On Apr 27, 2005, at 11:54 PM, Cris Daniluk wrote:
>>>> Just to make sure I understand this correctly... As technically
>>>> subclass
>>>> is never generated if it already exists in the target location, is
>>>> it
>>>> fair
>>>> to rephrase this as "'cgen' must allow a different destination
>>>> folder
>>>> for
>>>> subclasses and superclasses"?
>>>
>>> That would work too. That would be handy in that new classes would
>>> get
>>> generated there, and then I'd add them to my version control.
>>>
> Seems to me like this would be redundant on functionality that can
> just as easily be achieved using existing ant tasks. Rather than have
> the potentially confusing option to split the output dirs, I would
> have the option to simply not generate wrappers, or better, to leave
> it as it is... the more pure the generator the better, I think...
Having the output directories optionally split seems a good idea. If I
add a new table and new object mapping, I'm going to need the subclass
somehow, either manually created or generated. My first thought was to
avoid having subclasses generated and I'd manually create them, but
having them generated in a separate directory would save some time.
No argument that having the generator pure and clean is ideal and I
wouldn't want it any other way. But it also has to do what developers
need it to do :)
Erik
This archive was generated by hypermail 2.0.0 : Thu Apr 28 2005 - 08:26:28 EDT