michaelplatings marked an inline comment as done. michaelplatings added a comment.
Thanks @simon_tatham. I agree there's definitely scope to improve this. The behaviour is quite specific so I kept the code in the BareMetal toolchain rather than as part of the core Multilib class. Other toolchains may require different behaviour as you say. This change also doesn't address the layered multilib scenario. In that scenario multiple multilibs may be required to match to get a complete set of headers and libraries. But this change moves us in the right direction. ================ Comment at: clang/test/Driver/baremetal-multilib.yaml:88-90 +- Dir: arm-none-eabi/arm/v4t + Flags: [--target=armv4t-none-unknown-eabi] + ---------------- simon_tatham wrote: > I'm not sure what the purpose of this addition is. It doesn't seem to be > listed in the expected error messages (but it's presumably being printed > anyway, and skipped over because the checks don't say `-NEXT`?). Did it sneak > in here from another commit, perhaps? Sorry I forgot to explain this already. With the new error, CHECK-NO-EXECUTE-ONLY-ASM started failing because no multilib variant matched. I added this variant to fix that. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D153292/new/ https://reviews.llvm.org/D153292 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits