Ian: "optional parameters and their resolution/scoring"
Remo: Scoring has to be reworked anyway
Ian: any ideas you what you would want to do in a rework?
we have all thought about it. default ctor, no satisfiable ctors, ctors with default params that can't be otherwise satisfied, ctors with default params that can be fully or partially satisfied
what happens if you have two ctors of equal params. but one uses default and one can be injected, which do we prefer, or do we error and make them disambiguate? -and supplied parameters.
Joe: And ctors that can be satisifed with implicit/missing bindings.
Ian: from this thread: http://groups.google.com/group/ninject/browse_thread/thread/e725483345594bed/db3a2d18a407f0b2?lnk=gst&q=Let.Inject#db3a2d18a407f0b2
any thoughts joe?
Joe: reading...
Remo: I think default parameters have no influence
At least no other than self bindable types
Ian: given that they are compile time constants that are not self bindable
Remo: Can you have default params that are not constant?
Joe: No.
value types are constants, references types have to be null (IIRC)
Remo: Ok in that case default value => satisfied parameter, conts the same as every other parameter
so no change needed
But I think we still have poblems with the current implementation
Ian: but we have to detect them iirc
Remo: I rather think we should change the binding resolvers and askt them if they have a binding
rather than chacking for all this stuff in the heuristic checking
Ian: "optional parameters and their resolution/scoring"
Remo: Scoring has to be reworked anyway
Ian: any ideas you what you would want to do in a rework?
we have all thought about it. default ctor, no satisfiable ctors, ctors with default params that can't be otherwise satisfied, ctors with default params that can be fully or partially satisfied
what happens if you have two ctors of equal params. but one uses default and one can be injected, which do we prefer, or do we error and make them disambiguate? -and supplied parameters.
Joe: And ctors that can be satisifed with implicit/missing bindings.
Ian: from this thread: http://groups.google.com/group/ninject/browse_thread/thread/e725483345594bed/db3a2d18a407f0b2?lnk=gst&q=Let.Inject#db3a2d18a407f0b2
any thoughts joe?
Joe: reading...
Remo: I think default parameters have no influence
At least no other than self bindable types
Ian: given that they are compile time constants that are not self bindable
Remo: Can you have default params that are not constant?
Joe: No.
value types are constants, references types have to be null (IIRC)
Remo: Ok in that case default value => satisfied parameter, conts the same as every other parameter
so no change needed
But I think we still have poblems with the current implementation
Ian: but we have to detect them iirc
Remo: I rather think we should change the binding resolvers and askt them if they have a binding
rather than chacking for all this stuff in the heuristic checking