formula: anrieff/libcpuid - #519
MeteorsLiu wants to merge 2 commits into
Conversation
![fennoai[bot]](https://avatars.githubusercontent.com/in/1924301?s=60&v=4){kind=small}
There was a problem hiding this comment.
The formula structure, version thresholds, patched v0.5 build, CMake option handling, pkg-config metadata, and independent consumer tests were reviewed against the requested Conan recipe and upstream tags. One architecture-filter regression is called out inline; the shared/fPIC combinations and representative upstream CMake configurations otherwise match the inspected source behavior.
Verified the v0.7.0 shared build with fPIC=OFF directly against upstream; it succeeds.
| } | ||
| } | ||
| } | ||
| return slices.contains(target.require["arch"], "amd64") |
There was a problem hiding this comment.
[P1] Preserve 32-bit x86 support in the architecture filter
The Conan recipe's validate contract explicitly accepts both x86 and x86_64, and the upstream project publishes ix86 builds. LLAR represents 32-bit x86 as 386, so this filter rejects a supported target before onBuild runs; the same predicate is duplicated in the v0.6.0 and v0.7.0 thresholds. Include 386 alongside amd64 (or otherwise mirror the recipe's x86/x86_64 matrix) so the formula does not silently drop a supported architecture.
1 participant
Summary
anrieff/libcpuidLLAR Formula withv0.5.0,v0.6.0, andv0.7.0thresholds covering the complete upstream tag set throughv0.8.1.v0.5.1recipe, including its CMake patch, shared/fPIC options, x86-only matrix contract, license/tool installation, and relocatablelibcpuid.pcmetadata.cpuid_tool --report; no direct dependencies.Validation
v0.5.0andv0.5.1tests pass on nativedarwin/amd64under Rosetta, including static and shared builds.fPIC=OFF,v0.6.0,v0.6.5, default latestv0.8.1, and cache-hit consumer tests pass ondarwin/amd64.darwin/arm64is rejected by the Formula filter as unsupported by the pinned Conan recipe.9c062aa5c4b1724bfd30cb72f7e0bda136f9c078but could not start because this host has nold.lld.Closes #243