http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56696
--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> 2013-03-23
13:10:45 UTC ---
The following patch seems to fix the issue:
--- a/libgfortran/io/list_read.c
+++ b/libgfortran/io/list_read.c
@@ -1435,3 +1435,2 @@ read_real (st_parameter_dt *dtp, void * dest, int length)
CASE_SEPARATORS:
- case EOF:
unget_char (dtp, c); /* Single null. */
It's a very partial undo of the commit in PR 55818 comment 8. However, it does
not cause "eof_4.f90" to fail, which was added for that PR.
