[Grok-dev] Re: Patch to grokproject

[Philipp von Weitershausen]

On 12 Aug 2007, at 19:33 , Luciano Ramalho wrote:
>>> I also think that skeleton doctests (unit and functional) should be
>>> generated by grokproject, for the same reason a skeleton app module
>>> and index template are generated.
>>
>> Hmmm, perhaps. What do others think? Too much boilerplate?
>
> Boilerplate is always bad. At best, it's a necessary evil: code we are
> condemned to copy and paste forever for the sake of flexibilty, for
> those rare opportunities when we need some variation on the
> boilerplate.
>
> Once we assume that the boilerplate is needed, for instance, to
> properly setup tests, and to perform the tests themselves, I think it
> is natural for a user friendly tool like grokproject to provide the
> boilerplate ready to use. It least that is what every other major web
> framework these days is doing. Especially since the grok code base
> itself is not a very good source for the testing boilerplate code,
> since it has it's own particular testing needs.
>
> But I agree that a good framework should reduce the need for
> boilerplate code to a minimum.

Right. Grok should probably grow a little bit of code that would make  
the testing boilerplate in grokproject superfluous, or at least  
reduce it to a minimum...