IRC Logs of #dri-devel on for 2015-03-01

Previous dayChoose dateNext day Show menu

09:58 #dri-devel: < DanteLA> Hi. According to this: sandy bridge HD3000 supports OpenGL 3.3 / GLSL 3.3, but it doesn't seem to be supported with latest git mesa unless 'env MESA_GL_VERSION_OVERRIDE=3.3' and then most real-world apps fail, while some simple 3.3 tests do work. Should this work?
10:04 #dri-devel: < imirkin> DanteLA: it only supports OpenGL 3.3 Core profile, not the compat profile. Only up to 3.0 for compat (along with every other mesa driver)
10:04 #dri-devel: < imirkin> DanteLA: for full info on what hardware supports what, take a look at
10:05 #dri-devel: < DanteLA> imirkin: Hmm. But OpenGL 3.3 app requesting 3.3 core context should work?
10:05 #dri-devel: < imirkin> that's right
10:06 #dri-devel: < imirkin> if it doesn't, that's a bug in mesa
10:07 #dri-devel: < imirkin> if you're having to use MESA_GL_VERSION_OVERRIDE, that indicates that the app is requesting a compat context though
10:07 #dri-devel: < DanteLA> imirkin: I guess I should try with the latest stable. GLSL shaders failed to compile unless 'MESA_GL_VERSION_OVERRIDE=3.3' was set with (git 1.6.0)
10:07 #dri-devel: < DanteLA> imirkin: That was with app using core contedxt
10:08 #dri-devel: < imirkin> can you pastebin the output of 'glxinfo'?
10:08 #dri-devel: < imirkin> perhaps you forgot to build with --enable-texture-float
10:08 #dri-devel: < imirkin> although.... that wouldn't matter for i965
10:54 #dri-devel: < DanteLA> imirkin: you were right, core context was not properly requested. OpenGL 3.3 with HD3000 works now.
13:18 #dri-devel: < imirkin> if i do glTexImage2D(GL_RGB, GL_UNSIGNED_BYTE), each pixel will have a stride of 3 bytes, right? (not 4?) [and there's the UNPACK_ALIGNMENT thing for the next row which i'm also aware of]
13:22 #dri-devel: < glennk> 3 bytes/pixel for the external data
13:23 #dri-devel: < imirkin> grrr. i'm having the hardest time loading some data into a texture :(
13:25 #dri-devel: < mareko> there's also stride alignment
13:27 #dri-devel: < imirkin> mareko: can you elaborate? this is in gles2 btw
13:30 #dri-devel: < mareko> hm, seems like unpack_alignment is the stride
13:30 #dri-devel: < mareko> nevermind
13:31 #dri-devel: < imirkin> not really the stride, but kinda-sorta
13:31 #dri-devel: < imirkin> it can only be 1/2/4/8

Written by Christoph Brill © 2007-2015

Valid XHTML 1.0 Transitional

Available in Android Market