kadircet added a comment. In https://reviews.llvm.org/D51747#1227089, @ilya-biryukov wrote:
> Not sure if it's fine to hijack our own diagnostic-specific flags in to clang > command args. > > Cons that I see: > > 1. There is no way for the users to turn them off if they find them > non-useful. If we add a way, it would be more config parameters which overlap > with other mechanism that we have - compiler flags. > 2. Users who are used to having them as warnings will now see them as notes. > Again, no way to tweak this behavior. > > What's our use-case? Maybe we should ask the clients to add -Wdeprecated if > they care about those? > > PS In case I'm missing the context here, please let me know. Agree with you, I think it would be better to provide it as an option. https://reviews.llvm.org/D51724 with this one we added a way to show deprecated symbols on code completion responses and wanted to move forward with showing the ones that are already in existing code. ================ Comment at: clangd/Diagnostics.cpp:299 + D.Severity = + D.Category == "Deprecations" ? DiagnosticsEngine::Note : DiagLevel; return D; ---------------- ioeric wrote: > kadircet wrote: > > Couldn't find a better way to check for this one. Category types are coming > > from tablegen file > > https://github.com/llvm-mirror/clang/blob/master/include/clang/Basic/DiagnosticGroups.td#L128 > > which does not expose category id, at least I couldn't find even if it > > does. > Have you tried the Diagnostic ID (i.e. `Info.getID()`)? It matches the diag > kinds defined in `DiagnosticSemaKinds.inc` (e.g. `warn_deprecated`). Unfortunately these are also internal, https://github.com/llvm-mirror/clang/blob/master/lib/Basic/DiagnosticIDs.cpp#L97 Repository: rCTE Clang Tools Extra https://reviews.llvm.org/D51747 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits