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
From: Chris Faylor <cgf AT cygnus DOT com>
Date: Mon, 3 May 1999 22:26:25 -0400
To: Mumit Khan <khan AT xraylith DOT wisc DOT EDU>
Cc: Corinna Vinschen <corinna DOT vinschen AT cityweb DOT de>,
        cygwin-developers AT sourceware DOT cygnus DOT com
Subject: Re: (fixed patch) "%E" formatting for the humans
Message-ID: <19990503222625.A2251@cygnus.com>
References: <372E2625 DOT 791A1EA AT cityweb DOT de> <199905040206 DOT VAA25326 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: <199905040206.VAA25326@modi.xraylith.wisc.edu>; from Mumit Khan on Mon, May 03, 1999 at 09:06:48PM -0500

On Mon, May 03, 1999 at 09:06:48PM -0500, Mumit Khan wrote:
>Corinna Vinschen <corinna DOT vinschen AT cityweb DOT de> writes:
>> 
>> FormatMessage has an option, to allocate the needed buffer itself
>> (FORMAT_MESSAGE_ALLOCATE_BUFFER or similar). This buffer should
>> then be freed with LocalFree() after usage.
>> Would this be an alternative?
>
>That was what I had done at first, only to discover that it doesn't fit
>into the model used by small_vsprintf etc al very well (passed in pre
>allocated buffer, arbitrary number of formatting directives, etc make
>it impossible to inject a newly allocated buffer within). I'd welcome 
>any suggestions of course.

The small_vsprintf model is the vsprintf model.

In the case of %E, I believe that this is only actually used by strace
and api_fatal.  strace_printf has a 10,000 byte buffer, api_fatal uses a
4096 byte buffer.

I would welcome a patch if someone wants to modify cygwin to use the
vnsprintf model instead.

cgf