rnk added a comment. I see. You know, these things would be nicely handled if we just had a reliable asynch SEH implementation. >_>
What do you think of making `FrontendOptions::Inputs` be a `SmallVector<*, 0>`? It's janky and inconsistent with the rest of that struct, but it seems less invasive in the end. If we ever implement iterator invalidation checks for SmallVector, I expect we'll use some kind of epoch counting strategy that hopefully won't require locking. I assume it would be similar to what is done for DenseMap. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D69959/new/ https://reviews.llvm.org/D69959 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
