------- Comment #2 from burnus at gcc dot gnu dot org 2009-12-17 23:07 ------- I think that at the current result of gfortran is OK. However, I think a proper way would to send an interpretation request as we did before (PR39997, PR 40264). The question seems to be whether it is valid at all and, if so, which interpretation is correct. I think we have at least two or three examples which should be part in an interpretation request.
(I have filled PR 42418 for some issues found when testing the code in James last post, which is a bit unrelated to this PR.) -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=35913
