http://gcc.gnu.org/bugzilla/show_bug.cgi?id=54618
--- Comment #21 from Salvatore Filippone <sfilippone at uniroma2 dot it> 2013-04-05 12:29:11 UTC --- (In reply to comment #20) > (In reply to comment #18) > > I am seeing intermittent issues with CLASS,ALLOCATABLE,INTENT(OUT) variables > > that have a CLASS,ALLOCATABLE component; however when I tried to reduce to > > the > > following test code, it segfaults on the inner clone, whereas the full code > > only fails on the outer code. In any case, the attached code is supposed to > > work, and it fails (under 4.7.2, 4.8.0 and trunk) > > For the code in comment 18 and comment 19, I get the output > T T > T T > using the latest draft patch at > https://userpage.physik.fu-berlin.de/~tburnus/final/ (It doesn't help with > comment 3.) Thus, if you build GCC yourself, that might be a work around. Thanks for the heads up. After further reflection, I've come to realize that for what I want to do, I can use INTENT(OUT) only when FINAL works. So, from now on I can fork a code with INTENT(OUT) and test your patch, but the main stable code will have to do with INTENT(INOUT) until FINAL is fully tested. Thanks a lot
