i dont know if we want to go down this road. its a mixin that is not very obvious, how many people do you think know about IStyledColumn? besides it might be two objects that represent the pk - like two longs for composite pks. i really dont think it is that big a deal to implement equals/hashcode on a model, and if people have their own variants of entitymodels they only have one place to implement it. imho this is just bloat.
-igor On 8/27/07, Matej Knopp <[EMAIL PROTECTED]> wrote: > > Perhaps IEntityModel Object getEntityId/PK() would be better name? > ICompressedModel doesn't really say much, sounds too abstract to me. Also > I > could live with EqualsDecorator as long as it clearly states with CAPITAL > LETTERS that it can't be used on detachable models :) > > -Matej > > On 8/27/07, Eelco Hillenius <[EMAIL PROTECTED]> wrote: > > > > On 8/27/07, Igor Vaynberg <[EMAIL PROTECTED]> wrote: > > > the reason i did not provide this when i wrote the repeaters package > is > > that > > > it can lead down some pretty awful roads. an easy example where this > > goes > > > totally wrong is when you use detachable models for the items. if you > > use > > > this decorator all models from the previous page will be loaded > because > > they > > > will all be compared - but in case of the detachable model it is not > > > necesary to load the object, it is enough to only compare the id. > > > so while this makes it easy it also makes it easy to make something > > > inefficient because you dont have to think about it. > > > > Ok, I can follow that. If you are not using such models however, this > > is a very convenient class. I think that if we give a clear enough > > warning in the docs (and book) it should be ok. > > > > How about facilitating the case you described a little bit better as > > well? Every project probably has it's own variants of entity models by > > now. Wouldn't it be an idea to introduce an interface like > > ICompressedModel { Object getCompressed() } or something along those > > lines that people can implement and which would return e.g. the id > > which would then be used by the wrapper? > > > > Eelco > > >