[huntharo@msu.edu: Patched w32api headers]

[Chris Faylor <cgf at cygwin dot com>]

----- Forwarded message from Harold Hunt <huntharo@msu.edu> -----

From: "Harold Hunt" <huntharo@msu.edu>
To: <cgf@cygwin.com>
Subject: Patched w32api headers
Date: Mon, 26 Mar 2001 23:12:40 -0500
In-Reply-To: <20010326193340.F2921@redhat.com>

Chris,

To build the Test series of servers, which #include w32api headers, I have
to wrap a couple things in the w32api headers with #ifndef XFree86Server;
the things being wrapped are, fortunately, things that are either identical
to their X counterpart, such as INT32, or they are things that will never be
used by Cygwin/XFree86, such as ATOM, RT_CURSOR, and RT_FONT.

My question for you is whether I should ever submit these as patches to
Cygwin, or if I should instead distribute the diffs to people wishing to
compile Cygwin/XFree86.  I'm having a hard time with this question, because
it is unusual to see system headers get modified for a single application;
while on the other hand, this is Cygwin/XFree86, so maybe it would make
sense in this case.

The diffs are actually quite small, they just wrap the following with
#ifndef XFree86Server:
winuser.h	RT_CURSOR RT_FONT
basetsd.h	INT32
windef.h	ATOM BOOL BYTE
winbase.h	FreeResource

I also added a DUMMYUNIONNAME to the union in the structure
REPARSE_DATA_BUFFER around line 2472 in winnt.h.  That isn't essential, it
was just to get rid of a warning about how ANSI C doesn't support unnamed
structures/unions.

Well, what do you think?

Harold

----- End forwarded message -----

-- 
cgf@cygnus.com                        Red Hat, Inc.
http://sources.redhat.com/            http://www.redhat.com/