Re: g77 potential bug with "backspace" command

["Tim Prince" <tprince at computer dot org>]

Your example appears to be working as you intended on my current
installation of cygwin, and also with gcc-3.1/g77.  It's difficult for
anyone to help with such an old version as you quote.  Current cygwin is
several giant steps ahead of what it was 3 years ago.
----- Original Message -----
From: "Hudson, Peter O" <HudsonPO@navair.navy.mil>
To: <gcc-bugs@gcc.gnu.org>
Sent: Monday, December 31, 2001 2:43 PM
Subject: g77 potential bug with "backspace" command


> Gentlemen,
>
> 1st off, thanks for the outstanding FORTRAN compiler.   In the R&D world
we
> have to use and re-port a lot of legacy researchy codes ...often written
in
> FORTRAN.
>
> During one such instance is a slightly older NASA program that uses
namelist
> formatted input files along with some free format data.
>
> The free format data is read in a loop until it errors out of the loop
when
> it encounters the next namelist. It then uses BACKSPACE to back up and
read
> that namelist.  (probably not how I'd have done it).
>
> At any rate the backspace command errors out of the program rather than
> backing up.  The same result is duplicated in the enclosed test code
called
> backspace.f
>
>  <<backspace.f>>  <<backspace.dat>>
> My work around was to add a junk namelist after the free format data.  It
> could chew that up and spit out then read the next namelist for real...not
> perfect but it works.  This let me eliminate the backspace command in the
> source.
>
> I'm not sure if "backspace" is supported or if this bug report should
really
> be an "I'd like to see this added".  But, at any rate it seems like a
simple
> thing to fix.
>
> Specifics details:
> running on Intergraph Pentium 400
> Windows NT
> in a CYGNUS-B20 shell
>
> source compiles fine, and works fine.
>
> thanks for your attention:
>
> -Peter Hudson-
> Mechanical Engineer
> NAWCWD China Lake
>
>
>
>
>
>
>