Date: Wed, 26 Jun 1996 16:17:34 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Demmer AT LStM DOT Ruhr-Uni-Bochum DOT De
Cc: "Alexander V. Lukyanov" <lav AT video DOT yars DOT free DOT net>, djgpp AT delorie DOT com
Subject: Re: FSE, some questions
In-Reply-To: <C399321855@brain1.lstm.ruhr-uni-bochum.de>
Message-Id: <Pine.SUN.3.91.960626160908.8317C-100000@is>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII


On Wed, 26 Jun 1996, Tom Demmer wrote:

> > Why would you want to call `_close'?  After all, `fd' is not a file 
> > handle, so why call a function that only knows about true file handles?  
> > 
> Because fd _IS_ a true filehandle for DOS. Connected to "NUL". 

Who did connect it to "NUL" and why?  I don't think this is necessary for 
the operation of filesystem extensions to have a true file handle, it can 
be just an int.

Anyway, if for some reason I don't get `fd' *must* be a file handle, you 
should just call DOS function 3Eh via `__dpmi_int' instead of calling 
`_close'.

> For the second problem, Alexander Lukyanov proposed to add ioctl()
> to libc. Well, let's see.

I'm not sure this is such a good idea.  Under MS-DOS, `ioctl' means a
totally different thing, so it might be a source of subtle bugs (if
somebody tries to compile a program that comes from BC, say, and uses
`ioctl').  I remember some discussion about this about a year ago.  Maybe
its worthwhile to ask DJ.