Sender: nate AT cartsys DOT com
Message-ID: <35B25560.219B8E16@cartsys.com>
Date: Sun, 19 Jul 1998 13:21:52 -0700
From: Nate Eldredge <nate AT cartsys DOT com>
MIME-Version: 1.0
To: Andris Pavenis <pavenis AT lanet DOT lv>
CC: "John M. Aldrich" <fighteer AT cs DOT com>, djgpp-workers AT delorie DOT com
Subject: Re: Compile gives Abort! error
References: <Pine DOT A32 DOT 3 DOT 91 DOT 980717095017 DOT 54016A-100000 AT ieva05 DOT lanet DOT lv>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Precedence: bulk

Andris Pavenis wrote:
> 
> On Mon, 13 Jul 1998, Nate Eldredge wrote:

> > Then I think what we really need is for GCC to handle that more
> > gracefully.  A cursory glance through the FSF GCC sources didn't show
> > anyplace where it might abort if the file didn't open, but it still
> > might be there, or only in the DJGPP sources (which I don't have
> > handy).  This is where `abort' giving SIGABRT and a traceback would be
> > helpful...
> > --
> >
> 
> It is in original sources of gcc:
>   file:         choose_temp.c
>   procedure:    char * choose_temp_base ()    (it is very end of this file)
> 
> When gcc is not getting name of temporary file it's simply aborts.
> DJGPP port of gcc-2.8.1 now creates empty file when asked for name of
> temporary files to avoid conflicts between different DOS sessions
> running simultaneusly. Of course this fails if disk is read only.

What about doing at least a `perror' in that case?  That will help
identify this problem.
 
> If we'll not create file then we'll not be able to safely run gcc from 2 DOS
> sessions simultaneusly. In such situation some other program more likely
> cpp will complain.

Right.
-- 

Nate Eldredge
nate AT cartsys DOT com