shortmess/longmess were widely used before their current stint of being documented, which was done for [perl #114280]. FC suggested that we were forgetting something: https://rt.perl.org/Ticket/Display.html?id=114280#txn-1155408 Presumably this is that thing nobody could remember at the time. longmess and shortmess have always (in my digging) been documented as returning the same things as cluck and carp, either in those terms or by virtue of the same language being used to document each pair. I would rather that we fix longmess and shortmess. Almost certainly, anything that is "broken" by this change is actually being fixed, except for bad test expectations. The question is: how much is that? Where do we stand on doing comparison CPAN smokes? I think our current (or most recent) champion for that is Matthew Horsfall, who is, I believe, on vacation. -- rjbs