Bug 77084 - [Chrome] Incorrect rendering of anti-aliased strokes
Summary: [Chrome] Incorrect rendering of anti-aliased strokes
Status: UNCONFIRMED
Alias: None
Product: WebKit
Classification: Unclassified
Component: SVG (show other bugs)
Version: 528+ (Nightly build)
Hardware: PC Linux
: P2 Normal
Assignee: Nobody
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-01-26 02:42 PST by Pieter Senster
Modified: 2012-01-30 02:33 PST (History)
2 users (show)

See Also:


Attachments
Testcae (1009 bytes, image/svg+xml)
2012-01-26 02:42 PST, Pieter Senster
no flags Details
Screenshot of actual results (screenshot scaled) (1.53 KB, image/png)
2012-01-26 02:50 PST, Pieter Senster
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description Pieter Senster 2012-01-26 02:42:17 PST
Created attachment 124084 [details]
Testcae

Steps to reproduce:

1) Open the attached testcase (strokebug.svg) on Chrome/Chromium Linux

Expected results:
There should be a square with a black border with nothing in it.

Actual results:
Artifacts of a path (an 'S' upside down) appear within the square. When increasing the scale by zooming in, the artifacts will disappear.

Builds:
Chrome 17.0.963.38 beta on Linux

Also occurs in: Chromium r118842 on Linux
Works as expected on: All Chrome versions (including Canary) on Mac, Safari 5 on Mac, Firefox 8 on Linux

Additional information:
Strokes seem to be blended with a black background when the stroke is sufficiently small. The bug occurs with any color of fill or stroke, and will disappear when either the stroke-width or the path itself is increased.
Comment 1 Pieter Senster 2012-01-26 02:50:35 PST
Created attachment 124087 [details]
Screenshot of actual results (screenshot scaled)
Comment 2 komoroske 2012-01-27 10:57:24 PST
I just opened the test case and saw the artifacts on 18.0.1017.2 dev on Mac.  However, now no matter how many times I open it I can't reproduce.  Odd.

Pieter, I don't have a Linux box handy to test on. Did this ever work correctly on Linux? If so, can you run bisect-builds.py (http://www.chromium.org/developers/bisect-builds-py) and post the results here?  That will help identify when the regression happened and who might be a good owner.
Comment 3 Pieter Senster 2012-01-30 02:33:36 PST
Reproduced the bug on Chromium 2.0.180.0 (Developer Build 15734), so it seems this never worked.

I've noticed the same glimpse of artifacts on Chrome Mac as you describe, but no way to reproduce either.