tahonermann wrote: > Unless we document this and insist on it for future code reviews (which IMO > is not required), the consistency is not going to last. I don't think think > we need to mandate a specific triple because functionally none of it is wrong.
I was more concerned about encouraging consistency rather than mandating it. And more specifically about encouraging testing of `spirv64` vs `spir64`. I agree this isn't a case of right or wrong. I'm most interested in the question of whether we should provide an explicit target at all for tests that are not intended to be target dependent. I tend to think that we should exercise the default unless there is a good reason not too. For now, the default device target does not vary by host environment, but that could change in the future (it might be that the default of spirv64 doesn't change, but the selection of ABI might and therefore affect the "-unknown-unknown" part). I understand if you would prefer such a change to be done via a separate PR. https://github.com/llvm/llvm-project/pull/172366 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
