MaskRay added inline comments.
================ Comment at: clangd/FindSymbols.cpp:31 + + if (supportedSymbolKinds && + std::find(supportedSymbolKinds->begin(), supportedSymbolKinds->end(), ---------------- MaskRay wrote: > malaperle wrote: > > MaskRay wrote: > > > This std::find loop adds some overhead to each candidate... In my > > > experience the client usually doesn't care about the returned symbol > > > kinds, they are used to give a category name. You can always patch the > > > upstream to add missing categories. > > > > > > This is one instance where LSP is over specified. nvm I don't have strong > > > opinion here > > I have a client that throws an exception when the symbolkind is not known > > and the whole request fails, so I think it's worth checking. But if it's > > too slow I can look at making it faster. Unfortunately, I cannot patch any > > upstream project :) > https://github.com/gluon-lang/languageserver-types/blob/master/src/lib.rs#L2016 > > LanguageClient-neovim returns empty candidate list if one of the candidates > has unknown SymbolKind. Apparently they should be more tolerant and there is > an issue tracking it. If it was not an internal confidential client, I would like to know its name, unless the confidentiality includes the existence of the client. Repository: rCTE Clang Tools Extra https://reviews.llvm.org/D44882 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits