tbaederr wrote: Endianness is very well possible indeed, but it's a little weird this starts failing now, since the shifts done aren't vector specific. Are you testing this on a solaris host? When you say stage2-build, you mean you're building clang main with your system compiler (this is stage1, and the tests pass here?) and then use that compiler to build the same clang main again, but this time the tests fail?
https://github.com/llvm/llvm-project/pull/108949 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
