This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
[gfortran] patch for PR18398, problem with sequential formattedreads.
- From: Bud Davis <bdavis9659 at comcast dot net>
- To: "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>, gfortran <fortran at gcc dot gnu dot org>
- Date: Thu, 06 Jan 2005 00:12:40 -0600
- Subject: [gfortran] patch for PR18398, problem with sequential formattedreads.
the problem:
123456789012 (to see the columns of the data)
1.00000
2.00000
3.00000
read with F15.5 edit descriptor.
read_sf was detecting the eol, terminating the input of that item
and setting the flag; and then next_record_r was consuming
everything to the next eol, which means the second line of input
had disappeared.
i think the done flag here is not needed.. when eor has been
seen it is a bad idea to eat more input chars, period.
tested on i686/gnu linux with no additional testsuite failures.
no changes to the NIST results.
all examples posted with the bug (thanks guys !) are also corrected
by this patch.
--bud
2004-12-07 Bud Davis <bdavis9659@comcast.net
PR fortran/18398
* transfer.c (next_record_r): always skip the
eol search if it was found during sf_read.
! { dg-do run }
! pr18398, missing data on sequential formatted reads
! test contributed by Thomas.Koenig@online.de
open(7,status='scratch')
write (7,'(F12.5)') 1.0, 2.0, 3.0
rewind(7)
read(7,'(F15.5)') a,b
! note the read format is wider than the write
if (abs(a-1.0) .gt. 1e-5) call abort
if (abs(b-2.0) .gt. 1e-5) call abort
end
Index: gcc/libgfortran/io/transfer.c
===================================================================
RCS file: /cvs/gcc/gcc/libgfortran/io/transfer.c,v
retrieving revision 1.23
diff -c -3 -p -r1.23 transfer.c
*** gcc/libgfortran/io/transfer.c 24 Dec 2004 00:29:07 -0000 1.23
--- gcc/libgfortran/io/transfer.c 6 Jan 2005 05:56:03 -0000
*************** next_record_r (int done)
*** 1209,1215 ****
case FORMATTED_SEQUENTIAL:
length = 1;
! if (sf_seen_eor && done)
break;
do
--- 1209,1216 ----
case FORMATTED_SEQUENTIAL:
length = 1;
! /* sf_read has already terminated input because of an '\n' */
! if (sf_seen_eor)
break;
do