ilya-biryukov added a comment. Oh, sorry, I forgot to submit the comments yesterday :-(
================ Comment at: clangd/CodeComplete.cpp:245 + // Methods are simply grouped by name. + return hash_combine('M', IndexResult->Name); + case index::SymbolKind::Function: ---------------- Members never come from the index for completion, right? Maybe add an assert here instead? ================ Comment at: clangd/CodeComplete.cpp:265 + : StringRef(D->getDeclName().getAsString())); + return hash_combine('F', StringRef(D->getQualifiedNameAsString()), + headerToInsertIfNotPresent().getValueOr("")); ---------------- `getQualifiedNameAsString` has a fixme, suggesting it's deprecated. Maybe we should call `printName` directly instead? And we have `Scratch` for storage anyway, so it might be a little faster too. ================ Comment at: clangd/CodeComplete.cpp:366 + Opts.EnableSnippets ? (Name + "(${0})").str() : Name.str(); + First.label = (Name + "(...)").str(); + First.detail = llvm::formatv("[{0} overloads]", Bundle.size()); ---------------- Maybe use a unicode ellipsis char (…) here? Or even come up with something else. `...` is a valid syntax in C++, might be little confusing. Especially in VSCode, where detail is not shown. Repository: rCTE Clang Tools Extra https://reviews.llvm.org/D47957 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits