X-Authentication-Warning: delorie.com: mailnull set sender to djgpp-workers-bounces using -f
To: Richard Dawe <rich AT phekda DOT freeserve DOT co DOT uk>
Cc: djgpp-workers AT delorie DOT com
Subject: Re: Fileutils 4.1 alpha 1 uploaded
References: <Pine DOT SUN DOT 3 DOT 91 DOT 1020414190622 DOT 237A-100000 AT is>
	<3CB9A954 DOT 40EBBAEC AT phekda DOT freeserve DOT co DOT uk>
In-Reply-To: <3CB9A954.40EBBAEC@phekda.freeserve.co.uk> (Richard Dawe's message of "Sun, 14 Apr 2002 17:07:48 +0100")
From: Jim Meyering <jim AT meyering DOT net>
Date: Sun, 14 Apr 2002 23:30:52 +0200
Message-ID: <87g01yrmr7.fsf@pixie.eng.ascend.com>
Lines: 52
User-Agent: Gnus/5.090006 (Oort Gnus v0.06) Emacs/21.2.50 (i686-pc-linux-gnu)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Reply-To: djgpp-workers AT delorie DOT com

Richard Dawe <rich AT phekda DOT freeserve DOT co DOT uk> wrote:

> Hello.
>
> Eli Zaretskii wrote:
>>
>> On Sun, 14 Apr 2002, Richard Dawe wrote:
> [snip]
>> > Also test
>> > suite hasn't been modified to work on DOS (lots of tests use <stem>.$$ for
>> > temporary filenames, which clearly doesn't work on DOS for PIDs > 999.)

Hi Rich,

How does my creation of a stem.$$ temporary _directory_ (in the current
directory) cause trouble?  AFAICS (for the few tests I looked at), there
is no risk that more than one of those little scripts will run at the
same time, so there shouldn't be a problem on that front.

Of course, if some tests create such a temporary directory in $TMPDIR,
then that's a good reason to change it, and I'd appreciate patches that
do that.

I'm not too keen on mangling anything -- even temp-file names in test
scripts -- to accommodate DOS's brain-dead 8.3 limitation.  Of course,
if you find a clean way to fix all this without mangling things, I'll
be a lot more receptive.

>> The canonical way of solving this is to create a directory $TMPDIR/tmp$$ (up
>> to 8 characters for a 16-bit PID, so fits into DOS 8+3 limitations), and
>> then create temporary files in that directory.
>
> That's exactly what I did for Fileutils 4.0 (but thanks for the suggestion).
> Some of the tests may have been patched in the same way as the DJGPP port of
> 4.0 - I can't remember. But there are several more tests in 4.1, so I need to
> re-evaluate them.
>
>> Jim, will you accept patches for such an approach?
>
> I have a bunch of patches to feed back, now I've caught up with the latest
> stable release. ;) The primary set is to the test suite to cope with DJGPP
> programs reporting their names as e.g.: 'c:/djgpp/bin/cp' instead of just
> 'cp'. (We discussed this before.)

Sorry, but I'm really not inclined to spend time tweaking *many* of
the tests to work on systems where argv[0] != "cp" for commands like
`cp a b'.  I'm not even inclined to apply patches to allow that, since
it makes each little script more involved -- I went down that road quite
a ways, back before I changed things to set PATH and use simple command
names rather than using absolute file names.  The old way was too painful.

Jim