earnol wrote: First of all, thank you for looking into this PR. Yes. I also checked with other compilers and noticed the same with gcc, however what others do is not a point of authority here for me - the C standard and practical considerations are. And the practical considerations are the multidimensional array[m][n] is guaranteed to be stored as a whole block of size n × m, thus making pointers actually compatible. Also this specific condition was not tested at all. I had to changed zero lit tests when making my PR. This is the second factor which convinced me my change is correct.
https://github.com/llvm/llvm-project/pull/159896 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
