This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: [gfortran] patch for pr14942 -- list directed I/O
- From: Steve Kargl <sgk at troutmask dot apl dot washington dot edu>
- To: Paul Brook <paul at codesourcery dot com>
- Cc: fortran at gcc dot gnu dot org, Bud Davis <bdavis9659 at comcast dot net>, gcc-patches at gcc dot gnu dot org
- Date: Sun, 18 Apr 2004 10:55:26 -0700
- Subject: Re: [gfortran] patch for pr14942 -- list directed I/O
- References: <1082081238.9535.14.camel@localhost.localdomain> <200404181845.57956.paul@codesourcery.com>
On Sun, Apr 18, 2004 at 06:45:57PM +0100, Paul Brook wrote:
> On Friday 16 April 2004 03:07, Bud Davis wrote:
> > This patch corrects the specific examples listed in the PR. It
> > does not break anything else I am aware of, but I doubt if list
> > directed I/O works 100%, it just works better :)
> >
> > Change Log
> >
> > 2004-04-15 Bud Davis <bdavis9659@comcast.net>
> >
> > PR fortran/14942
> > * io/list_read.c (read_real,read_complex,read_integer
> > read_logical): continue reading when encountering a
> > separator.
> > * io/list_read.c (list_formatted_read): continue after
> > a separator.
>
> This looks wrong. If we do find a seperator we need to loop back to the start
> of the function, not keep going. This is particularly obvious in the logical
> case, where we use v uninitialised.
>
I can confirm that the patch does not fix all cases of list-directed
IO. This patch permits the lapack testsuite to procede beyond the
original failure point to a new list-directed IO failure. My wrong
patch for save_length was generated while I tried to determine the
the cause of the new failure.
--
Steve