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

Issue 118755 link

Starred by 148 users

Issue metadata

Status: Fixed
Owner:
Email to this user bounced
Closed: Mar 2012
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Mac
Pri: 1
Type: Bug-Regression

Restricted
  • Only users with EditIssue permission may comment.



Sign in to add a comment

REGRESSION: Extension buttons visual corruption and incorrect badges (and download button corruption)

Reported by amiag...@ghostery.com, Mar 17 2012

Issue description

Started with today's Canary update.

I'm also seeing an extension's (Ghostery's) button badge appear over an unrelated extension's button.

Screenshots attached.

Chrome Version       : 19.0.1072.0
OS Version: OS X 10.6.8
UserAgentString: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_6_8) AppleWebKit/536.4 (KHTML, like Gecko) Chrome/19.0.1072.0 Safari/536.4
 
Screen shot 2012-03-17 at 11.50.15 AM.png
96.0 KB View Download
Screen shot 2012-03-17 at 11.45.53 AM.png
14.9 KB View Download

Comment 1 by dhw@chromium.org, Mar 17 2012

Labels: -Area-Undefined -Pri-2 -Type-Bug Area-UI Pri-1 Type-Regression Feature-Extensions Mstone-19 ReleaseBlock-Beta
Status: Untriaged
Summary: REGRESSION: Extension buttons visual corruption and incorrect badges

Comment 2 by dhw@chromium.org, Mar 17 2012

Cc: aa@chromium.org meh...@chromium.org erikkay@chromium.org

Comment 3 by dhw@chromium.org, Mar 17 2012

 Issue 118756  has been merged into this issue.

Comment 4 by dhw@chromium.org, Mar 17 2012

Summary: REGRESSION: Extension buttons visual corruption and incorrect badges (and download button corruption)
From dup'd  Issue 118756 , visual corruption also affects the download percentage spinner.

Comment 5 by dhw@chromium.org, Mar 17 2012

 Issue 118771  has been merged into this issue.

Comment 6 by meh...@chromium.org, Mar 17 2012

Owner: meh...@chromium.org
Checking the regression range ...

Comment 7 by meh...@chromium.org, Mar 17 2012

Cc: abarth@chromium.org rsesek@chromium.org
Owner: ----
It seems that there are two regressions (one in the extension pop-up menu and another one on the toolbar extension icon.)

-- Extension pop-up menu:

http://build.chromium.org/f/chromium/perf/dashboard/ui/changelog.html?url=&range=126812%3A126822&mode=html

Maybe caused by Webkit role /trunk/src/DEPS ?


-- Toolbar extension icon:

http://build.chromium.org/f/chromium/perf/dashboard/ui/changelog.html?url=&range=127195%3A127213&mode=html

Maybe caused by http://codereview.chromium.org/9703084 ?

Comment 8 by meh...@chromium.org, Mar 17 2012

 Issue 118753  has been merged into this issue.

Comment 9 by meh...@chromium.org, Mar 17 2012

 Issue 118752  has been merged into this issue.
 Issue 118743  has been merged into this issue.

Comment 11 by dhw@chromium.org, Mar 18 2012

 Issue 118773  has been merged into this issue.
 Issue 118809  has been merged into this issue.
 Issue 118811  has been merged into this issue.

Comment 14 Deleted

Comment 15 Deleted

Cc: mark@chromium.org
 Issue 118826  has been merged into this issue.
Owner: rsesek@chromium.org
Status: Assigned
I doubt it's me, but I'll find the owner.
Cc: jayakrishnat@chromium.org gailh@chromium.org arsids@chromium.org ranavdk@chromium.org vandanashah@chromium.org
Labels: GoogleFeedback
Chrome version: 19.0.1074.0
User states that icons of the installed extensions, beside the wrench icon in Google chrome are not rendered properly. 
Graphics on the extension buttons are distorted with blank square boxes or displayed in broken lines. Users reported that this issue happens only to Extensions icons and the default Chrome toolbar icons display fine.

For more details, refer the user reports below:
http://goto.google.com/186590124
http://goto.google.com/218777693
http://goto.google.com/187026439
http://goto.google.com/187013454
http://goto.google.com/186334214

For further more user reports, refer the below cluster URL:
http://goto.google.com/2847914

~40 users have reported this issue in 'GoogleFeeback' since this weekend.
Reproduced in Lion, 10.7.3 (macbook pro 13-inch, Mid 2010, Mac OS X Lion 10.7.3 (11D50b), NVIDIA GeForce 320M 256 MB). 
 Issue 118847  has been merged into this issue.
 Issue 118901  has been merged into this issue.
 Issue 118905  has been merged into this issue.
Owner: twiz@chromium.org
I reverted http://src.chromium.org/viewvc/chrome?view=rev&revision=127196 locally and could no longer repro this.
 Issue 118947  has been merged into this issue.
 Issue 118988  has been merged into this issue.

