Use pathlib in cuda.pathfinder._static_libs (part 2 of #2410) - #2493
Use pathlib in cuda.pathfinder._static_libs (part 2 of #2410)#2493LeSingh1 wants to merge 2 commits into
Conversation
github-actions
Bot
added
the
cuda.pathfinder
mdboom
added
enhancement
|
/ok to test 8ed361d |
|
@LeSingh1: This PR has a non-trivial merge conflict. Would you mind resolving it? |
Part 2 of the series proposed in NVIDIA#2410, following the same conversion style as part 1 (NVIDIA#2489). Path construction, joining, and filesystem predicates in find_static_lib.py and find_bitcode_lib.py now go through pathlib.Path instead of os.path string manipulation. Both modules keep importing os solely for os.environ.get("CONDA_PREFIX"). Compatibility is preserved: every entry point still accepts str, and every function that documents or returns str still returns str. Path is used strictly as the internal representation and converted back with str() at each return, so LocatedStaticLib.abs_path, LocatedBitcodeLib .abs_path, find_static_lib() and find_bitcode_lib() are unchanged in both type and value. No signature changes. Signed-off-by: LeSingh1 <sshaurya914@gmail.com>
Follow-up to the review feedback on NVIDIA#2489: the str-compatibility constraint applies only to the public API. The try_* methods and _no_such_file_in_dir now work in Path throughout. str() is applied once, where abs_path is stored on the public LocatedStaticLib and LocatedBitcodeLib. The relative-path constants go from os.path.join(...) to forward-slash literals, matching how site_packages_dirs is already written in the same dicts; Path normalizes the separator on Windows. One behavior change: a CUDA_PATH or CONDA_PREFIX containing redundant separators ("//", "/.") now produces a normalized abs_path, because Path collapses them. Differential fuzzing against the pre-revision code (16k lookups over randomized trees, comparing located paths and full error text) shows no other difference, and none at all when those variables are free of redundant separators. Signed-off-by: LeSingh1 <sshaurya914@gmail.com>
Head branch was pushed to by a user without write access
8ed361d to
11916d6
Compare
|
Rebased onto current The conflict was "ctk_rel_paths": (str(Path("lib", arch_dir)),),
"conda_rel_paths": (str(Path("lib", arch_dir)), *conda_fallback_dirs),so the arch-aware behaviour from One thing I should flag rather than let you assume otherwise: I could not run the test suite locally to confirm. Happy to adjust if you would rather the helper stay on |
2 participants
Follow-up to the review on #2489: the str-compatibility constraint applies only to the public API, so the finder internals now pass and return
Path.str()is applied once, whereabs_pathis stored onLocatedStaticLib/LocatedBitcodeLib.The rel-path constants are now forward-slash literals, matching
site_packages_dirsin the same dicts.One behaviour change: a
CUDA_PATHorCONDA_PREFIXcontaining redundant separators now gives a normalizedabs_path, becausePathcollapses them.Verified on Linux CI: full pytest output byte-identical to the base commit apart from elapsed time (1276 passed, 4 skipped, 0 errors). Differential fuzzing against the previous revision — 16k lookups over randomized trees, comparing located paths and error text — shows no other difference.
Part 2 of #2410.