nathanchance added a comment. In D117717#3256293 <https://reviews.llvm.org/D117717#3256293>, @MaskRay wrote:
> Assuming scan-build is clang/tools/scan-build, which is in tree and can be > fixed instead. Yes, that is the one. > I am not familiar with it, but from > >> However, this caused issues with scan-build, as GCC is used as the compiler >> but clang is invoked after it with all the -f options that GCC was. > > if scan-build+GCC takes all GCC recognized options and forwards them to > Clang, I think it is a generally infeasible model. > GCC has dozens of options not recognized by Clang... I can barely read Perl but it seems like that is what the script does: https://github.com/llvm/llvm-project/blob/058d2123792da54ae7460fea0946d2c90a032e1c/clang/tools/scan-build/libexec/ccc-analyzer#L661-L664 I don't really know what the fix is for that on the `scan-build` side of things (how do you decide what flags should be passed through or not?). If there is someone who should be looped in, please do so. I agree that this is an infeasible model to support so I have commented on the kernel patch that brought this up: https://lore.kernel.org/r/YeiaAgQ+gbZYTMwD@archlinux-ax161/ I'll abandon this depending on what others have to say. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D117717/new/ https://reviews.llvm.org/D117717 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits