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

Issue 659790 link

Starred by 3 users

Issue metadata

Status: Fixed
Owner: ----
Closed: Dec 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux , Windows , Mac
Pri: 2
Type: Bug-Regression



Sign in to add a comment

Transparent elements with clip-path having an additional clip-path render as black

Reported by marcelja...@gmail.com, Oct 26 2016

Issue description

UserAgent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_11_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/54.0.2840.71 Safari/537.36

Steps to reproduce the problem:
1. Have an SVG element that references a clip-path that references another clip-path (test case: the top left quadrant)
2. Set the color to transparent or fill to none

What is the expected behavior?
Element should be invisible

What went wrong?
Element is drawn black

Did this work before? Yes 53

Does this work in other browsers? Yes

Chrome version: 54.0.2840.71  Channel: stable
OS Version: OS X 10.11.6
Flash Version: Shockwave Flash 23.0 r0

- Setting fill to `rgba(0,0,0,0.01)` works (0.001 does not).
- `transparent` and `fill` don't work.
- Making the clip-path not reference another clip-path prevents the issue (lower column in test case)
 
svg-bug-with-experiments.xml
946 bytes View Download
chrome54.png
4.1 KB View Download
safari-ff-chrome53.png
4.1 KB View Download
This is a quick JS fiddle to test this: https://jsfiddle.net/37qd0s7t/2/

Comment 2 by pdr@chromium.org, Oct 27 2016

Labels: -Type-Bug Needs-Bisect Type-Bug-Regression
Status: Available (was: Unconfirmed)
Thanks for taking the time to file this with a nice and small testcase.

@test team, please test with https://jsfiddle.net/37qd0s7t/2/. This test is correct if it matches firefox/safari (aka there is no black).

Comment 3 by hdodda@chromium.org, Oct 27 2016

Cc: hdodda@chromium.org
Labels: -Needs-Bisect hasbisect-per-revision OS-Linux OS-Windows
Owner: wkorman@chromium.org
Status: Assigned (was: Available)
Using the per-revision bisect providing the bisect results,
Good build:54.0.2827.0 (Revision: 411497).
Bad build: 54.0.2828.0  (Revision:411828).

You are probably looking for a change made after 411559 (known good), but no later than 411560 (first known bad).

CHANGELOG URL:
 
The script might not always return single CL as suspect as some perf builds might get missing due to failure.
  https://chromium.googlesource.com/chromium/src/+log/3831b72130f11ebac151fd4945719f6d7489e048..971a9c9725e293bd89b7cb1475acdc502065e6b3

From the CL above, assigning the issue to the concern owner 

@wkorman - Could you please check whether this is caused with respect to your change, if not please help us in assigning it to the right owner.

Review-Url: https://codereview.chromium.org/2225563002

Note : Issue is seen in Windows and Linux OS also.
Thanks!
Components: Blink>Paint
I can repro with jsfiddle on Linux 56.0.2906.0 dev. It is likely due to my change. I plan to look into this next week.
Cc: wkorman@chromium.org
Owner: ----
Status: Available (was: Assigned)
Have not had chance to look at this and not clear I will get to it soon. Unassigning in case someone else on Paint team can pick up before I can.
Having the same problem here. @wkorman Any update?
Symptom seems to have changed now, too: the bottom half is black as well.
black-and-blue.png
4.1 KB View Download
Cc: f...@opera.com pdr@chromium.org
Sorry, I have not had a chance to look at this. I am on other things currently and this is SVG specific where I have only limited background. +pdr and fs in case they have insight or can help route.

Comment 9 by f...@opera.com, Aug 22 2017

The intersection of the CL bisected to and the testcase would seem to indicate the visual rect derived from the <clipPath> with a clipped child element. Presumably the indicated CL only exposed previously "bad" data by somehow starting to rely on it (being correct.) (There was similar issues before, like  issue 598051 , which may not have ironed out this particular kink.)

Comment 10 by f...@opera.com, Dec 1 2017

Cc: trchen@chromium.org
This appears to have been fixed/papered over by fc2679dfa73c1dc102d86832ef049359f433c314.
Status: Fixed (was: Available)
Yes that CL fixed the exact root cause of it.

Sign in to add a comment