From: Michelle S O. <ms...@ny...> - 2008-08-03 03:00:16
|
That is weird, now that you point it out. The argument types to 'put' are bound at compile-time but 'get' seems to ignore generics altogether.. anyone know why that is? Michelle ----- Original Message ----- From: Paul Munson Bethe <pm...@ny...> Date: Saturday, August 2, 2008 10:20 pm Subject: Re: GitFileStatus changes To: Michelle S Osborne <ms...@ny...> Cc: git...@li... > What is interesting about this is the following: > > Map<K,T>.get takes type Object, not K... > > So ResourceLabelProvider.java:86 - I would have expected a compiler > error, since I changed resource.status to return GitResourceStatus - > but no, both compatible with Object... thus the problem. > > Fixed, and checked in now. > > -Paul > > > ----- Original Message ----- > From: Michelle S Osborne <ms...@ny...> > Date: Saturday, August 2, 2008 6:58 pm > Subject: Re: GitFileStatus changes > To: pm...@ny... > Cc: git...@li... > > > > Okay, I guess that makes sense for now. I'm still not sure that the > > > different uses of this enum are ultimately going to represent the > same > > abstraction, but we can always separate them again if that becomes > > apparent. It would have been nice to say something about making the > > > change though, given our discussion in the code review. > > > > The broken functionality is my fault too, because there really > should > > be tests. Once commit and add are release-ready I'm going to focus > > solely on that. We should be able to refactor with confidence. > > > > Michelle > > > > ----- Original Message ----- > > From: pm...@ny... > > Date: Saturday, August 2, 2008 6:31 pm > > Subject: Re: GitFileStatus changes > > To: Michelle S Osborne <ms...@ny...> > > Cc: git...@li... > > > > > Sorry. I thought I tested that, but I guess not... I changed it > > > > because I needed to know about Untracked files, and it didn't make > > > > sense to extend the functionality when there was the similar class. > > > > > > I will fix this tomorrow - I had to go out of town last > night/today, > > > > > and I forgot to make a code review. > > > > > > Also, I think that failing test is my fault. Generally me messing > > > > things up here :( > > > > > > I will sort it all out tomorrow. > > > > > > -Paul > > > > > > ------Original Message------ > > > From: Michelle S Osborne > > > Sender: > > > To: pm...@ny... > > > Cc: git...@li... > > > Sent: Aug 2, 2008 5:37 PM > > > Subject: GitFileStatus changes > > > > > > Paul, > > > > > > Why did you change all uses of GitFileStatus to GitResourceStatus? > > > You > > > asked in a code review if there was a reason why I didn't use it > and > > I > > > answered you. If you weren't satisfied with the answer you could > > have > > > discussed it further rather than taking it upon yourself to > replace > > > > > uses of one with the other. You didn't even create a code review > for > > > > > that changeset, so I didn't get a chance to give you feedback, > like > > > > > for example to tell you that your changes broke the commit > dialog's > > > > > status display. > > > > > > I may add more fields to that enum. I don't have the time to > change > > > > > all of the commit dialog's uses of GitResourceStatus back to > > > GitFileStatus, so if I do need to add fields I'm going to make the > > > > getParentTransition method throw exceptions for them, because I > > doubt > > > that the method will make sense in the context of how I'm planning > > > on > > > using the enum. > > > > > > Michelle > > > > > > > > > Sent via BlackBerry |