I agree that breaking the 1.0 code-base would be a bad thing.
However, I believe the 1.0 release should be as "correct" as possible.
Following my own definition of "correct", and recognizing that Struts
openly claims Beta status I believe this feature should be included.
Following Roland, I have trouble imagining a scenario where this
should cause problems. Basically, you know that the framework invokes
your forms reset method, therefore you need to make sure it represents
the state you intend. If someone has another scenario, please chime in.
Regards,
Levi Cook
> -----Original Message-----
> From: Roland Huss [mailto:[EMAIL PROTECTED]]
> Sent: Wednesday, June 13, 2001 4:27 PM
> To: [EMAIL PROTECTED]
> Subject: Re: Calling ActionForm.reset() from <html:form> when
> creating a
> bean
>
>
> "Craig R. McClanahan" <[EMAIL PROTECTED]> writes:
>
> > Bugzilla #2108 <http://nagoya.apache.org/bugzilla/show_bug.cgi?id=2108>
> > asks a very reasonable question -- whey isn't the reset() method called
> > when a new form bean is created in the <html:form> tag? After all, the
> > controller servlet will always call reset() on the form bean (whether or
> > not it is newly created) before calling populate() to copy the request
> > parameters.
>
> Recently I would indeed have needed this feature to initialize the
> possible choices for multi-choice select box using <html:options>
> attribute "property" to return a collection for the multiselect
> box. But it since it wasn't called upon form creation, the list has
> never been initialized. My solution was to use the "collection"
> attribute which takes a collection put into the request by the
> previous action, which after all was the cleaner solution anyway (no
> logic within forms !).
>
> > This seems like a perfectly reasonable thing to add for 1.1. My
question
> > is, what do folks think about making this change for 1.0 as well? I'm
> > more than a little concerned that it would break existing code, but the
> > consistency argument might make doing this worthwile anyway. (Of
course,
> > we would highlight this change in the Release Notes.).
> >
> > What do you think? Should we do this in 1.0 as well?
>
> At the moment, I can't imagine any situation where it would harm to
> call reset() upond form creation (or why it would be reasonable to
> rely that it is *not* called). So, I would vote to include it in
> 1.0 as well....
>
> cu...
> --
> ...roland huss
> consol.de
>
