erichkeane wrote: > > This brings me to conclusion that hand-written style of codegen tests > > presented above is a better fit for purposes of C++ defect report codegen > > tests. > > I share that conclusion. I think the hand-written IR is easier for people to > reason because it strips out unrelated details that make the test somewhat > more brittle than it needs to be for DR testing. I agree that we should use > the script whenever possible, but DR testing is a bit different from typical > codegen testing IMO. So I'm fine with the hand-crafted IR unless I'm missing > something.
Agreed, I find that using the script makes the tests WAY too fragile and hides the important parts of the test. I DISAGREE we should use the script when possible, as I basically never see a time where it improves a CFE test. https://github.com/llvm/llvm-project/pull/77637 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
