MagmaCollection API

[Chris Muller]

Before, #read:at: that was on MagmaCollection and returned a reader;

 	| reader |
 	reader _ cases read: #id at: id.
 	reader size > 1 ifTrue: [self error: 'Multiple objects with the same
 UUID!'].
 	^reader size = 1 ifTrue: [reader at: 1] ifFalse: [nil]

so you still had to do the size checking, nothing has changed w.r.t.
that.  The only difference is now we have just one #where: method for
all querying rather than multiple methods.  Everyone is concerned with
code bloat, so I think #where: handling all cases was addition by
subtraction.

The other thing to watch out for with the above code is, because the
reader is based on a collection that is always changing, its really
best to use #at:ifOutOfBounds: since, theoretically, another user could
remove objects causing the size to change.

> PS. I am now experimenting with a shared magma session in combination
> with one per user - I drowned in too much materialization otherwise.

By "otherwise", do you mean the strict 1:1?  And even after employing
correct read-strategys?

thanks..



--- goran at krampe.se wrote:

> Howdy all, and Chris of course!
> 
> Ok, let me show you a method in Gjallar:
> 
> findObjectById: id
> 	"Find any object by id. Return nil if not found."
> 
> 	| reader |
> 	reader _ cases where: [:each | each read: #id at: id ].
> 	reader size > 1 ifTrue: [self error: 'Multiple objects with the same
> UUID!'].
> 	^reader size = 1 ifTrue: [reader at: 1] ifFalse: [nil]
> 
> So "cases" is a MagmaCollection with an index #id. Now... what is the
> preferred simplest way of doing the #read:at: ? Nowadays this method
> is
> not on MagmaCollection but on the reader. So.... the above looks
> needlessly complex IMHO - or is there a point I am missing?
> 
> regards, Göran
> 
> PS. I am now experimenting with a shared magma session in combination
> with one per user - I drowned in too much materialization otherwise.
> _______________________________________________
> Magma mailing list
> Magma at lists.squeakfoundation.org
> http://lists.squeakfoundation.org/mailman/listinfo/magma
> 
>