agozillon wrote:
Most recent updates were a rebase for the entire stack and an update of this current PR (Fortran level changes) to try and address the very good feedback. However, first time doing this using SPR, so if anyone notices any weirdness please don't hesitate to point it out! The main changes from the update are: * Attempting to remove the few cases of hard-coded SSA values from tests * Early exitting where suggested and where possible * Emission of assert when certain expected conditions are not met * Movement of mapSymbols to a required argument for processMap, as we always end up generating the list in any case now as with the current implementation for finding prior mapped parents we need to keep track of the symbol list * Removal of seperate getOmpObj... function for parent, I've opted to just use getFirstName for this PR, I think it may need something a little more robust for nested derived types (I made an attempt at making something that'd retrieve the parent one up from the current element, but it's not required for this PR and I'd like to test it a bit more first), but no sense getting ahead of myself if I can keep it simple for now * Better/clearer method of getting the derived type definition symbol that we utilise to access the member symbols for calculating the member indices https://github.com/llvm/llvm-project/pull/81511 _______________________________________________ llvm-branch-commits mailing list llvm-branch-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
