--- Comment #4 from Joost VandeVondele <Joost.VandeVondele at pci dot> 
2010-09-24 17:59:40 UTC ---
(In reply to comment #3)
> With Crayftn, there is no abort.

that also holds for NAG and g95, BTW.

> Thinking a bit about the program, I believe it is valid, i.e. gfortran has a
> wrong-code bug.

yes, I also think this testcase is valid. 

One little bit more tricky (in my mind) is the case where one declares T as
'TYPE(T1) :: T' in S1. It could be that that is not valid.

Configure bugmail:
------- You are receiving this mail because: -------
You are on the CC list for the bug.

Reply via email to