Date: Sun, 24 Jan 1999 13:47:52 -0500 (EST)
From: Michael Callahan <mjc AT rodagroup DOT com>
X-Sender: mjc AT evelyn DOT smith DOT edu
To: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
cc: djgpp-workers AT delorie DOT com, "Peter J. Braam" <braam AT cs DOT cmu DOT edu>
Subject: Re: Coda has TCP/IP and annonymous mmap for DJGPP on W95
In-Reply-To: <Pine.SUN.3.91.990124105741.4501O-100000@is>
Message-ID: <Pine.SGI.3.94.990124134257.10191A-100000@evelyn.smith.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp-workers AT delorie DOT com

On Sun, 24 Jan 1999, Eli Zaretskii wrote:

> On Fri, 22 Jan 1999, Michael Callahan wrote:
> 
> > (Note that this means that I do not support using mmap() to map the
> > contents of a file.)
> 
> Please note that this means the limited version of mmap you are
> suggesting should be called by some other name, not mmap.  Otherwise,
> packages that auto-configure themselves by trying to link in the mmap
> function will think that DJGPP supports it, and try to use it.

Hmm.  There are real Unices that support mmap() only for device-special
files (where it was used to support mapping the framebuffer in the X
server), so such autoconfig scripts are likely to come to grief already.
This was actually pretty common in late 1980s Unices, IIRC.

For Coda's purposes it would be a lot nicer not to have to modify all the
call sites where mmap() is used to use a different name.

However, I don't feel strongly enough about this to object much if the
community feels otherwise.

Note again: using mmap requires you to use a different crt0.o.  We have
set up djgpp so that this is an entirely separate -b machine option to the
gcc driver.  So programs that autoconfigure using the default djgpp-gcc
command don't see mmap at all.

Michael