Bug 25041 - gfortran internal compiler error: Segmentation fault when compiling 'rksuite_90'
Summary: gfortran internal compiler error: Segmentation fault when compiling 'rksuite_90'
Status: RESOLVED DUPLICATE of bug 24266
Alias: None
Product: gcc
Classification: Unclassified
Component: fortran (show other bugs)
Version: 4.1.0
: P3 major
Target Milestone: ---
Assignee: Not yet assigned to anyone
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2005-11-26 00:00 UTC by Todd Hay
Modified: 2006-01-01 06:53 UTC (History)
5 users (show)

See Also:
Host: i386 FreeBSD 6.0
Target: i386 FreeBSD 6.0
Build:
Known to work:
Known to fail:
Last reconfirmed: 2005-12-27 23:52:32


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Todd Hay 2005-11-26 00:00:29 UTC
The file 'rksuite_90.f90' from the archive 'rksuite_90.tar' from www.netlib.org or  http://math.la.asu.edu/~eric/mat420/rksuite_90.tar produces an internal compiler error:

$ gfortran41 -c rksuite_90.f90       
rksuite_90.f90: In function 'interpolate_r1':
rksuite_90.f90:2646: internal compiler error: Segmentation fault: 11
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.

I found bug #22160 but I believe that this is a different problem. I can compile the file mentioned in bug #22160 just fine.
Comment 1 kargls 2005-11-26 00:11:44 UTC
How are you building this?  If I fix and run the make_rk script
for all possibilities, I get a 16000+ line program.
Comment 2 kargls 2005-11-26 02:49:53 UTC
Reduce test case.

module z
   type b
     character(len=80) :: r(10)
   end type b
     contains
     subroutine a(c)
       type(b), intent(inout), target :: c
       write (c%r,"(a)") "death"
     end subroutine a
end module z
Comment 3 Todd Hay 2005-11-26 06:03:38 UTC
(In reply to comment #1)
> How are you building this?  If I fix and run the make_rk script
> for all possibilities, I get a 16000+ line program.
> 

I built it with the "r1" option ($make_rk r1) for 1D array dependent variables. Sorry for not being specific previously.
Comment 4 Andrew Pinski 2006-01-01 06:53:47 UTC
This is actually a dup of bug 24266.

*** This bug has been marked as a duplicate of 24266 ***