New issue
Advanced search Search tips

Issue 863037 link

Starred by 2 users

Issue metadata

Status: Fixed
Owner:
Closed: Aug 31
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Windows
Pri: 2
Type: Bug



Sign in to add a comment

Composite SVG filter parameters affect previous drawImage call

Reported by ivan.pop...@gmail.com, Jul 12

Issue description

UserAgent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/66.0.3359.181 Safari/537.36

Steps to reproduce the problem:
Open https://jsfiddle.net/yeh1u5cv/3/ or attached file

What is the expected behavior?
Left square has red shadow, right square has green shadow.

What went wrong?
Both have red shadow.

Did this work before? N/A 

Does this work in other browsers? Yes

Chrome version: 66.0.3359.181  Channel: stable
OS Version: 10.0
Flash Version: 

if you uncomment RaF, you'll see how that both shadows are green.

if you uncomment try-catch , you'll see that canvas is tainted, i dont know if its the same bug or another.

It doesnt happen with simple ColorMatrix.
It works in firefox.
 
dropshadow-params.js
3.6 KB View Download
Components: Blink>SVG
Cc: chrishtr@chromium.org f...@opera.com
Components: Blink>Canvas
Status: Available (was: Unconfirmed)
Tainting is probably part of the same bug.

I have a suspicion this is deliberate to prevent leakage of user information.
I think the tainting is expected here per the limitations of the current implementation.

Given that seems to be specific to the feFlood@flood-color in the example (mutating the stdDeviation in the push() function reflects fine), and said attribute is a presentation attribute, it'd seem likely that the <canvas> code is not ensuring style is clean before resolving the filter.
(Throwing in a statement that forces layout seems to fix rendering, increasing the likelihood of that being the issue.)
> I think the tainting is expected here per the limitations of the current implementation.

Its fine. Just tell me whether to create new long-living issue for it after feFlood@flood-color gets resolved.
No need for that, we have one already: issue 792518. But thanks for the offer.
Owner: aaronhk@chromium.org
Status: Assigned (was: Available)
Cc: fs...@chromium.org

Sign in to add a comment