Comment 27 Deleted

Comment 28 by Deleted ...@, Mar 19 2012

Well I had a strange build number of 180.1.1025.40 Canary and I was no longer seeing it, but I upgraded again and now it's on 19.0.1074.0 Canary and it's showing it again.
Labels: Restrict-AddIssueComment-Commit
This has been confirmed and it has not yet been fixed.

Comment 30 by dhw@chromium.org, Mar 20 2012

 Issue 119025  has been merged into this issue.
 Issue 119034  has been merged into this issue.

Comment 32 by twiz@chromium.org, Mar 20 2012

This is likely related to my change, noted in comment #24.

I'm having a look.  Hopefully it will be a simple fix for the mac, otherwise we can revert my optimization if the M19 branch point gets too close.
 Issue 119085  has been merged into this issue.
 Issue 119079  has been merged into this issue.
 Issue 119077  has been merged into this issue.
 Issue 119120  has been merged into this issue.

Comment 37 by twiz@chromium.org, Mar 20 2012

I can confirm that this problem is related to my change:  http://src.chromium.org/viewvc/chrome?view=rev&revision=127196

The simple fix is to re-introduce the memory clear removed by my change for the mac platform:   http://codereview.chromium.org/9416017/diff/92010/skia/ext/bitmap_platform_device_mac.cc

This fix will cause a performance degradation.  The real fix is to find the call site on Mac that is assuming an initialized buffer when displaying the extension icons.

I'm OOO tomorrow (03/21/12), but will address the problem on Thurs. 
 Issue 119190  has been merged into this issue.
 Issue 119263  has been merged into this issue.
 Issue 119365  has been merged into this issue.
Project Member

Comment 41 by bugdroid1@chromium.org, Mar 22 2012

The following revision refers to this bug:
    http://src.chromium.org/viewvc/chrome?view=rev&revision=128271

------------------------------------------------------------------------
r128271 | twiz@chromium.org | Thu Mar 22 12:43:47 PDT 2012

Changed paths:
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/browser/ui/cocoa/extensions/browser_action_button.mm?r1=128271&r2=128270&pathrev=128271


Correct the rendering of BrowserActionIcons on mac that was introduced by CL: http://src.chromium.org/viewvc/chrome?view=rev&revision=127196

The cocoa rendering code was assuming that the CanvasSkiaPaint object was initialized to 0, which is no longer true.  This CL explicitly initializes the canvas before rendering the icon.

BUG= 118755 
TEST=None

Review URL: http://codereview.chromium.org/9839012
------------------------------------------------------------------------

Comment 42 by twiz@chromium.org, Mar 22 2012

Status: Fixed
 Issue 119567  has been merged into this issue.
Cc: ligim...@chromium.org dharani@chromium.org pavanv@chromium.org anan...@chromium.org
 Issue 119565  has been merged into this issue.

Comment 45 by levin@chromium.org, Mar 23 2012

It looks like the fix was for browser action buttons.

Although that was the primary place that people noticed it, it also occurred in the download icons. See http://code.google.com/p/chromium/issues/detail?id=118756

Did the fix address those as well? Is there some deeper issue (since this affects multiple places)?


Comment 46 by levin@chromium.org, Mar 23 2012

Status: Assigned
To answer my own question, here's a code search for places that use CanvasSkiaPaint in Objective C code:

http://code.google.com/p/chromium/source/search?q=gfx%3A%3ACanvasSkiaPaint+file%3A%5C.mm&origq=gfx%3A%3ACanvasSkiaPaint+file%3A%5C.mm&btnG=Search+Trunk

This patch fixes one of them.  download_item_cell.mm must be the place where the download icon is drawn.

In the third place it uses "true" as the second parameter so maybe that alleviates the need for this fix. I may have missed some places in my search above so feel free to search for more :).

Comment 47 by twiz@chromium.org, Mar 23 2012

Levin@: Thanks for the notification. 

Yes, my previous fix was specific to browser action buttons.

I'm worried about the possibility for similar regressions hiding on windows.  My original change had the intent to remove the unnecessary clear from PlatformCanvas, but propagating that change through to users of CanvasSkiaPaint caused these regressions.  A safe fix that would prevent all of these regressions, on all platforms, would be to introduce the clear at the level of CanvasSkiaPaint.  

By doing so, the performance improvement by avoiding clears for PlatformCanvas users is preserved, and these regressions are prevented.

I'll submit a change for review today.

