Date: Sat, 06 Jan 2001 15:18:55 +0200
From: "Eli Zaretskii" <eliz AT is DOT elta DOT co DOT il>
Sender: halo1 AT zahav DOT net DOT il
To: Martin Str|mberg <ams AT ludd DOT luth DOT se>
Message-Id: <3995-Sat06Jan2001151855+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, ceo AT nbensacomputers DOT com
In-reply-to: <200101061244.NAA18812@father.ludd.luth.se> (message from Martin
	Str|mberg on Sat, 6 Jan 2001 13:44:07 +0100 (MET))
Subject: Re: Fw: Patch for statfs.c
References:  <200101061244 DOT NAA18812 AT father DOT ludd DOT luth DOT se>
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

> From: Martin Str|mberg <ams AT ludd DOT luth DOT se>
> Date: Sat, 6 Jan 2001 13:44:07 +0100 (MET)
> 
> According to Eli Zaretskii:
> > > From: "Norberto Alfredo Bensa" <ceo AT nbensacomputers DOT com>
> > > Date: Fri, 5 Jan 2001 18:40:47 -0300
> > > 
> > > I've just tested on Win98 plain DOS mode and 213600 gives:
> > > blocks = 32768, bsize = 4096, resulting in 134,217,728 or 128MB...
> > 
> > Expected: 213600 is limited to 128MB, by its very definition.
> 
> That is a little surprising as it manages to report drive sizes up to
> ~4GiB in DOZE...

For a CD, it reports the correct cluster size (2K), and this limits it
to 128MB, because the maximum number of clusters on a FAT system is
64K.

> > > Under Windows (DOS box) with 217303, I get:
> > > blocks = 19275, bsize = 32768, resulting in 631,603,200 or ~602.3MB
> > 
> > That's a lie: the block size is reported incorrectly (2048 is the
> > right value).  This clearly shows that we shouldn't use 217303 for
> > CDs.
> 
> I realise that a bsize of 32768 is a lie, however ...
> 
> > > Properties for my E: drive gives 631,603,200, that's what 217303 is
> > > reporting.
> > 
> > Small wonder: I guess Windows uses 217303 itself.
> 
> ... shouldn't we let WINDOZE have it's way. The INT21 AX=0x7303 calls
> do give reasonably accurate values when they succeed and it's what
> WINDOZE itself reports.

You assume that the only value the application is interested in is the
total size.  What if it wants to know the cluster size?

I think a small inaccuracy in the total size is worth reporting an
accurate cluster size.  (I'm also quite sure the value reported by the
CD-specific calls is also correct, but Windows might add to it
something, like the TOC or whatsnot.)