mordante wrote: > > @ilya-biryukov this is the new attempt to land this patch, the version you > > committed breaks the libc++ CI. I'll try to fix it this week. > > Wow, I didn't expect that. Is libc++ ok with disabling this test until the > nightly build catches up? (The PR is in the Draft state, so it's hard to tell > if that's acceptable).
It's not ideal, but I can't think of a better solution. Typically nightly apt builds are built nightly so then we can update the Docker CI quickly. If the nightly builds are red it might take a bit longer. The patch is in draft since I wanted to test with the bootstrap build which required temporary disabling other builds (the second commit). When the CI is green I'll do a final review and publish the patch. > In the long-term, how should people land changes like this? Compiler > diagnostics are expected to change sometimes, so we should figure out some > way that allows landing these without reverts. Any thoughts? Typically we work together with whom changes the tests. When there are libc++ changes it is visible in our review queue. For changes in diagnostics it's often possible to use a regex matcher. For example, the messages of `static_asserts` changed a while back. In general we prefer to minimize the tests depending on the compiler diagnostics. https://github.com/llvm/llvm-project/pull/76833 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
