Date: Sun, 21 Jan 2001 20:07:15 +0200
From: "Eli Zaretskii" <eliz AT is DOT elta DOT co DOT il>
Sender: halo1 AT zahav DOT net DOT il
To: Richard Dawe <rich AT phekda DOT freeserve DOT co DOT uk>
Message-Id: <2427-Sun21Jan2001200715+0200-eliz@is.elta.co.il>
X-Mailer: Emacs 20.6 (via feedmail 8.3.emacs20_6 I) and Blat ver 1.8.6
CC: djgpp-workers AT delorie DOT com
In-reply-to: <3A6B1BCF.FDCB78E7@phekda.freeserve.co.uk> (message from Richard
	Dawe on Sun, 21 Jan 2001 17:26:39 +0000)
Subject: Re: Moving _invent_inode() to a libc/*.h header file
References:  <3A6B1BCF DOT FDCB78E7 AT phekda DOT freeserve DOT co DOT uk>
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

> Date: Sun, 21 Jan 2001 17:26:39 +0000
> From: Richard Dawe <rich AT phekda DOT freeserve DOT co DOT uk>
> 
> The /dev/zero and /dev/full code needs to use _invent_inode() from
> src/libc/posix/sys/stat/xstat.c, to ensure that stat() or fstat() on
> either of these devices returns a unique inode. Currently I have a
> declaration in src/libc/fsext/fse_zero.c for the function, but I want to
> use a header. Now I could use a line like:
> 
>     #include "../sys/stat/xstat.h"
> 
> but these sort of includes always give me a bad feeling. I'd much rather
> do:
> 
>     #include <libc/something.h>
> 
> I'm not proposing to make src/libc/posix/sys/stat/xstat.h a header, but
> perhaps there should be a header of things useful to file system
> extensions - e.g. libc/fsext.h. What do you think?

I don't mind, but making a header for a single function seems
excessive.  You can always put the prototype inside the source.