[Bug 96971] invariant qualifier is not valid for shader inputs

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=96971

--- Comment #5 from Qiankun Miao <qiankun.miao at intel.com> ---
(In reply to Kenneth Graunke from comment #4)
> No.  The wording isn't great, but
> 
> "Only variables output from a shader (including those that are then input to
> a subsequent shader) can be candidates for invariance."
> 
> implies to me that you can only mark outputs as invariant.  This would also
> make it match the GLSL ES 3xx rules.  GLSL ES 3.20 has similar text, but
> also says in the issues section:
> 
> "Do invariance qualifiers for declarations in the vertex and fragment
> shaders need to match?
> Option 1: Only allow invariance declarations on outputs. If a vertex shader
> output is declared as invariant, it implies that the corresponding input to
> the fragment shader is also invariant.
> Option 2: Specify that they must match.
> RESOLUTION: Only allow invariant declarations on outputs."
> 
> The resolution is pretty clear there.  You only get to use 'invariant' on
> outputs with the new rules.
> 
> Are you running into an issue with the new code I implemented?  Or just spec
> reading and wondering if we should have done something else?

I just found some differences between ESSL and GLSL. These differences should
be handled by WebGL since webgl is based on ES 3.0 API, but WebGL also should
run on top of OpenGL driver. WebGL soloved the differences, and run well on
Mesa.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20161125/1ac05ea5/attachment.html>