Mesa 10.0.4 Release Notes / (March 12, 2014)

Mesa 10.0.4 is a bug fix release which fixes bugs found since the 10.0.3 release.

Mesa 10.0.4 implements the OpenGL 3.3 API, but the version reported by glGetString(GL_VERSION) or glGetIntegerv(GL_MAJOR_VERSION) / glGetIntegerv(GL_MINOR_VERSION) depends on the particular driver being used. Some drivers don't support all the features required in OpenGL 3.3. OpenGL 3.3 is only available if requested at context creation because compatibility contexts not supported.

MD5 checksums

5a3c5b90776ec8a9fcd777c99e0607e2  MesaLib-10.0.4.tar.gz
8b148869d2620b0720c8a8d2b7eb3e38  MesaLib-10.0.4.tar.bz2
da2418d25bfbc273660af7e755fb367e  MesaLib-10.0.4.zip

New features

None

Bug fixes

This list is likely incomplete.

• Bug 71870 - Metro: Last Light rendering issues
• Bug 72895 - Missing trees in flightgear 2.12.1 with mesa 10.0.1
• Bug 74251 - Segfault in st_finalize_texture with Texture Buffer
• Bug 74723 - main/shaderapi.c:407: detach_shader: Assertion `shProg->Shaders[j]->Type == 0x8B31 || shProg->Shaders[j]->Type == 0x8B30' failed.

Changes

The full set of changes can be viewed by using the following git command:

  git log mesa-10.0.3..mesa-10.0.4

Anuj Phogat (4):

• mesa: Generate correct error code in glDrawBuffers()
• mesa: Add GL_TEXTURE_CUBE_MAP_ARRAY to legal_get_tex_level_parameter_target()
• glsl: Fix condition to generate shader link error
• i965: Fix the region's pitch condition to use blitter

Brian Paul (8):

• r200: move driContextSetFlags(ctx) call after ctx var is initialized
• radeon: move driContextSetFlags(ctx) call after ctx var is initialized
• gallium/auxiliary/indices: replace free() with FREE()
• draw: fix incorrect color of flat-shaded clipped lines
• st/mesa: avoid sw fallback for getting/decompressing textures
• mesa: update assertion in detach_shader() for geom shaders
• mesa: do depth/stencil format conversion in glGetTexImage
• softpipe: use 64-bit arithmetic in softpipe_resource_layout()

Carl Worth (4):

• docs: Add md5sums for 10.0.3 release
• main: Avoid double-free of shader Label
• get-pick-list: Update to only find patches nominated for the 10.0 branch
• Update version to 10.0.4

Chris Forbes (1):

• i965: Validate (and resolve) all the bound textures.

Christian König (1):

• radeon/uvd: fix feedback buffer handling v2

Daniel Kurtz (1):

• glsl: Add locking to builtin_builder singleton

Emil Velikov (3):

• dri/nouveau: Pass the API into _mesa_initialize_context
• nv50: correctly calculate the number of vertical blocks during transfer map
• dri/i9*5: correctly calculate the amount of system memory

Fredrik Höglund (3):

• mesa: Preserve the NewArrays state when copying a VAO
• glx: Fix the default values for GLXFBConfig attributes
• glx: Fix the GLXFBConfig attrib sort priorities

Hans (2):

• util: don't define isfinite(), isnan() for MSVC >= 1800
• mesa: don't define c99 math functions for MSVC >= 1800

Ian Romanick (6):

• meta: Release resources used by decompress_texture_image
• meta: Release resources used by _mesa_meta_DrawPixels
• meta: Fallback to software for GetTexImage of compressed GL_TEXTURE_CUBE_MAP_ARRAY
• meta: Consistenly use non-Apple VAO functions
• glcpp: Only warn for macro names containing __
• glsl: Only warn for macro names containing __

Ilia Mirkin (3):

• nv30: report 8 maximum inputs
• nouveau/video: make sure that firmware is present when checking caps
• nouveau: fix chipset checks for nv1a by using the oclass instead

Julien Cristau (1):

• glx/dri2: fix build failure on HURD

Kenneth Graunke (2):

• glsl: Don't lose precision qualifiers when encountering "centroid".
• i965: Create a hardware context before initializing state module.

Kusanagi Kouichi (1):

• targets/vdpau: Always use c++ to link

Marek Olšák (1):

• st/mesa: fix crash when a shader uses a TBO and it's not bound

Matt Turner (1):

• glsl: Initialize ubo_binding_mask flags to zero.

Paul Berry (2):

• glsl: Make condition_to_hir() callable from outside ast_iteration_statement.
• glsl: Fix continue statements in do-while loops.

Tom Stellard (1):

• r600g/compute: PIPE_CAP_COMPUTE should be false for pre-evergreen GPUs

Topi Pohjolainen (1):

• i965/blorp: do not use unnecessary hw-blending support