[ZPT] Mini proposal: repeat-define and repeat-condition

[Dieter Maurer]

Evan Simpson wrote at 2003-1-24 17:05 -0600:
 > Dieter Maurer wrote:
 > > Casey Duncan wrote at 2003-1-23 10:24 -0500:
 > >  > Invent a new repeat tag that happens first and leave the existing repeat 
 > >  > tag alone.
 > > I like this proposal.
 > > 
 > > One minor problem: how do old and new repeat tag interact with one another.
 > > Probably, we can forbid that both occur on the same tag.
 > 
 > Hmm.  I'm beginning to like the original proposal better.  After all, 
 > while we would probably want to forbid the combination of old and new 
 > repeat statements on a tag, any combination of define, condition, 
 > repeat-define, and repeat-condition is sensible.
I think, we can forget about the "condition" altogether:

  Python provides list comprehension in a very intuitive way.
  Therefore, we do not need "repeat-condition".

  If you already have one way to do it, think at least thrice
  before you invent another one.

 > Not only that, all five statements are quite likely to be useful on a 
 > single tag:
 > 
 > <tr tal:define="items here/getItems"
 >      tal:condition="options/itemize"
 >      tal:repeat="item items"
 >      tal:repeat-define="name item/name;price item/price"
 >      tal:repeat-condition="python:name and price">
 > 
 > How about calling them "define2" and "condition2" instead, and allowing 
 > them even in the absence of a tal:repeat?  It's a little ugly, and could 
 > be taken to imply the existence of "condition3", etc., but would allow:
 > 
 > <div tal:define="x here/getX"
 >       tal:condition="python:len(x) == 2"
 >       tal:define2="x0 python:x[0];x1 python:x[1]"
 >       tal:condition2="python:x0 < x1">
 > 
 > The fact that they sandwich tal:repeat in the order of operations is 
 > sufficient rationale (for me) for having two sets without 
 > over-generalizing into "defineN" and "conditionN".
One of the original aims for TAL has been: KISS, keep it as simple
as possible. This was a good aim.


Dieter