diff options
| author | Nicholas Nethercote <n.nethercote@gmail.com> | 2025-03-12 13:45:48 +1100 |
|---|---|---|
| committer | Nicholas Nethercote <n.nethercote@gmail.com> | 2025-03-17 08:20:56 +1100 |
| commit | 2b431f768b283319d52be7353024c8426fe8e9fe (patch) | |
| tree | 64e2bc0745ce75b6405c0b1a3b80be7a8b371179 /compiler/rustc_llvm/llvm-wrapper/CoverageMappingWrapper.cpp | |
| parent | f3579934b1a66da5b21cd26bccd578982c773c75 (diff) | |
| download | rust-2b431f768b283319d52be7353024c8426fe8e9fe.tar.gz rust-2b431f768b283319d52be7353024c8426fe8e9fe.zip | |
Fix `is_relevant_impl`.
It determines if a function should have any `inline` attributes checked. For `ItemKind::Fn` it returns true or false depending on the details of the function; for anything other item kind it returns *true*. This latter case should instead be *false*. (In the nearby and similar functions `is_relevant_impl` and `is_relevant_trait` the non-function cases return false.) The effect of this is that non-functions are no longer checked. But rustc already disallows `inline` on any non-function items. So if anything its a tiny performance win, because that was useless anyway.
Diffstat (limited to 'compiler/rustc_llvm/llvm-wrapper/CoverageMappingWrapper.cpp')
0 files changed, 0 insertions, 0 deletions
