On 16 September 2011 10:59, Paul Berry <span dir="ltr"><<a href="mailto:stereotype441@gmail.com">stereotype441@gmail.com</a>></span> wrote:<br><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div><div></div><div class="h5">On 16 September 2011 09:01, Marek Olšák <span dir="ltr"><<a href="mailto:maraeo@gmail.com" target="_blank">maraeo@gmail.com</a>></span> wrote:<br></div></div><div class="gmail_quote">
<div><div></div><div class="h5"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi Paul,<br>
<br>
this commit breaks the piglit test<br>
shaders/glsl-arb-fragment-coord-conventions on softpipe and r600g:<br>
<br>
bin/glsl-arb-fragment-coord-conventions -auto<br>
Regular gl_FragCoord<br>
Pixel center half integer<br>
Pixel center integer<br>
Probe at (0,0)<br>
Expected: 0.250000 0.250000 0.000000<br>
Observed: 0.749020 0.749020 0.000000<br>
Probe at (99,99)<br>
Expected: 0.250000 0.250000 0.000000<br>
Observed: 0.749020 0.749020 0.000000<br>
Pixel origin upper left<br>
Probe at (0,0)<br>
Expected: 0.000000 1.000000 0.000000<br>
Observed: 0.003922 0.003922 0.000000<br>
Probe at (99,99)<br>
Expected: 1.000000 0.000000 0.000000<br>
Observed: 0.996078 0.996078 0.000000<br>
Pixel origin upper left and pixel center integer<br>
PIGLIT: {'result': 'fail' }<br>
<br>
Reverting the commit fixes this. It looks like<br>
layout(pixel_center_integer) is ignored, because "Pixel center half<br>
integer" and "Pixel center integer" return the same color in the test.<br>
<br>
Do you have any idea what's wrong?<br>
<br>
Marek<br>
<div><div></div><br></div></blockquote></div></div><div><br>Wow, that's surprising, since the commit should only affect what is considered an error condition; it shouldn't affect the GLSL IR that gets generated.<br>
<br>I'll do some investigation and let you know what I find.<br><font color="#888888">
<br>Paul <br></font></div></div>
</blockquote></div><br>Marek,<br><br>I've reproduced the behavior you saw, and I believe it is due to a build dependency problem of some sort, unrelated to the array_lvalue fix.<br><br>If I check out commit 373234ce17555376c5cb57a2cdce3038df3f51cd (the commit before my array_lvalue fix) and do a clean build of Gallium, glsl-arb-fragment-coord-conventions passes.<br>
If I then check out commit 00792e3586746c833ffc9bb65712e38038916e06 (my array_lvalue fix) and do an *incremental* build, glsl-arb-fragment-coord-conventions fails.<br>If I then do a clean build of Gallium, glsl-arb-fragment-coord-conventions passes again.<br>
<br>Unfortunately, I'm not familiar enough with the Gallium build system to track this bug further. But I'm confident that it's unrelated to this patch since a clean build makes it go away.<br><br>Hope that helps,<br>
<br>Paul<br>