On Wed, Apr 17, 2002 at 12:01:00PM +0200, Stefan H. Holek wrote:
At 16.04.2002 15:44 -0400, Brian Lloyd wrote:
I am, as the author of the dtml-set tag, of course willing to commit to the implementation of this tag for 2.6
Ivo - I don't have a problem with the spelling of this. I _do_ have a problem with the fact that it (your existing release) actually stores the variable in REQUEST. If it were to store them somewhere more appropriate in the DTML namespace stack, I'd be happy to OK it.
Hm, but the REQUEST is where we want the variables to be set! Otherwise we'd use dtml-let, wouldn't we?. It is my understanding that the dtml-set tag is just REQUEST.set() with sugar on it (and sweet sugar that is!). Changing that semantics might (will?) break existing projects that already use dtml-set (several of mine, FWIW) where the REQUEST is later passed from DTML to e.g. PythonScripts...
What Ivo does is *exactly* what I want. What would "somewhere more appropriate" look like? What then would be the difference to dtml-let?
I understand that REQUEST.set()/dtml-set is to be used sparingly in general, but in some situations there is no alternative. And it appears to be a widely used Zope(2) idiom as well.
Exactly. dtml-set is just a user-friendly replacement for those horrible <dtml-call "REQUEST.set(..)"> constructs. I could make an optional attribute that makes the tag use the dtml namespace in stead, though I do not know what complications that would have at this moment. Cheers Ivo -- Drs. I.R. van der Wijk -=- Brouwersgracht 132 Amaze Internet Services V.O.F. 1013 HA Amsterdam, NL -=- Tel: +31-20-4688336 Linux/Web/Zope/SQL/MMBase Fax: +31-20-4688337 Network Solutions Web: http://www.amaze.nl/ Consultancy Email: ivo@amaze.nl -=-