[Petal] [PATCH] Options via constructor rather than package globals

[Steve Purkis]

On Thursday, June 5, 2003, at 01:24  pm, Grant McLean wrote:

>>> 2. cascade-able accessor methods which let you do:
>>>
>>> 	$template
>>> 	  ->language('fr')
>>> 	  ->input('HTML')
>>> 	  ->output('XHTML');
>> Now I find this really, really useless :)
>> I disagree with the idea of constructing an object that is in
> > an inconsistent state and setting the state afterwards through
> > some strange syntaxic sugar.
>
> I tend to agree with Jean Michel on this one - for exactly the same
> reasons.  I have seen it before but don't find it at all intuitive.

I cottoned onto it after reading "Smalltalk Best Practice Patterns".  
Excellent idea, IMHO.  But to each their own.


>>> Additionally, you might get rid of a lot of cut-n-paste work in 
>>> _initialize() by doing something like:
>>>
>>> [snip!]
>
> I actually started out with something a bit like that but the
> exceptions cases started to mount up and I decided the cut-n-
> paste version was clearer (although some of the lines ended up
> uncomfortably long).

Fair enough.


>> It might seem like a bit of a hack but I think it's acceptable and it
>> fits in better with the way Petal was designed to start with. Also, 
>> the
>> patch seemed to fail one test...
>
> It was passing all the tests when it left here :-)  Did the patch
> apply cleanly?  I haven't generated patches much before so I may have
> screwed that up.

I'd say it's prolly just tests doing their job, that's all.

Here's how I generate patches, if that helps:

	% diff -ruN Petal-0.92.orig Petal-0.92.new > Petal-0.92.some.patch

-Steve