New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 760807 link

Starred by 2 users

Issue metadata

Status: Fixed
Owner:
Closed: Sep 2017
Cc:
EstimatedDays: ----
NextAction: ----
OS: Mac
Pri: 1
Type: Bug-Regression



Sign in to add a comment

Iframes on HiDPI machine in low-res

Project Member Reported by a...@chromium.org, Aug 31 2017

Issue description

OS: Seen on macOS 10.11, 10.12

What happens is that an iframe is shown in low-res. The contents fill the extents of the iframe but it is fuzzy, as if it had been rendered in 1x and then upscaled to 2x to fit. I have seen it in G+ popups in the bell bubble on Google sites, but the easiest way is to watch an embedded YouTube video.

What steps will reproduce the problem?
(1) Go to https://twitter.com/based_ES/status/902305420471087104
(2) Click the embedded video to start playing the video

What is the expected result?
What should happen is that the video plays at full resolution.

What happens instead?
What actually happens is that the video plays blurry.

I bisected:

You are probably looking for a change made after 491725 (known good), but no later than 491734 (first known bad).
CHANGELOG URL:
  https://chromium.googlesource.com/chromium/src/+log/4ad909a931bd1f3e2b1a9861b251096c7b8f9b71..229c603a647d42f4754e6517e4376cf7f3805908

There are three graphics changes in that range, but none obviously stand out as what's broken:
https://chromium.googlesource.com/chromium/src/+/e47105d8341c6c7b21696e13c0cdcc5f6f714c13
> A Skia roll containing bot changes.
https://chromium.googlesource.com/chromium/src/+/a5f7079f3c63aa9cec376886fddb1141527bfd53
> A change to cc's SolidColorAnalyzer.
https://chromium.googlesource.com/chromium/src/+/229c603a647d42f4754e6517e4376cf7f3805908
> This is ANGLE which I don't believe is used on the Mac where I'm seeing this.

My best guess is cc's SolidColorAnalyzer though I don't understand why. Assigning.
 

Comment 1 by a...@chromium.org, Aug 31 2017

Labels: -Type-Bug Type-Bug-Regression

Comment 2 by ajha@chromium.org, Aug 31 2017

Cc: gov...@chromium.org ligim...@chromium.org wkorman@chromium.org
sunxd@: Could you please get the CL reverted if your change is related, as we are very close to M-62 branch.
M62 is branching today, Please take a look ASAP.

Comment 4 by kenrb@chromium.org, Aug 31 2017

I just tried a local revert of the CL that avi suggested in the first post, and it did indeed resolve the problem.
Great, could you please revert the same in trunk.
Labels: -ReleaseBlock-Dev ReleaseBlock-Beta
Doesn't seem to be to critical for Dev channel. But would like to have a fix before Beta. 

Comment 7 by sunxd@chromium.org, Sep 1 2017

I'll revert the change.

Comment 8 by ajha@chromium.org, Sep 5 2017

Friendly ping to get an update on the revert.
Your bug is tagged as RBB, please have a fix/ revert ASAP.
Cc: xidac...@chromium.org
Hi xidachen@, can you help me with reverting the CL as I'm travelling? There is a merge conflict.
Project Member

Comment 12 by bugdroid1@chromium.org, Sep 6 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/8d1900b0fd708cbe8b298a4f5d8a4cb566007255

commit 8d1900b0fd708cbe8b298a4f5d8a4cb566007255
Author: Xida Chen <xidachen@chromium.org>
Date: Wed Sep 06 19:19:19 2017

Manually revert: cc: Support ClipRRect in SolidColorAnalyzer

This CL reverts a previous CL here:
https://chromium-review.googlesource.com/c/chromium/src/+/591913

The previous CL caused a regression on Mac. GL renderer has
difficulties dealing with RPDQs converted from
SolidColorDrawQuads, this CL will make mask layer produce
solid quads. We need to keep it reverted before we fix the
gl renderer bug.

