Re: pattern variable coercion

To: info-dylan@ai.mit.edu
Subject: Re: pattern variable coercion
From: Hugh Greene <q@tardis.ed.ac.uk>
Date: Sat, 6 May 2000 14:00:01 -0400 (EDT)
Organization: Division of Informatics, The University of Edinburgh
References: <8f10jr$eg$1@rznews2.rrze.uni-erlangen.de>
Xref: traf.lcs.mit.edu comp.lang.dylan:11962

On Sat, 6 May 2000, Gabor Greif wrote:
> While reading "Dylan Programming" I stumbled over coercion of pattern 
> variables [e.g.,]
> 
> define macro symbolize
>  { symbolize(?:name) } => { ?#"name" }
> end macro symbolize;
> 
> I could not derive this syntax from the BNF in the DRM. Is this a
> Harlequinism?

Nope, it's in the DRM, just not easy to find (like a few other things ;-). 
See the production /name-string-or-symbol/, under /substitution/, under
/template-element/. 

> .. AFAIK ##-style concatenation preserves the syntactical quality (being
> string, symbol or name) ...

Yes, it does -- see the section "Templates" in Chapter 10 of the DRM.

Hope that puts your mind at ease :-)
Hugh

References:

pattern variable coercion

From: "Gabor Greif" <gabor@mac.com>