aaron.ballman added a comment. Multichar literals are implementation-defined in C and conditionally supported with implementation-defined semantics in C++. I agree that it may make sense to warn about their use for portability reasons, but I'm not certain whether it makes sense to promote their use to be always-on diagnostics. I'm curious to know if this change causes any issues with system headers (which may or may not still define four char codes) or popular libraries.
I was curious as to why this was an extension in the first place and found the original commits (https://github.com/llvm/llvm-project/commit/74c95e20af4838152a63010292d1063835176711 and https://github.com/llvm/llvm-project/commit/8577f62622d50183c7413d7507ec783d3c1486fc) but there's no justification as to why this was picked as an extension. ================ Comment at: clang/include/clang/Basic/DiagnosticLexKinds.td:109 "multi-character character constant">, InGroup<MultiChar>; -def ext_four_char_character_literal : Extension< +def ext_four_char_character_literal : Warning< "multi-character character constant">, InGroup<FourByteMultiChar>; ---------------- One potential reason why we don't want to warn on this by default is that four char codes were quite popular back in the Mac Classic days. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D87962/new/ https://reviews.llvm.org/D87962 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
