sammccall added inline comments.
================ Comment at: clang-tools-extra/clangd/SemanticHighlighting.cpp:608 for (const auto &Line : Tokens) { - llvm::SmallVector<char, 128> LineByteTokens; + llvm::SmallVector<char> LineByteTokens; llvm::raw_svector_ostream OS(LineByteTokens); ---------------- njames93 wrote: > Looks like this is referring to how many bytes are in a line, having 128 > seems like a good amount, most coding standards don't like lines longer than > that. As a follow up refractor, this could be extracted out the loop to reuse > the buffer on the case it does need to allocate. Nit; This is rather the encoded data for all the highlighted tokens on a line of source code. Which appears to be a constant 8 bytes for token. So previously 16 tokens, now 32 tokens. Either seems fine to me given we don't highlight most punctuation I think. (Incidentally, this is the old pre-standard semantic highlighting protocol that will go away soon, so I wouldn't worry too much about its performance) Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D92788/new/ https://reviews.llvm.org/D92788 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits