kbobyrev added a comment. In D119130#3300780 <https://reviews.llvm.org/D119130#3300780>, @kadircet wrote:
> Regarding the include mapping generator, I think it would've been better if > we had some sort of list directly from libc++ (as this is now being part of > clang rather than just clangd), but having the current symbol mapping > available for other tools too is definitely a useful improvement and > implementation details can change later on. > I think we should have some "more public" documentation available around the > limitations of current generator though, so that people are not surprised and > aware of the caveats (like symbols might be missing, or in case of ambiguity > they might be dropped, etc). Not sure where that belongs though, maybe header > of the `.inc` file, or if we want it to be only used through the recognizer > interface, maybe we can make the `inc` file "private" and document it there. > > As for stdlib symbol/header/recognizer, I've got a couple questions around > the functionality we are exposing: > > - (briefly mentioned above) should we just make raw symbol mappings an > implementation detail of stdlib recognizer and have applications depend on it? > - do we want headers/symbols to be created outside of the recognizer? Right, I agree that using libc++ directly would be better, but for now this just makes the functionality public. We can change it and update as we want afterwards. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D119130/new/ https://reviews.llvm.org/D119130 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
