User-Agent: Thunderbird 2.0.0.22 (Windows/20090605)
MIME-Version: 1.0
References: <4A7921CF.5020803@morningstar2.co.uk>           
            <4A7A1505.4040604@residenset.net>
            <4A7A762A.3040105@morningstar2.co.uk>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Message-ID:  <4A7A8688.9030808@morningstar2.co.uk>
Date:         Thu, 6 Aug 2009 08:30:16 +0100
Reply-To: Mailing list for the LaTeX3 project
              <LATEX-L@LISTSERV.UNI-HEIDELBERG.DE>
Sender: Mailing list for the LaTeX3 project <LATEX-L@LISTSERV.UNI-HEIDELBERG.DE>
From: Joseph Wright <joseph.wright@MORNINGSTAR2.CO.UK>
Subject: Re: xparse
To: LATEX-L@LISTSERV.UNI-HEIDELBERG.DE
In-Reply-To:  <4A7A762A.3040105@morningstar2.co.uk>
Precedence: list
Status: R

I wrote:
> [Discussion about \IfNoValue]

I think there are two sensible ways to approach things:

1) Insist that \NoValue can only be used directly inside xparse
functions, document this and provide an example of how to pass data
along (using \q_no_value, as in my example). In this case, \IfNoValue
does not need to be expandable.

2) Document that \NoValue may be passed along, and as a result ensure
that \IfNoValue is expandable.

The current situation seems to fall between the two (\IfNoValue is not
expandable *but* there is no guidance on any restriction).  As Lars
says, there are cases where an expandable \NoValue would be very useful
(although I wonder if my suggestion of effectively converting \NoValue
into \q_no_value at the xparse stage "fits" better the separation of
argument parsing from coding).
-- 
Joseph Wright