[Grok-dev] Re: Template pluggability and megrok.genshi finished.

Martijn Faassen faassen at startifact.com
Tue Oct 9 06:33:10 EDT 2007


Tres Seaver wrote:
> Martijn Faassen wrote:
>> Lennart Regebro wrote:
>> [snip]
>>
>>> megrok.genshi, Genshi templates in Grok:
>>> ------------------------------------------------------
>>> http://svn.zope.org/megrok.genshi/trunk/
>>>
>>> After feedback on the list, we now have no way of traversing from
>>> inside a template. I agree on the principle, but suspect it's going to
>>> be unwieldy, so maybe a "traverse" method or similar could be an idea.
>>> Maybe we should start using it first and see. Maybe not. :-)
>> I think we should end up with two versions of the Genshi integration:
>>
>> * the unwieldy, strictly controlled version where nothing is in the 
>> namespace except exactly what people pass in from their view. As far as 
>> I understand it, we have the infrastructure to build this now.
>>
>> * the "wieldy" one, where we *definitely* should pass in the 'path' 
>> function (as known from ZPT) into the Genshi template.
>>
>> Since we're working on the "wieldy" version first, I think we *should* 
>> pass in 'path' (and call it that way).
> 
> Your terms imply an obvious loading:  how about "well-factored" vs.
> "wild west" as alternatives?  Examples of the mess which which
> pull-model templates organically grow are not exactly hard to find in
> ZPT land.

Whatever terms you like. I think I understand the potential benefits of 
a push model:

http://faassen.n--tree.net/blog/view/weblog/2005/4/15/0

> Nevertheless, for *Grok*, the restrictions of push mode may be
> inappropriate.  Much of Grok's charm comes from its facilities for "solo
> mode" development (saving typing is one of Grok's key memes).  In such
> an environment, as opposed to multi-person, separation-
> of-responsibilities kind of projects, the advantages of push-mode are
> considerably less attractive.

Yes, exactly. I think Grok should scale up, which is why I said I 
thought it needed to support both. I think we figured out a way to do 
that at the sprint, but we'll just have to try this a bit more. The wild 
west approach seemed easiest to make progress on as it's so similar to 
ZPT, so in the interests of making progress, I wanted to focus on this 
initially.

I think we're just violently agreeing here. :)

Regards,

Martijn



More information about the Grok-dev mailing list