time being.

Maybe another possibility to consider would be to always set antialias to false
(for SP) instead in InlinePainter? Means outlines might "pop" instead during
zooming... So "seams" (bleeding) vs. "popping"...

Regardless, could ve get a bug about fixing this? (If there isn't one already.)
Regardless of direction of choice.

[chrishtr, 2015-06-20 02:44:12]

On 2015/06/19 at 23:52:59, fs wrote:
> On 2015/06/19 21:52:25, chrishtr wrote:
> > See regressions in painting diagonals for inline element outlines in
> > fast/css/outline-narrowLine-diffs.html and
> > fast/borders/outline-alpha-inline-diffs.html. Maybe this is uncommon enough
to
> > be acceptable?
> 
> Well, all of them have diagonal seaming, it's just more apparent on the wider
outlines I guess. I have a hard time imagining that wide outlines are that
common, so it's probably ok to choose this path for the time being.
> 
> Maybe another possibility to consider would be to always set antialias to
false (for SP) instead in InlinePainter? Means outlines might "pop" instead
during zooming... So "seams" (bleeding) vs. "popping"...

Good idea, done. All tests now pass.

> 
> Regardless, could ve get a bug about fixing this? (If there isn't one
already.) Regardless of direction of choice.

Done, crbug.com/502531.

[f(malita), 2015-06-20 05:04:55]

On 2015/06/20 at 02:44:12, chrishtr wrote:
> On 2015/06/19 at 23:52:59, fs wrote:
> > Regardless, could ve get a bug about fixing this? (If there isn't one
already.) Regardless of direction of choice.
> 
> Done, crbug.com/502531.

LGTM.

We can switch outline AA back on after we fix that bug.

[chrishtr, 2015-06-20 06:22:22]

The CQ bit was checked by chrishtr@chromium.org

[commit-bot: I haz the power, 2015-06-20 06:22:33]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1195933002/20001

[commit-bot: I haz the power, 2015-06-20 06:25:50]

Committed patchset #2 (id:20001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=197528