NOTRY=true

Bug:  760807 
Cq-Include-Trybots: master.tryserver.blink:linux_trusty_blink_rel
Change-Id: I844d02886329acbbbebd98200f8d2f90681adb8a
Reviewed-on: https://chromium-review.googlesource.com/653317
Commit-Queue: Xida Chen <xidachen@chromium.org>
Reviewed-by: Vladimir Levin <vmpstr@chromium.org>
Reviewed-by: Xianda Sun <sunxd@chromium.org>
Cr-Commit-Position: refs/heads/master@{#500036}
[modify] https://crrev.com/8d1900b0fd708cbe8b298a4f5d8a4cb566007255/cc/paint/solid_color_analyzer.cc
[modify] https://crrev.com/8d1900b0fd708cbe8b298a4f5d8a4cb566007255/cc/paint/solid_color_analyzer_unittest.cc

This likely need to be merged to M62.
Please let me know which branch and I will merge

Comment 14 by a...@chromium.org, Sep 6 2017

The M62 branch is 3202.
Project Member

Comment 15 by bugdroid1@chromium.org, Sep 6 2017

Labels: merge-merged-3202
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/c4aff051f7e47011ace905719ea3898ea58d4d17

commit c4aff051f7e47011ace905719ea3898ea58d4d17
Author: Xida Chen <xidachen@chromium.org>
Date: Wed Sep 06 20:14:33 2017

Manually revert: cc: Support ClipRRect in SolidColorAnalyzer

This CL reverts a previous CL here:
https://chromium-review.googlesource.com/c/chromium/src/+/591913

The previous CL caused a regression on Mac. GL renderer has
difficulties dealing with RPDQs converted from
SolidColorDrawQuads, this CL will make mask layer produce
solid quads. We need to keep it reverted before we fix the
gl renderer bug.

NOTRY=true
TBR=xidachen@chromium.org

(cherry picked from commit 8d1900b0fd708cbe8b298a4f5d8a4cb566007255)

Bug:  760807 
Cq-Include-Trybots: master.tryserver.blink:linux_trusty_blink_rel
Change-Id: I844d02886329acbbbebd98200f8d2f90681adb8a
Reviewed-on: https://chromium-review.googlesource.com/653317
Commit-Queue: Xida Chen <xidachen@chromium.org>
Reviewed-by: Vladimir Levin <vmpstr@chromium.org>
Reviewed-by: Xianda Sun <sunxd@chromium.org>
Cr-Original-Commit-Position: refs/heads/master@{#500036}
Reviewed-on: https://chromium-review.googlesource.com/653803
Reviewed-by: Xida Chen <xidachen@chromium.org>
Cr-Commit-Position: refs/branch-heads/3202@{#55}
Cr-Branched-From: fa6a5d87adff761bc16afc5498c3f5944c1daa68-refs/heads/master@{#499098}
[modify] https://crrev.com/c4aff051f7e47011ace905719ea3898ea58d4d17/cc/paint/solid_color_analyzer.cc
[modify] https://crrev.com/c4aff051f7e47011ace905719ea3898ea58d4d17/cc/paint/solid_color_analyzer_unittest.cc

merged
Cc: abdulsyed@chromium.org
Cl listed at #15 got merged to M62 without merge request and approval.

+ abdulsyed@ (M62 Desktop TPM) as FYI

Comment 18 by sunxd@chromium.org, Sep 11 2017

Chris has reverted the mask tiling patch which also fixes this issue, I think we no longer need to merge this patch into M61.

Comment 19 by sunxd@chromium.org, Sep 11 2017

Labels: Merge-Request-61
On the other bug, Chris has requested merging into M62. But I think the issue exists in M61.
Project Member

Comment 20 by sheriffbot@chromium.org, Sep 11 2017

Labels: -Merge-Request-61 Merge-Review-61 Hotlist-Merge-Review
This bug requires manual review: Request affecting a post-stable build
Please contact the milestone owner if you have questions.
Owners: amineer@(Android), cmasso@(iOS), ketakid@(ChromeOS), govind@(Desktop)

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Comment 21 by sunxd@chromium.org, Sep 11 2017

Labels: -Merge-Review-61
Sorry we don't need to merge into M61, the reverted cl was landed after M61.

