Re: [Jython-checkins] SF.net SVN: jython: [4345] trunk/jython/src

[Philip J. <pj...@un...>]

On Apr 17, 2008, at 12:07 AM, Samuele Pedroni wrote:
>
> wouldn't it make more sense for toString to correspond to __str__  
> and to __repr__ only if the first is not defined,
> notice that this would remove the need of special casing. Also I  
> think __str__ is really the conceptual correspondent
> to toString,   indeed both Python str.__str__ and java  
> String.toString are nops.

I agree, I was trying to avoid too big of a change in behavior from  
what we had before. I guess most of our builtin types don't actually  
do anything special with __str__ and __repr__ -- those methods are  
typically doing the same thing, so what you're proposing wouldn't  
change much there.

Our whole __str__/__repr__ setup is a little odd to me -- like for  
example PyObject's __repr__ is exposed as __str__.

--
Philip Jenvey