Re: Fwd: How to cast generic collections..

Endre,

Your helper methods need to accept "Collection<? extends Helped>".  Then yo=
u
should be fine, and you don't even need to cast.

K




On Fri, Aug 29, 2008 at 5:58 AM, Endre St=F8lsvik <[email protected]> wro=
te:

> Hi.
>
> I have bumped into a problem a couple of times which I can't seem to
> quite understand how to handle best. It revolves around casting the
> type of a generic collection.
>
> My situation is something like this: An interface DiskResource, whose
> instances users get from a DiskProvider extends Provider instance.
>
> The DiskProvider has these two methods:
> releaseDiskResource(DiskResource diskResource);
> releaseDiskResources(Collection<DiskResource> diskResources);
>
> I have this helper class for implementing Providers, called Helper.
> This helper requires that the Resources implement Helped. I therefore
> make a DiskResourceImpl which implement DiskResource and Helped. I
> thus need to implement the DiskProviderImpl as such:
>
> releaseDiskResource(DiskResource diskResource) {
>  // Cast, with unchecked-warning:
>  helper.doInternalStuff((Helped) diskResource)
> }
>
> releaseDiskResources(Collection<DiskResorce> diskResources) {
>  // PROBLEM HERE, this cast is an ERROR, not a warning:
>  helper.doInternalStuff((Collection<Helped>) resourceImpls);
> }
>
> How am I supposed to do this cast? See, I know why it isn't allowed to
> cast like that, but I want to nevertheless! It seems strangely
> non-symmetric that I just cast it in the singular version (with a
> unchecked-warning), but end up with an error in the other. Is the
> latter that more dangerous than the first?
>
> Both these are "solutions":
> releaseDiskResources(Collection<DiskResorce> diskResources) {
>  Collectioc<Helped> c =3D new HashSet<Helped>();
>  for(DiskResource diskResource : diskResources) {
>    c.add((Helped) diskResource);
>  }
>  helper.doInternalStuff(c);
> }
>
> releaseDiskResources(Collection<DiskResorce> diskResources) {
>  helper.doInternalStuff((Collection) c);
> }
>
> .. of which I refuse to use #1, while #2 seems like a just-to-weird
> hack around the compiler error, replacing the error with two warnings
> (!). Raw types are the evilest there is, are they not?
>
> I hope this example points out my annoyance while not being too convolute=
d.
>
> Where is my thinking derailing? I have a feeling it is a couple of
> steps before these methods' implementation, but I've tried different
> attacks now. I don't want the users of this API to see internal stuff
> that is useless for them.
>
> Regards.
> Endre.
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> This list is hosted by DevelopMentor(R)  http://www.develop.com
>
> View archives and manage your subscription(s) at
> http://discuss.develop.com
>



--=20
Kevin Bourrillion @ Google
internal: go/javalibraries
google-collections.googlecode.com
google-guice.googlecode.com

=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
This list is hosted by DevelopMentor=AE  http://www.develop.com

View archives and manage your subscription(s) at http://discuss.develop.com