to me, dummy stub implementations of an interface method is a bad smell... if a dummy implementation is good enough then that method should not be on the interface. One should always have to go and implement the method in a meaningful way anyway. Generating stubs is generating smell imho.
On the other hand, I would welcome some way to make it easier to find all the places where you have to implement the new interface method so you don't have to hit compile to get all the errors. cheers, Gauti ----- Original Message ----- From: "Ian Pojman" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Thursday, April 25, 2002 5:30 PM Subject: Re: SV: [Eap-features] Better Support for Interfaces +1 on interface intention On Thursday, April 25, 2002, at 02:19 AM, Jan wrote: > +1 for navigation > Well thought out, and good arguments. Yes, I think this would be an > IDEA-'natural' implementation of interface navigation. > > +1 for the stub-implementation, though I think this one might need a > little bit more thought from people on the list, so that it doesn't end > up conflicting with different peoples coding styles. > > Regards Jan Krag > > > -----Oprindelig meddelelse----- > Fra: Dwight Rodgers [mailto:[EMAIL PROTECTED]] > Sendt: 25. april 2002 02:25 > Til: [EMAIL PROTECTED] > Emne: [Eap-features] Better Support for Interfaces > > The following would be useful for quicker manipulation of interfaces: > > 1) ctrl-clicking on a method declaration in an interface should popup a > list of implementations, and clicking on a particular one should > navigate to the method declaration in that interface > Rational: usually we can hold down ctrl and click on methods or > variables to navigate to their definitions. However, if we have an > object of an interface type, we will very likely wish to navigate to > one of the implementations. Currently, the fastest way to do this is to > navigate to the interface, then hit ctrl-H for the heirarchy, and pick > the implementation -- but at this point the memory of which method we > were viewing has been discarded, so we must hit ctrl-F12 to popup the > method list and pick the right method. Alternatively, we could pull up > the find-usages dialog, choose implementing methods, and then select > the appropriate one from the find-pane. This is not efficient enough -- > it should be possible in two-clicks using the same familiar > "ctrl-click" approach as general declaration-finding. > > 2) program-by-intent: when declaring a new method on an interface, > intellij should offer to construct an implementation of the method in > every class that implements the interface. > Rational: The code won't compile if the implementing classes don't > implement the new method. Inserting stub implementations is rote work. > Might as well be automatic. > > Logical extensions of these suggestions apply to abstract classes, > abstract methods, and to some degree overriden methods. > > Thanks, > > -Dwight > > > _______________________________________________ Eap-features mailing list [EMAIL PROTECTED] http://www.intellij.com/mailman/listinfo/eap-features
