jeanPerier wrote: The determinism of the failure with this patch is disturbing and I think I now know why: you are adding a test with `module m1` (in flang/test/Lower/zero_init.f90), which is likely the source of the clash since `module_use.f90` is reading a module also named m1.
This means your patch to `module_use.f90` test did not fix the issue. I am just realizing `-J` in bbc is the path for intrinsic modules, so `-I` is probably the one needed, but maybe the best is just to `cd` into `%t` and run the test there (there is a precedent for doing cd %t in test/Lower/OpenMP/cfg-conversion-omp.private.f90). https://github.com/llvm/llvm-project/pull/123330 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
