Skip to content

v2026.10

Choose a tag to compare

@github-actions github-actions released this 29 May 03:13
· 155 commits to master since this release
679800a
=== Breaking changes ===
f8e65c2d3 Add function and branch shader coverage (#11214)
38aaf2e7d Use variadic layer sizes for neural shared memory (#11255)
1d3d07772 Extend shader coverage source metadata (#11177)

=== All changes for this release ===
b83f133a1 Bump DXC to v1.10.2605.24 (#11348)
bd5e45064 Fix printf in LLVM IR emitter when it is passed structs that are not from MakeStruct (#11159)
095a18a58 Fix and test infinite recursion in LLVM debug data emission (#11108)
29cd3866f Fix command line reference regeneration setup (#11342)
c5ede96fb Fix matrix swizzle out argument lowering (#11275)
79ac457f0 Consolidate generic constraint solving in a monolithic worklist (#11210)
b6ca5682d Emit macro expansion stack in diagnostics (#6165) (#11116)
e4c7947a2 Use VS 2026 for CMake options workflow (#11321)
67d4692d8 Fix inherited associated type hoisting (#11325)
be7ab29f2 Update generated design docs (#11322)
[BREAKING] f8e65c2d3 Add function and branch shader coverage (#11214)
2e84bc029 ir-layout: don't deref non-literal vector element count; skip unresolved reinterprets (#11290)
4ae6eee33 Language reference: literal expressions (#11217)
9cc1ac7cb reflection: unwrap ModifiedType at API boundary (#11280)
39cebbf01 Add LLM-generated Slang architectural design documentation (#11207)
888fb5daf Enable KHR barycentric features during Vulkan device init (#11298)
74459f665 Remove deprecated SPIRV-Tools optimizer passes (#11312)
d50a8340f Add Metal neural cooperative matrix support (#11099)
a02423373 Validate VM bytecode operand bounds (#11309)
c01b0b63f ci: skip Claude PR review for draft PRs (#11295)
13701f1ff Suppress Clang -Wnontrivial-memcall warnings from vendored imgui (#10858)
37f0f9d15 Support stdin input in slangc (#11305)
bd95ee134 Open existential through ModifiedType in member lookup (#11273)
43ec1a308 Tighten CI Git trust and sccache setup (#11302)
28aecc53e Validate repro offsets before loading to prevent out-of-bounds reads (#11250)
9e0acfecc Limit replay stream array decoding (#11256)
cdf0bd1a0 Add PR review mode to worktree helper (#11301)
65d2b9e3a Implement SP #039: Add `__func_extension` for flexible custom derivatives (#10827)
[BREAKING] 38aaf2e7d Use variadic layer sizes for neural shared memory (#11255)
944da8249 Validate replay blob hashes (#11259)
e6f99696b Guard OffsetContainer allocation against overflow and 32-bit offset truncation (#11257)
15291555f Fix spurious W40004 on non-decimal integer literals (#11216 Case A) (#11220)
fe5ce5ef6 Fix #11278: out-of-bounds read in getStringLiteralTokenValue (#11281)
73697735e Limit zip uncompressed file size to mitigate decompression bombs (#11258)
f9ec8ac29 Fix/git worktree add 2 (#11288)
ac9c20bea Guard serialized IR operand indices against out-of-bounds access (#11251)
a5cc7ebdf Fix buffer overflow and dead code in RTTI list copyArray (#11239)
3b9d0d627 ci: bump linux-gpu-ci container to v1.6.1 (#11274)
98067e481 Fix submodule sync error (#11283)
18e72e1b6 Fix spurious GL_EXT_texture_shadow_lod requirement for baseline shadow LOD types (#11156)
9c1b6be74 Reject oversized replay strings (#11260)
803a4180b Validate VM bytecode offsets (#11261)
627c6a817 Diagnose non-float matrix element types in SPIR-V entry-point varyings (#11150)
a99af6e7f Initialize worktree submodules concurrently (#11269)
9589cd407 Add tmux option to worktree helper (#11268)
e35f2870c Add git worktree helper (#11264)
fd6aa89a2 Use slang-build skill in agent guidance (#11263)
acfb1379f Verify Vulkan SDK checksum in GPU CI image (#11238)
2566c5da5 ci: harden slash command and format workflows against untrusted input (#11241)
184d9aa41 Require fp8 render feature for CUDA tests (#11222)
9b044ad46 Fix SPIR-V void pointer array stride (#11110)
a2ad34d7a Enable optional WebGPU development on Linux (#11206)
bf714d55b Emit default ray payload access qualifiers for HLSL SM 6.7+ (#11224)
569377024 Language reference: expressions reorganization, main page, operator precedence table (#11194)
4c4e8d388 ci/rtx-remix-nightly: drop ngx_sdk_dlfg from packman pull (#11229)
47ef298bc chore: ignore .claude/scheduled_tasks.lock (#11091)
38411f60f Add Claude local settings and task files to .gitignore
36deb481e ci: bake gdb into linux-gpu-ci container (v1.6.0) (#11219)
f4bf8c8ae Enable WGSL ByteAddressBuffer.Load and related buffer access support (#11167)
1bb7cf9f0 Add GL_KHR_shader_subgroup_basic to the SPIR-V branch of sm_6_0 (#11145)
[BREAKING] 1d3d07772 Extend shader coverage source metadata (#11177)
423b83fd8 ci/coverage-nightly: add merged + new-renderer reports (#10973)
c6d49b8ab Fix slang-test launching problem on WSL (#11212)
6a5ae04d0 linalg: fix implementation of cooperative vector VectorAccumulate (#11213)
704de06f5 Fix WGSL emitter: emit u64 instead of uint64_t for UInt64Type (#11164)
7dd70f15b Add WaveGetWaveIndex builtin (#11192)
e89ff95f4 ci_health: detach Queued series from hosted-runner in-use stack (#11205)