Closed Bug 898129 Opened 6 years ago Closed 6 years ago

Assertion "Bad uniform location." in LayerManagerOGLProgram.cpp

Categories

(Core :: Graphics: Layers, defect)

x86_64
Linux
defect
Not set

Tracking

()

RESOLVED FIXED
mozilla25

People

(Reporter: billm, Assigned: mattwoodrow)

Details

Attachments

(2 files)

I get an assertion here:
  http://mxr.mozilla.org/mozilla-central/source/gfx/layers/opengl/LayerManagerOGLProgram.cpp#310

for the "uTextureTransform" location. According to Matt,

<mattwoodrow> so uTextureTransform is used in the vertex shader to compute vTexCoord
<mattwoodrow> and the normal fragment shaders read from vTexCoord
<mattwoodrow> *but* the ColorLayer frag shader ignores vTexCoord (since there's no texture to read from, so no need for texture coords)
<mattwoodrow> so your GL implementation is optimising the whole thing out
Bill, does this patch make a difference?
Attachment #783295 - Flags: feedback?(matt.woodrow)
Flags: needinfo?(wmccloskey)
Attachment #783295 - Attachment is patch: true
Attachment #783295 - Attachment mime type: text/x-pascal → text/plain
Attached patch bug898129Splinter Review
I think we need to get rid of the callers to SetTextureTransform when we have a ColorLayer program too.
Attachment #783303 - Flags: review?(ncameron)
Attachment #783295 - Flags: feedback?(matt.woodrow) → feedback-
I'm assuming you don't want me to try this out anymore.
Flags: needinfo?(wmccloskey)
Comment on attachment 783303 [details] [diff] [review]
bug898129

Review of attachment 783303 [details] [diff] [review]:
-----------------------------------------------------------------

::: gfx/layers/opengl/CompositorOGL.cpp
@@ +1158,5 @@
>  
>        program->Activate();
>        program->SetTextureUnit(0);
>        program->SetLayerOpacity(aOpacity);
> +      program->SetTextureTransform(gfx3DMatrix());

I think you need similar additions for the other two effect types.
Attachment #783303 - Flags: review?(ncameron) → review+
Oops, I didn't realize there was a new patch. It fixes the problem for me.
I don't which of bug 893824, bug 900133 or bug 898129 caused it, but inbound has been closed for long enough.
https://hg.mozilla.org/integration/mozilla-inbound/rev/5e1009e4b1e1
Assignee: nobody → mh+mozilla
Assignee: mh+mozilla → matt.woodrow
https://hg.mozilla.org/mozilla-central/rev/8808316e03bb
Status: NEW → RESOLVED
Closed: 6 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla25
You need to log in before you can comment on or make changes to this bug.