aaron.ballman added inline comments.
================ Comment at: clang/www/hacking.html:279 + + <p>Some changes in Clang affect <a href="https://libcxx.llvm.org";>libc++</a>, + for example:</p> ---------------- ldionne wrote: > This follows my previous comment, but we really don't want to encourage Clang > folks to start using a `.DELETE_ME` file for testing Clang changes. Indeed, > roughly 54/55 of our CI jobs don't use the just-built Clang. The only > relevant CI job for a Clang change is the Bootstrapping build. As a result, > using a `.DELETE_ME` file is extremely wasteful of libc++ CI, since almost > all the CI jobs are completely irrelevant to the change being tested. > Furthermore, it will also provide a false sense of security for Clang folks, > who will think "hey but I ran the CI under all configurations", not > understanding the details. > > So I think we need to set up libc++ testing for Clang changes, but we should > discourage folks from running the libc++ CI itself as a poor proxy for > testing Clang changes. Most of the jobs are just not useful at all from the > perspective of testing a Clang change. > > So I'd rather not add this part of the documentation at all, to avoid > creating a problem for ourselves. > So I'd rather not add this part of the documentation at all, to avoid > creating a problem for ourselves. FWIW, I love the idea of setting up libc++ testing for Clang changes so we don't need a .DELETE_ME file. I think that's ultimately where we all would love to land. I'm fine either pulling these docs in anticipation of that work or leaving the docs here until that work is done. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D133249/new/ https://reviews.llvm.org/D133249 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
