MaskRay added a comment. > -gsplit-dwarf -gmlt -fno-split-dwarf-inlining => 1 > > This last one currently produces split-dwarf (if there's any DWARF worth > splitting - if there are any subprogram descriptions, etc, otherwise it saves > the indirection and produces an empty .dwo file).
Thanks for catching this. It should be 1 + split instead of 1. I've updated the description. This patch doesn't change the behavior of this rule. > Seems there's going to be confusion either way, though - either the > presence/absence of -fno-split-dwarf-inlining changes whether -gsplit-dwarf > is respected/ignored (in the presence of -gmlt), or changes whethe -gmlt > composes with -gsplit-dwarf or overrides it? Seems these are both a bit > confusing, no? See the updated description. The lower half of the rules can be summaries as `-g* -gsplit-dwarf => 2 + split` with this patch. I think that is how this patch makes the rules less confusing. Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D59923/new/ https://reviews.llvm.org/D59923 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
