Revision 1.64 of libgfortran/io/transfer.c causes a regression for the
following testcase:
% cat <<EOF > regression.f90
program demonstration
implicit none
character(len=1) :: saux
open(unit = 1, FILE = "foo.conf", STATUS = 'OLD', ACTION = 'READ')
do
10 read(unit = 1, fmt = '(a)', advance = 'no', end = 99, eor = 11) saux
print *, "'", saux, "'"
cycle
11 print *, "Newline"
end do
99 close(1)
end program
EOF
% cat <<EOF > foo.conf
foo1: bar
1.2
EOF
ifort and g95 compiled binaries give:
'f'
'o'
'o'
'1'
':'
' '
'b'
'a'
'r'
Newline
'1'
'.'
'2'
Newline
while gfortran 4.1.0 20051024 gives only the characters from the first line of
'foo.conf', all characters after the first 'end of record' are whitespace:
'f'
'o'
'o'
'1'
':'
' '
'b'
'a'
'r'
Newline
' '
' '
' '
Newline
For sake of compliance with the bug report policy:
* gfortran from cvs via
TZ=GMT cvs -q update -D'2005.10.24.03.00.00'
works,
* gfortran from cvs via
TZ=GMT cvs -q update -D'2005.10.24.04.00.00'
does not.
--------
By the way, there is another small issue with the above testcase. I'm not sure
if I should file a separate bug report for it:
Changing line 6 of regression.f90 to
10 read(unit = 1, fmt = '(a)', advance = 'no', end = 99, eor = 10) saux
gives the incorrect warning:
In file regression.f90:6
10 read(unit = 1, fmt = '(a)', advance = 'no', end = 99, eor = 10) saux
1
Warning: Branch at (1) causes an infinite loop
while ifort and g95 do not complain.
--
Summary: Summary: [4.1 Regression] Nonadvancing read does not
read more than 1 line
Product: gcc
Version: 4.1.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: fortran
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: sven dot buijssen at math dot uni-dortmund dot de
GCC build triplet: i686-pc-linux-gnu
GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=24719
