Align image sizes for SVG with raster image size

Align image sizes for SVG with raster image size

The background geometry calculations has heuristics for tiling image
sizes optimized for integer image sizes (at least in effective zoom ==
1). Rounding SVG image sizes before zoom application makes SVG images
fit better into the existing heuristics.

BUG=601011
Committed: https://crrev.com/c53a0d93dd13a025e65668dda879f78142459e04
Cr-Commit-Position: refs/heads/master@{#389158}

[davve, 2016-04-22 14:26:01]

Description was changed from

==========
Round SVG size used for tiling similar to the actual image size

BUG=601011
==========

to

==========
Align image sizes for SVG with raster image size

The background geometry calculations has heuristics for tiling image
sizes optimized for integer image sizes (at least in effective zoom ==
1). Rounding SVG image sizes before zoom application makes SVG images
fit better into the existing heuristics.

BUG=601011
==========

[davve, 2016-04-22 15:04:41]

davve@opera.com changed reviewers:
+ fs@opera.com, leviw@chromium.org

[davve, 2016-04-22 15:04:41]

Off-by-one tiling bugs are fun!

Note for sensitive readers: Epic editing fail lurking in the red/removed line
that happened to not ruin the world and actually didn't have anything to do with
the bug I'm addressing here. But fixing the double call to concreteObjectSize
seemed prudent.

It's all about the rounding in this patch. The main problem I attempt to address
is when the aspect ratio is used in combination with a default object size to
paint a background, e.g. a 1:0.99 aspect ratio SVG background for a 16x16 div
should cover 16x16 and not 16:15. The latter was sometimes the case before this
patch.

Also, this patch addresses a new bug but the problem seems old. We used to do
the calculations for SVG backgrounds in LayoutUnits but when we recently merged
algorithms, SVG background sizes are computed using floats instead (and
converted to LayoutUnits as the last step). LayoutUnits having worse precision
hid the bug on the example URL in the bug notably.

The rebaselines results should be visible in patch set #1 (for linux).

[davve, 2016-04-22 15:04:51]

The CQ bit was checked by davve@opera.com to run a CQ dry run

[commit-bot: I haz the power, 2016-04-22 15:05:13]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1912063004/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1912063004/20001

[fs, 2016-04-22 15:38:53]

On 2016/04/22 at 15:04:41, davve wrote:
> Off-by-one tiling bugs are fun!
> 
> Note for sensitive readers: Epic editing fail lurking in the red/removed line
that happened to not ruin the world and actually didn't have anything to do with
the bug I'm addressing here. But fixing the double call to concreteObjectSize
seemed prudent.
> 
> It's all about the rounding in this patch. The main problem I attempt to
address is when the aspect ratio is used in combination with a default object
size to paint a background, e.g. a 1:0.99 aspect ratio SVG background for a
16x16 div should cover 16x16 and not 16:15. The latter was sometimes the case
before this patch.
> 
> Also, this patch addresses a new bug but the problem seems old. We used to do
the calculations for SVG backgrounds in LayoutUnits but when we recently merged
algorithms, SVG background sizes are computed using floats instead (and
converted to LayoutUnits as the last step). LayoutUnits having worse precision
hid the bug on the example URL in the bug notably.
> 
> The rebaselines results should be visible in patch set #1 (for linux).

LGTM, but if subpixel-pixie Levi could have a look to that'd be great.

[commit-bot: I haz the power, 2016-04-22 16:44:08]

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

[commit-bot: I haz the power, 2016-04-22 16:44:09]

Dry run: This issue passed the CQ dry run.

[leviw_travelin_and_unemployed, 2016-04-22 17:49:51]

leviw@chromium.org changed reviewers:
+ eae@chromium.org, schenney@chromium.org

[leviw_travelin_and_unemployed, 2016-04-22 17:49:51]

Makes perfect sense. LGTM.

FYI: I'm flying the coop in 2 weeks, off to all manner of adventure (living in a
van, building houses, and motorcycling across Africa)!

In my absence, schenney@ is the background image guru, and for other various
sub-pixel issues eae@ is your reviewer in shining armor.

[leviw_travelin_and_unemployed, 2016-04-22 17:49:54]

The CQ bit was checked by leviw@chromium.org

[commit-bot: I haz the power, 2016-04-22 17:50:41]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1912063004/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1912063004/20001

[eae, 2016-04-22 17:54:06]

LGTM

[commit-bot: I haz the power, 2016-04-22 17:56:13]

Description was changed from

==========
Align image sizes for SVG with raster image size

The background geometry calculations has heuristics for tiling image
sizes optimized for integer image sizes (at least in effective zoom ==
1). Rounding SVG image sizes before zoom application makes SVG images
fit better into the existing heuristics.

BUG=601011
==========

to

==========
Align image sizes for SVG with raster image size

The background geometry calculations has heuristics for tiling image
sizes optimized for integer image sizes (at least in effective zoom ==
1). Rounding SVG image sizes before zoom application makes SVG images
fit better into the existing heuristics.

BUG=601011
==========

[commit-bot: I haz the power, 2016-04-22 17:56:14]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-04-22 19:49:29]

Description was changed from

==========
Align image sizes for SVG with raster image size

The background geometry calculations has heuristics for tiling image
sizes optimized for integer image sizes (at least in effective zoom ==
1). Rounding SVG image sizes before zoom application makes SVG images
fit better into the existing heuristics.

BUG=601011
==========

to

==========
Align image sizes for SVG with raster image size

The background geometry calculations has heuristics for tiling image
sizes optimized for integer image sizes (at least in effective zoom ==
1). Rounding SVG image sizes before zoom application makes SVG images
fit better into the existing heuristics.

BUG=601011

Committed: https://crrev.com/c53a0d93dd13a025e65668dda879f78142459e04
Cr-Commit-Position: refs/heads/master@{#389158}
==========

[commit-bot: I haz the power, 2016-04-22 19:49:31]

Patchset 2 (id:??) landed as
https://crrev.com/c53a0d93dd13a025e65668dda879f78142459e04
Cr-Commit-Position: refs/heads/master@{#389158}