Created attachment 75444 [details]
Setting glEdgeFlag to GL_FALSE, then checking glGetBooleanv of GL_EDGE_FLAG returns GL_TRUE.
This patch should fix it. Posted for review.
diff --git a/src/mesa/main/get_hash_params.py b/src/mesa/main/get_hash_params.py
index 1b7d6ad..ff6e108 100644
@@ -412,7 +412,7 @@ descriptor=[
[ "DEPTH_SCALE", "CONTEXT_FLOAT(Pixel.DepthScale), NO_EXTRA" ],
[ "DOUBLEBUFFER", "BUFFER_INT(Visual.doubleBufferMode), NO_EXTRA" ],
[ "DRAW_BUFFER", "BUFFER_ENUM(ColorDrawBuffer), NO_EXTRA" ],
- [ "EDGE_FLAG", "LOC_CUSTOM, TYPE_BOOLEAN, 0, NO_EXTRA" ],
+ [ "EDGE_FLAG", "LOC_CUSTOM, TYPE_BOOLEAN, 0, extra_flush_current" ],
[ "FEEDBACK_BUFFER_SIZE", "CONTEXT_INT(Feedback.BufferSize), NO_EXTRA" ],
[ "FEEDBACK_BUFFER_TYPE", "CONTEXT_ENUM(Feedback.Type), NO_EXTRA" ],
[ "FOG_INDEX", "CONTEXT_FLOAT(Fog.Index), NO_EXTRA" ],
The proposed patch does not work for me, it still returns true.
Hmm, I retested here and it works. Are you sure everything was rebuilt properly? The get_hash.h file gets regenerated from the get_hash_params.py file (which is what we're patching). Could you try rm'ing get_hash.h and rebuilding?
Hmm, I probably rebuilt it incorrectly, I used ArchLinux's PKGBUILD file to auto build it into a package (and apply the patch beforehand), the build system probably splits these packages differently (I ended up with mesa as well as several driver packages). I'll try it again, and if it doesn't work, I'll manually build it.
Ah yes, apparently it was in the "ati-dri" package that arch built. Works properly now, thanks! By the way, I have some more unresolved bugs (Regarding OpenGL 1.x-2.1) on a opengl ruby binding library (https://github.com/archSeer/opengl), so be on a lookout, I'll report probably 2-3 more in the next few days.
Brian, has the patch been merged yet?
Thanks for the reminder. I just pushed it.