Sender: |
|
Date: |
Sat, 28 May 2011 18:16:09 +0100 |
Reply-To: |
|
Message-ID: |
|
Subject: |
|
MIME-Version: |
1.0 |
Content-Transfer-Encoding: |
7bit |
In-Reply-To: |
|
Content-Type: |
text/plain; charset=ISO-8859-1 |
From: |
|
Parts/Attachments: |
|
|
On 28/05/2011 18:08, Bruno Le Floch wrote:
> Hello all,
>
> on the TeX.stackexchange Q&A website [1], was asked the following
> question: how do I get with xparse a macro with three optional
> arguments: <any number of *> <any number of '> <any positive integer>
> (without the <>). This can be achieved by providing a new argument
> type, which collects tokens as long as they obey a user-defined
> condition. Namely,
>
> \DeclareDocumentCommand{\X}
> {
> W { \token_if_eq_meaning:NNTF * }
> W { \token_if_eq_meaning:NNTF ' }
> W { \token_if_digit:NTF }
> }
> { \tl_show:n { #1 | #2 | #3 | #4 } }
>
> would declare \X to first grab tokens as long as they are equal in
> meaning to *, then grab tokens as long as they are equal in meaning to
> ', and finally grab digit tokens.
>
> Do you think that it would be useful to add yet another argument type
> for that? (In the code I have now, I decided that the argument
> grabbing would stop at the first { or space or }, or implicit such
> token.)
I have to say my initial reaction is 'not keen'. While xparse is a
'glue' layer between LaTeX2e and LaTeX3, and may not make it to a LaTeX3
format, I think we don't want to encourage irregular input if we can
avoid it. Do we _really_ want to suggest that this may even be
considered as a good idea for a new format?
On the other hand, I guess it does fall within the 'parsing' remit of
xparse, and may be something that is occasionally useful. I'd like to
see a real use case, I guess.
Joseph
|
|
|