3.1 Blocks - Why doesn't it crash with web lock error?

Hi,

great book. Thank you. I am reading it on Readmill so I am not sure what the actual page number is but the source section I am concerned with is entitled Concurrency/PRPFirstProjectTweeter04/PRPFirstProjectTweeter/PRPViewContoller.m

Shouldn’t the [self reloadTweets] inside the block have the same thread problems (i.e. it might not be main) as is later discussed in 3.2. GCD?

Im surprised that it never (I tried 5 times) fails. According to the docs the completion handler block is not guaranteed to be called on any particular thread. Shouldn’t I also see crash reports of the form [..] