------- Comment #18 from paulthomas2 at wanadoo dot fr 2006-12-07 17:33 ------- Subject: Re: Problem with handling optional and entry master arguments
elizabeth, > > We are talking about the same machine and the same operating system. All the > x86_64 binaries are from http://quatramaran.ens.fr/~coudert/gfortran. I am > surprised you said you didn't think anything changed since January 2006. Nothing changed in respect of this bit of code - that it ever works is down to luck. The entry accesses, or tries to access, the value of a pointer that is not set. Use -fdump-tree-original and look at the code that is produced. > > > By the way, Jakub's code works on the June version without optimization. It > fails if -O is used. It fails with "gcc version 4.3.0 20061114" even without > optimization. > Pure chance, I think. > I understand perfectly that you need to fix the bug, and the bug involves much > more than my applications, and I wouldn't feel any better if someone told me > an > earlier version of my code works. I just hope extra information helps. > Yes, I am sorry if I seemed grumpy; as the monkey said, "every little bit helps..." :-) The bug is there and could lead to segfaults at random - that I do not like! Thanks Paul -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25818
