X-Authentication-Warning:	acp3bf.physik.rwth-aachen.de: broeker owned process doing -bs
Date:	Wed, 19 Jan 2000 14:23:48 +0100 (MET)
From:	Hans-Bernhard Broeker <broeker AT physik DOT rwth-aachen DOT de>
X-Sender:	broeker AT acp3bf
To:	djgpp-workers AT delorie DOT com
Subject:	Re: (patch) updated protoize patch
In-Reply-To:	<3885476F.51242B02@hmc.edu>
Message-ID:	<Pine.LNX.4.10.10001191419090.15464-100000@acp3bf>
MIME-Version:	1.0
Reply-To:	djgpp-workers AT delorie DOT com
Errors-To:	dj-admin AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

On Tue, 18 Jan 2000, Nate Eldredge wrote:

> Mumit Khan wrote:
> 
> > Thanks for your comments. Since you mention that foo.c.X is not a legal
> > DOS filename, there is another one to deal with -- currently a foo.c is
> > saved as foo.c.save. I've tentatively changed that to `foo.csaved'
> > instead, which looks ugly but at least workable.
> 
> Mmm... and what do we do about foo.bar?  foo.barsaved is truncated to
> foo.bar and overwrites the original file.
> 
> foo.sav might be more reasonable.

My recommendation would be to look at the methods in the DJGPP port of GNU
'patch'. It has a similar dilemma (saving 'foo.c.orig' and possibly
'foo.c.rej' style files, within DOS 8.3 limitations), and IIRC, the
solution was rather thoughtfully set up by Eli, with some additional input
from me:

	foo.c.orig    becomes foo.c~
	foo.c.rej             foo.c#
	foo.bar.orig          foo.ba~
        foo.bar.rej           foo.ba#

(or something like that). Of course, every such pattern must break, in
some cases, like if you have two files named

	12345678.abc
        12345678.abd

or similar, i.e. full-length names, and the only difference in the very
position that would be replaced by '#' or '~'... But at least, it's 
trying to do as much as can.

Hans-Bernhard Broeker (broeker AT physik DOT rwth-aachen DOT de)
Even if all the snow were burnt, ashes would remain.

- Raw text -