aaron.ballman added a comment. Given that we already have other setters to set global state, I'm less concerned about adding another one. I hadn't realized we already introduced the dangers here. But we should document the expectation that the call be made before creating the index.
In terms of the C API, I think it'd make more sense to name in terms of "override" rather than "set", but I don't feel as strongly about it given the other setters. In terms of the C++ file system API, I think "override" makes the most sense though (we don't have setters to follow the naming convention for) because some systems do allow you to set the system directory. In terms of memory ownership, WDYT of requiring the caller to handle this? e.g., calling `set_system_temp_directory_erased_on_reboot` will `strdup` a nonnull pointer and `free` the stored pointer when given nullptr. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D139774/new/ https://reviews.llvm.org/D139774 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
