This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug fortran/18481] [g77 regression] ICE with assigned integer variable format
- From: "Thomas dot Koenig at online dot de" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 4 Dec 2004 10:37:13 -0000
- Subject: [Bug fortran/18481] [g77 regression] ICE with assigned integer variable format
- References: <20041114203659.18481.Thomas.Koenig@online.de>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Additional Comments From Thomas dot Koenig at online dot de 2004-12-04 10:37 -------
This does indeed appear to be the problem
with quite a few failing NIST tests. Here's
a reduced testcase from NIST 111. The test case
in question has the comment
C***** - USE AS A FORMAT AN INTEGER VARIABLE WHOSE VALUE 10.3 02280111
C***** IS ASSIGNED USING AN ASSIGNMENT STATEMENT. 12.4(2) 02290111
so I guess this must be legal. It also works with g77.
$ cat assign.f
program main
assign 9000 to i
print i,3.14
9000 format ('real value = ', g12.5)
end
$ g77 assign.f && ./a.out
real value = 3.1400
$ gfortran assign.f
In file assign.f:2
assign 9000 to i
1
Warning: Obsolete: ASSIGN statement at (1)
assign.f: In function 'MAIN__':
assign.f:2: internal compiler error: in gfc_add_modify_expr, at fortran/trans.c:154
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
Same error for print instead of write.
--
What |Removed |Added
----------------------------------------------------------------------------
Summary|ICE with integer variable |[g77 regression] ICE with
|"format" |assigned integer variable
| |format
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18481