Date:	Thu, 19 Mar 1998 15:08:45 +0200 (IST)
From:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To:	Vik Heyndrickx <Vik DOT Heyndrickx AT rug DOT ac DOT be>
cc:	Andris Pavenis <pavenis AT laima DOT acad DOT latnet DOT lv>, djgpp AT delorie DOT com
Subject:	Re: Rebuilding config.in in gcc-2.8.1
In-Reply-To:	<3510D96E.3E35@rug.ac.be>
Message-ID:	<Pine.SUN.3.91.980319150338.26251U-100000@is>
MIME-Version:	1.0

On Thu, 19 Mar 1998, Vik Heyndrickx wrote:

> > Does this mean that if Autoheader is run with an argument this problem is
> > avoided?  Can you try?
> 
> I can, but I don't understand what you want me to do exactly

It seems from that fragment that Autoheader is only doing this 
redirection when it is invoked without arguments.  I suggested to call 
it with an argument to work around the problem.  But without seeing the 
entire script (no, don't send it to me ;-), I can't tell if that indeed 
will avoid the problem.

> > Anyway, if the part of the script which mv'es $tmpout does that while
> > output is still redirected to $tmpout, then this indeed will not work on
> > non-Unix systems.  
> 
> I'm not so sure whether moving a file that is open is allowed on Unix
> systems.

Unix just marks the file for being moved/deleted and does that when it is
closed.  So it works there, and Unix programmers have come to expect that
behavior and rely on it.  It is simpler than to remember to delete/rename
it after it is closed, which might be in a totally different place.

> > Renaming or deleting files which are open by some program is
> > nothing but trouble on MS-DOS/MS-Windows.
> 
> I don't blame M$ for this. It's more than sensible not to allow that.

That's probably because we both come from the MS world.  Unix people 
would say we were brainwashed.

- Raw text -