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

Issue 793307 link

Starred by 2 users
Status: Duplicate
Merged: issue 792966
Owner:
kainino@chromium.org
Closed: Dec 2017
Cc:
█ kbr@chromium.org
kainino@chromium.org
geoffl...@chromium.org
jdarpinian@chromium.org
jmad...@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: Windows
Pri: 1
Type: Bug-Regression



Sign in to add a comment

WebGL2: Uniform Buffer Objects do not work correctly

Reported by pdae...@gmail.com, Dec 8 2017 
UserAgent: Mozilla/5.0 (Windows NT 6.3; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/63.0.3239.84 Safari/537.36

Steps to reproduce the problem:
Open the file "webgl2_ubo_bug.html" attached to this bug report with version 63 on Windows. The machine must support WebGL2, and the rendering must be via ANGLE.

What is the expected behavior?
I expect to see a rectangle consisting of two triangles, one red and one green.

What went wrong?
Starting with version 63, I see a red rectangle.

This worked as expected until version 62. This also works as expected with version 63 on Mac OS, so probably it is a bug in ANGLE.

Did this work before? Yes 

Does this work in other browsers? N/A

Chrome version: 63.0.3239.84  Channel: stable
OS Version: 6.3
Flash Version:
webgl2_ubo_bug.html
2.6 KB View Download

Comment 2 by pdae...@gmail.com, Dec 8 2017 

Corresponding Firefox ticket:
https://bugzilla.mozilla.org/show_bug.cgi?id=1424258

Comment 3 by zmo@chromium.org, Dec 8 2017

Cc: kbr@chromium.org geoffl...@chromium.org jdarpinian@chromium.org kainino@chromium.org jmad...@chromium.org

Comment 4 by kainino@chromium.org, Dec 8 2017

Components: Internals>GPU>ANGLE
Owner: jmad...@chromium.org
Status: Assigned (was: Unconfirmed)
Thanks for the regression range, that's really helpful.
Here's the log for that range:
https://crrev.com/501743..501756
which includes this ANGLE roll:
https://chromium.googlesource.com/angle/angle.git/+log/9df395c..715e7f1
Almost every patch in that range is +jmadill's; assigning.

Comment 5 by kbr@chromium.org, Dec 8 2017

Blockedon: 792966
Labels: -Pri-2 Pri-1
This may be the same issue as  Issue 792966 . Linking the two together and upgrading to P1. It sounds like a lot of WebGL 2.0 content may be broken and we should consider a quick fix for M63.

Comment 6 by zmo@chromium.org, Dec 8 2017 

Thanks for the simple test case. Let me add it to WebGL's conformance test suite (I am surprised the current suite fails to catch such regression!)

Comment 7 by zmo@chromium.org, Dec 8 2017

Labels: M-64 M-63

Comment 8 by kbr@chromium.org, Dec 8 2017

Owner: kainino@chromium.org
Kai said he'd own this at least for the next couple of days.

Comment 9 by zmo@chromium.org, Dec 8 2017 

conformance test added: https://github.com/KhronosGroup/WebGL/pull/2559

Comment 10 by kainino@chromium.org, Dec 8 2017

Mergedinto: 792966
Status: Duplicate (was: Assigned)
Pretty confident these bugs are the same.

Comment 11 by kainino@chromium.org, Dec 8 2017

Blockedon: -792966
►

Sign in to add a comment