Mailing-List: contact cygwin-developers-help AT sourceware DOT cygnus DOT com; run by ezmlm
Sender: cygwin-developers-owner AT sourceware DOT cygnus DOT com
Delivered-To: mailing list cygwin-developers AT sourceware DOT cygnus DOT com
Date: Sat, 24 Apr 1999 21:12:44 -0400
From: Chris Faylor <cgf AT cygnus DOT com>
To: Mumit Khan <khan AT xraylith DOT wisc DOT EDU>
Cc: cygwin-developers AT sourceware DOT cygnus DOT com
Subject: Re: [fortinj AT ibm DOT net: New snapshots]
Message-ID: <19990424211244.A18028@cygnus.com>
References: <199904250021 DOT UAA29450 AT envy DOT delorie DOT com> <199904250108 DOT UAA09494 AT modi DOT xraylith DOT wisc DOT edu>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 0.95.3i
In-Reply-To: <199904250108.UAA09494@modi.xraylith.wisc.edu>; from Mumit Khan on Sat, Apr 24, 1999 at 08:08:51PM -0500

On Sat, Apr 24, 1999 at 08:08:51PM -0500, Mumit Khan wrote:
>DJ Delorie <dj AT delorie DOT com> writes:
>> It's because of a recent egcs change that aligns the stack to 8-byte
>> boundaries, even stdcall functions that *must* pop 12 bytes from the
>> stack (i.e. the dll entry function).  I've complained to our internal
>> gcc group, but no response yet.
>> 
>> Chris - you have a patch for this.  Could you check it in with a
>> suitable comment, or forward it to me and I'll check it in?
>> 
>
>Could you please forward the patch to me as well? This is a side effect
>of the recent stack alignment patch John Wehle.

I don't have a patch.  I'm not sure why DJ thinks I have one.

The assembly output for stdcall functions is "interesting" now.  A
caller of a stdcall function apparently knows that the stdcall function
might pop too many things off the stack and fixes things up
appropriately.  That's interesting but it certainly isn't the correct
semantics for stdcall.

cgf