On Sep 21, 2011, at 6:09 PM, Conrad Shultz wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > On 9/21/11 2:22 PM, Charles Srstka wrote: >> Why not? I find it to be safer than checking for the existence of >> a method, because you never know if that method might have actually >> have existed before it went public, but in a less stable form or >> with different behavior. If this is the case, then just checking >> whether the method exists and calling it could produce who knows >> what behavior on older systems. > > I have always been under the impression that Apple recommends weak > linking and run-time checks. It's what's covered in the (aging) Tech > Note 2064 for OS X and, IIRC, is what is discussed in, for example, > the original iPad/iOS 3.2 release notes. > > What are some examples of breakage happening under this scenario? I'd > like to be on the lookout myself if I indeed did put too much faith in > the documentation.
One relevant case I encountered was a leak in CFFileDescriptor which occurred on 10.5 but was fixed in 10.6. Under 10.5 (when detected), I had to manually close a leaked file descriptor. Cheers, M_______________________________________________ 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]
