[PATCH] D116085: [clang-tidy] Performance improvements for NOLINTBEGIN/END blocks

[Salman Javed via Phabricator via cfe-commits]

salman-javed-nz added a comment.

Friendly ping.

Would be good to get these performance improvements into trunk soon, so that we're not prolonging the time that people are putting up with the current slow implementation. Also, I believe that LLVM 14.0.0 will be up for a release candidate soon, and all going well, the NOLINTBEGIN feature would be debuting in it. It would be nice if the feature shipped in this faster incarnation. It would be a pity if users forgo using this feature because they are put off by the slowness of it.

@kadircet as the clangd expert, I'm especially interested in your thoughts on this patch, as clangd is sensitive to clang-tidy's performance. I compiled clangd and tested it with the vscode-clangd extension. Everything worked as expected when moving lines around, applying fixits, and switching tabs. Performance looked OK to me too, but I haven't used clangd all that much so I don't have a baseline on what is normal performance.

Thanks a lot.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D116085/new/

https://reviews.llvm.org/D116085