[Bug 83380] Linking fails when not writing gl_Position.

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Tue Sep 2 01:06:15 PDT 2014


https://bugs.freedesktop.org/show_bug.cgi?id=83380

--- Comment #2 from Tapani Pälli <lemody at gmail.com> ---
I might have to take back comment #1 though. In the discussion section of GLSL
1.10 spec there is still mention that writing to gl_Position is mandatory.

--- 8< ---
82) Is it really necessary to require writing of gl_Position, gl_FragColor or
gl_FragDepth?

DISCUSSION: This can be difficult to handle error cases for when the writes are
conditional. There was also discussion that either kill should be called, or a
gl_ output be written in a fragment shader. However, that is irrelevant, as
it’s okay to neither kill nor write any outputs in a fragment shader. For the
vertex shader, it still makes no sense to not write a position, so this should
still be required.

RESOLUTION: For the fragment shader, there are no rules; either kill can be
called or not, and if not, nothing need be written to, existing values are
picked up from the pipeline. For the vertex shader, gl_Position should still be
written, with the compiler giving a diagnostic when possible.
--- 8< ---

If the test uses ES 3.0, why does it not use GLSL ES 3.0 in the shaders? Test
should have line "#version 300 es" to utilize correct version.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/intel-3d-bugs/attachments/20140902/2a42e740/attachment.html>


More information about the intel-3d-bugs mailing list