On Dec 28, 2009, at 18:41, Graham Cox wrote: > I was wondering if fast enumeration operated on a copy internally. I'm > thinking 'no', since ...
The implementation of the fast enumeration protocol is different in every class that conforms to it, but typically it would be expected not to copy anything. However, I *think* it would be functionally just fine for an implementer of the protocol to do exactly that: copy enough of the collection storage structure to be able to keep iterating after deletions were made from the actual collection. Detection of the mutation is completely in the hands of the implementer, so it could simply not report any change. It would no longer be a "fast" enumeration, but it would be a "mutable" enumeration, with a nice concise syntax. Unfortunately, there'd be no way to have both possibilities in a single class. But maybe I haven't thought through all the issues ... _______________________________________________ Cocoa-dev mailing list ([email protected]) Please do not post admin requests or moderator comments to the list. Contact the moderators at cocoa-dev-admins(at)lists.apple.com Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com This email sent to [email protected]
