jdenny-ornl wrote: > > So we = clang-linker-wrapper? Yes, we could. But I figured this problem > > might prove not to be a one-off, and -foutput-file-base is a more general > > solution that could be used elsewhere. > > I'm just hesitant to add a new flag that's used just for this (Though I've > done similar hacks in the past.) I'll add some more clang people to get their > opinion.
Thanks. Looking through clang/lib/Driver/ToolChains/CommonArgs.cpp, I see other places where the main output file appears to be used to compute new file names (OPT_gsplit_dwarf, getStatsFileName). I'm not familiar with those features, and I haven't investigated them specifically yet to prove the same problem arises. Maybe someone else already knows. https://github.com/llvm/llvm-project/pull/149003 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