Comment 22 by sunxd@chromium.org, Sep 11 2017

Status: Fixed (was: Assigned)
Can you verify if the issue is fixed?
Labels: TE-Verified-M63 TE-Verified-63.0.3212.0
Verified this issue on MacBookPro 10.12.6 using chrome latest canary #63.0.3212.0 by following steps mentioned in the original comment, observed the video plays in full resolution as expected. Hence adding TE-Verified label for M-63.

Note: Attaching screen-shot for reference, comparison between #62.0.3182.0 and 63.0.3212.0.

Thanks!  
760807.png
1.3 MB View Download
Project Member

Comment 24 by bugdroid1@chromium.org, Sep 29 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/71fe8359e1fff12a6bec2384a83b6440bca62b55

commit 71fe8359e1fff12a6bec2384a83b6440bca62b55
Author: sunxd <sunxd@chromium.org>
Date: Fri Sep 29 17:51:25 2017

Relanding: cc: Support ClipRRect in SolidColorAnalyzer

We previously reverted the SolidColorAnalyzer patch because
of mask tiling regressions:

Some texts are rendered blurry when mask tiling is enabled
along with this patch. This patch was reverted because
bisect points at this one.

But actually, it turns out that the root cause is that ca
layers cannot work with solid color mask layer correctly,
there is nothing wrong with this patch. So we are relanding
it.

We also disabled mask tiling in
https://chromium-review.googlesource.com/c/chromium/src/+/661072,
and are fixing the bug in
https://chromium-review.googlesource.com/c/chromium/src/+/685716.

This reverts commit 8d1900b0fd708cbe8b298a4f5d8a4cb566007255.

Reason for revert: The original bug will not reproduce again since mask tiling is disabled. We're also landing the patch that fixes the root cause of the bug. The reverted patch proves to be correct. The solid color analyzer change can be re-land.

Original change's description:
> Manually revert: cc: Support ClipRRect in SolidColorAnalyzer
> 
> This CL reverts a previous CL here:
> https://chromium-review.googlesource.com/c/chromium/src/+/591913
> 
> The previous CL caused a regression on Mac. GL renderer has
> difficulties dealing with RPDQs converted from
> SolidColorDrawQuads, this CL will make mask layer produce
> solid quads. We need to keep it reverted before we fix the
> gl renderer bug.
> 
> NOTRY=true
> 
> Bug:  760807 
> Cq-Include-Trybots: master.tryserver.blink:linux_trusty_blink_rel
> Change-Id: I844d02886329acbbbebd98200f8d2f90681adb8a
> Reviewed-on: https://chromium-review.googlesource.com/653317
> Commit-Queue: Xida Chen <xidachen@chromium.org>
> Reviewed-by: Vladimir Levin <vmpstr@chromium.org>
> Reviewed-by: Xianda Sun <sunxd@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#500036}

TBR=flackr@chromium.org,vmpstr@chromium.org,chrishtr@chromium.org,wkorman@chromium.org,xidachen@chromium.org,sunxd@chromium.org

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug:  760807 
Change-Id: I96969d458f14bc1129d8a9b4b7b22bd562806a92
Cq-Include-Trybots: master.tryserver.blink:linux_trusty_blink_rel
Reviewed-on: https://chromium-review.googlesource.com/674790
Reviewed-by: vmpstr <vmpstr@chromium.org>
Reviewed-by: Robert Flack <flackr@chromium.org>
Reviewed-by: Xianda Sun <sunxd@chromium.org>
Commit-Queue: Xianda Sun <sunxd@chromium.org>
Cr-Commit-Position: refs/heads/master@{#505417}
[modify] https://crrev.com/71fe8359e1fff12a6bec2384a83b6440bca62b55/cc/paint/solid_color_analyzer.cc
[modify] https://crrev.com/71fe8359e1fff12a6bec2384a83b6440bca62b55/cc/paint/solid_color_analyzer_unittest.cc
[modify] https://crrev.com/71fe8359e1fff12a6bec2384a83b6440bca62b55/cc/trees/draw_property_utils.cc

Comment 25 by sunxd@chromium.org, Sep 29 2017

The original bug is caused by mac overlay, relanding the previous reverted patch since that one does not break it.

Sign in to add a comment