gamesh411 wrote: I have considered the following approach first: https://github.com/llvm/llvm-project/commit/23514c5262e28f3cae352a0cfbe5c7de6b186a9e That one increases test coverage at the cost of introducing some duplication in diagnostics expectations, as we have discussed in #186802. The current PR makes the tests simpler (technically reducing coverage, but IMO, in practice, coverage stays the same: we only lose the non-header-guarded code's coverage from run-lines that *do* use those preprocessor symbols). So there is a tradeoff here. I am not heavily invested in either solution. @steakhal, if you think that direction is better than this, I'll just update this PR to go that way.
https://github.com/llvm/llvm-project/pull/188709 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
