1081497 - Rendering error on http://madebyevan.com/webgl-water/

Status: RESOLVED DUPLICATE of bug 1125445

(Core :: Graphics: CanvasWebGL, defect)

Description

[Alice0775 White]

Attachment: screenshot

[Tracking Requested - why for this release]:

Reported on http://forums.mozillazine.org/viewtopic.php?p=13814123#p13814123


STR
1. Open http://madebyevan.com/webgl-water/

Actual Results:
Nightly35.0a1 doesn't render it correctly.
See attached screenshot.

Expected Results:
Render like Aurora34.0a2 and Chrome38

Comment 1

[Alice0775 White]

In local build 
Last Good: 45472bfa3f16
First Bad: e17f9996e092

Regressed by:
  e17f9996e092	Jeff Gilbert — Bug 1048108 - Workaround ANGLE not having no-alpha surfaces. - r=kamidphish

Comment 2

[Lukas Blakk [:lsblakk] use ?needinfo]

Jeff - can you take a look at this and see if you can do a backout or forward fix for this regression?

Comment 3

[Alice0775 White]

Please ignore comment #1

Comment 4

[Alice0775 White]

Oops
Sorry, please ignore Comment #3.

Comment #1 is correct.

Comment 5

[Lawrence Mandel [:lmandel] (use needinfo)]

Jeff - We're almost out of time on beta. Do you still intend to work on this? If not, bug 1048108 is going to ship in 34.

Comment 6

[Kelsey Gilbert [:jgilbert]]

(In reply to Lawrence Mandel [:lmandel] (use needinfo) from comment #5)
> Jeff - We're almost out of time on beta. Do you still intend to work on
> this? If not, bug 1048108 is going to ship in 34.

I'll see what I can do.

Comment 7

[Lawrence Mandel [:lmandel] (use needinfo)]

Jeff - I need to ping you on this one as we're going to build with 34 beta9 (the last scheduled beta) Thursday. Any progress on this bug?

Comment 8

[Kelsey Gilbert [:jgilbert]]

(In reply to Lawrence Mandel [:lmandel] (use needinfo) from comment #7)
> Jeff - I need to ping you on this one as we're going to build with 34 beta9
> (the last scheduled beta) Thursday. Any progress on this bug?

Nothing new yet. Still trying to figure out what it is.

Comment 9

[Kelsey Gilbert [:jgilbert]]

Attachment: gl command dump

Comment 10

[Lawrence Mandel [:lmandel] (use needinfo)]

Beta10 goes to build today and our final beta will now be Beta11, which will go to build on Thursday. Depending on whether Jeff is able to find a fix and the risk of accepting the fix, we may or may not be able to fix this bug in 34.

Comment 11

[Kelsey Gilbert [:jgilbert]]

(In reply to Lawrence Mandel [:lmandel] (use needinfo) from comment #10)
> Beta10 goes to build today and our final beta will now be Beta11, which will
> go to build on Thursday. Depending on whether Jeff is able to find a fix and
> the risk of accepting the fix, we may or may not be able to fix this bug in
> 34.

Thanks for the update. I think I'm most of the way there.

Comment 12

[Lawrence Mandel [:lmandel] (use needinfo)]

We're going to build with the final beta11 in ~1.5 hours. I take it this fix isn't ready and that we'll be shipping this issue in 34. I don't think this is stop ship so we'll need to live with this for at least one cycle.

Comment 13

[Kelsey Gilbert [:jgilbert]]

(In reply to Lawrence Mandel [:lmandel] (use needinfo) from comment #12)
> We're going to build with the final beta11 in ~1.5 hours. I take it this fix
> isn't ready and that we'll be shipping this issue in 34. I don't think this
> is stop ship so we'll need to live with this for at least one cycle.

Alright, just go without it.

Comment 14

[Lukas Blakk [:lsblakk] use ?needinfo]

Hey Jeff - anything new that could be considered for FF35?  We have one more beta on Mon Dec 29th.

Comment 15

[Kelsey Gilbert [:jgilbert]]

(In reply to Lukas Blakk [:lsblakk] use ?needinfo from comment #14)
> Hey Jeff - anything new that could be considered for FF35?  We have one more
> beta on Mon Dec 29th.

No.

Comment 16

[Lukas Blakk [:lsblakk] use ?needinfo]

Wontfixing for 35 in that case, and also removing tracking as this - like bug 1089140 - doesn't appear to have large impact and we can take followup fixes on a per-case basis for uplift as they are available.

Comment 17

[Milan Sreckovic [:milan] (needinfo for best results)]

This seems to have gotten worse - see bug 1118284.

Comment 18

[Alice0775 White]

It works if alpha contextAttributes has been explicitly defined when initialize webgl context.

"lightgl.js"

   create: function(options) {
     options = options || {};
     var canvas = document.createElement('canvas');
     canvas.width = 800;
     canvas.height = 600;
+    try {
+      options.alpha = true;
+    }  catch (e) {}
     if (!('alpha' in options)) options.alpha = false;
     try { gl = canvas.getContext('webgl', options); } catch (e) {}
     try { gl = gl || canvas.getContext('experimental-webgl', options); } catch (e) {}
     if (!gl) throw new Error('WebGL not supported');

Comment 19

[Milan Sreckovic [:milan] (needinfo for best results)]

Does setting a preference webgl.default-no-alpha to true change the behaviour here?

Comment 20

[Loic]

(In reply to Milan Sreckovic [:milan] from comment #19)
> Does setting a preference webgl.default-no-alpha to true change the
> behaviour here?

No (tested with 35 & 38).

Comment 21

[Loic]

I tested with the latest Nightly, it's fixed by bug 1125445, let's close it as dupe.