kbobyrev added inline comments.
================
Comment at: clang-tools-extra/clangd/Headers.cpp:205
CurrentLevel.push_back(Child);
- const auto &Name = RealPathNames[Child];
// Can't include files if we don't have their real path.
+ if (!RealPathNames[static_cast<unsigned>(Child)].empty())
----------------
sammccall wrote:
> kbobyrev wrote:
> > sammccall wrote:
> > > This is no longer true, we don't need the check.
> > Wait, why not? We still have unresolved includes, e.g. preamble patches are
> > like that, their `RealPathName`s stay empty.
> Right, but why do we need to filter them out here?
> Can't we just drop them when we use them to seed the proximity sources?
I feel like producing them in the first place is kind of redundant, what signal
does it give to the user through such API?
Getting some "hidden"/"unresolved" includes certainly feels rather confusing to
me, post-filtering outside probably requires understanding of the internals
which does not look like a good idea. WDYT?
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D110386/new/
https://reviews.llvm.org/D110386
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