Project Member

Comment 48 by bugdroid1@chromium.org, Mar 23 2012

The following revision refers to this bug:
    http://src.chromium.org/viewvc/chrome?view=rev&revision=128574

------------------------------------------------------------------------
r128574 | twiz@chromium.org | Fri Mar 23 14:54:24 PDT 2012

Changed paths:
 M http://src.chromium.org/viewvc/chrome/trunk/src/skia/ext/canvas_paint_x.h?r1=128574&r2=128573&pathrev=128574
 M http://src.chromium.org/viewvc/chrome/trunk/src/skia/ext/canvas_paint_wayland.h?r1=128574&r2=128573&pathrev=128574
 M http://src.chromium.org/viewvc/chrome/trunk/src/skia/ext/canvas_paint_win.h?r1=128574&r2=128573&pathrev=128574
 M http://src.chromium.org/viewvc/chrome/trunk/src/skia/ext/canvas_paint_gtk.h?r1=128574&r2=128573&pathrev=128574
 M http://src.chromium.org/viewvc/chrome/trunk/src/skia/ext/canvas_paint_mac.h?r1=128574&r2=128573&pathrev=128574
 M http://src.chromium.org/viewvc/chrome/trunk/src/chrome/browser/ui/cocoa/extensions/browser_action_button.mm?r1=128574&r2=128573&pathrev=128574

Default initialize CanvasPaintT<T> bitmaps to 0, 0, 0, 0.

PlatformCanvas objects are constructed uninitialized.  (Since change http://src.chromium.org/viewvc/chrome?view=rev&revision=127196).  This change had propagated uninitialized semantics to CanvasPaint<..>, and CanvasPaintSkia, which introduced the regressions mentioned in the bug below.

This change re-introduces the initilized memory behaviour, but at the CanvasPaint<T> level.  This preserves the performance boost from rev 127196, and prevents regressions in displayed UI elements.

Change also removes a now-unnecessary clear in browser_action_button.mm introduced by http://codereview.chromium.org/9839012.

BUG= 118755 
TEST=None
Review URL: https://chromiumcodereview.appspot.com/9837043
------------------------------------------------------------------------
 Issue 119874  has been merged into this issue.
 Issue 119898  has been merged into this issue.

Comment 51 by dhw@chromium.org, Mar 24 2012

 Issue 119935  has been merged into this issue.

Comment 52 by dhw@chromium.org, Mar 24 2012

 Issue 119932  has been merged into this issue.

Comment 53 by dhw@chromium.org, Mar 24 2012

 Issue 119933  has been merged into this issue.

Comment 54 by dhw@chromium.org, Mar 24 2012

 Issue 119936  has been merged into this issue.

Comment 55 by dhw@chromium.org, Mar 24 2012

 Issue 119931  has been merged into this issue.

Comment 56 by dhw@chromium.org, Mar 25 2012

 Issue 119955  has been merged into this issue.

Comment 57 by dhw@chromium.org, Mar 25 2012

 Issue 119952  has been merged into this issue.

Comment 58 by dhw@chromium.org, Mar 25 2012

 Issue 119891  has been merged into this issue.
 Issue 119993  has been merged into this issue.
Issue 119980 has been merged into this issue.
 Issue 119972  has been merged into this issue.
 Issue 119941  has been merged into this issue.
 Issue 119998  has been merged into this issue.
Cc: srihariraju@chromium.org
 Issue 120088  has been merged into this issue.

Comment 65 by dhw@chromium.org, Mar 26 2012

 Issue 120109  has been merged into this issue.

Comment 66 by dhw@chromium.org, Mar 26 2012

 Issue 120120  has been merged into this issue.

Comment 67 by dhw@chromium.org, Mar 26 2012

 Issue 120045  has been merged into this issue.

Comment 68 by dhw@chromium.org, Mar 26 2012

 Issue 120015  has been merged into this issue.

Comment 69 by dhw@chromium.org, Mar 26 2012

 Issue 120089  has been merged into this issue.

Comment 70 by dhw@chromium.org, Mar 26 2012

 Issue 120138  has been merged into this issue.

Comment 71 by twiz@chromium.org, Mar 26 2012

Status: Fixed
 Issue 120187  has been merged into this issue.
 Issue 120166  has been merged into this issue.
 Issue 120233  has been merged into this issue.
 Issue 120231  has been merged into this issue.
 Issue 120227  has been merged into this issue.
 Issue 120283  has been merged into this issue.
Cc: backer@chromium.org kbr@chromium.org thakis@chromium.org
 Issue 120345  has been merged into this issue.

Comment 79 by dharani@google.com, Mar 27 2012

Cc: jeffreyc@chromium.org jhurwich@chromium.org
 Issue 120364  has been merged into this issue.
 Issue 120413  has been merged into this issue.
 Issue 120505  has been merged into this issue.

Comment 82 by a...@chromium.org, Mar 27 2012

 Issue 120523  has been merged into this issue.

Comment 83 by dhw@chromium.org, Mar 27 2012

 Issue 120524  has been merged into this issue.
 Issue 120655  has been merged into this issue.
Project Member

Comment 85 by bugdroid1@chromium.org, Mar 9 2013

Labels: -Area-UI -Type-Regression -Feature-Extensions -Mstone-19 Type-Bug-Regression Cr-Platform-Extensions Cr-UI M-19
Project Member

Comment 86 by bugdroid1@chromium.org, Mar 14 2013

Labels: -Restrict-AddIssueComment-Commit Restrict-AddIssueComment-EditIssue

Comment 87 by laforge@google.com, Jul 24 2013

Cc: -jeffreyc@chromium.org

Sign in to add a comment