Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 10 Oct 2001 12:57:07 -0700
From:      Jordan Hubbard <jkh@freebsd.org>
To:        anarcat@anarcat.dyndns.org
Cc:        alex@big.endian.de, freebsd-libh@freebsd.org
Subject:   Re: adding Disk::freespace to lib[h]disk
Message-ID:  <20011010125707J.jkh@freebsd.org>
In-Reply-To: <20011010152124.A1233@shall.anarcat.dyndns.org>
References:  <20011010124341.C531@shall.anarcat.dyndns.org> <22961.1002736882@critter.freebsd.dk> <20011010152124.A1233@shall.anarcat.dyndns.org>

next in thread | previous in thread | raw e-mail | index | archive | help
You're making perfect sense.  We've long known that libdisk(3) needed
to be taken out in back of the barn and shot through the head, but there's
just been nobody around to do it so we're still stuck with that evil piece
of.... code.  The person who re-implements libdisk correctly will get
a whole box of chocolate chip cookies in the mail from me. :)

- Jordan

From: The Anarcat <anarcat@anarcat.dyndns.org>
Subject: Re: adding Disk::freespace to lib[h]disk
Date: Wed, 10 Oct 2001 15:21:25 -0400

> On Wed Oct 10, 2001 at 08:01:22PM +0200, Poul-Henning Kamp wrote:
> > 
> > Just for the record:  I'm not actively interested in libdisk...
> 
> Noted. phk removed from cc: :)
> 
> Ok, then I guess libdisk is left to itself now?
> 
> I think libdisk should be rewritten, maybe keeping a compatibility layer
> with the current libdisk.
> 
> This is because I think there are some fundamental designs flaws in
> libdisk. For example, it mixes "undo management" with low-level disk
> writing routines. libdisk should be low-level, that is, not an "image"
> of the current that gets written only when write() is called, but a set
> of clearly defined operations that operate directly on the disk.
> 
> An additional "upper level" layer might be implemented to provide the
> kind of facilities libdisk provides (undo(), write()).
> 
> The problem with current implementation is that the disk contains
> implementation-specific data (ie, subtype codes) that are undocumented
> and sometimes conflict with library specific data (ie type codes:
> [0-6]).
> 
> So what do you people think? Am I making any sense?
> 
> A.

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-libh" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20011010125707J.jkh>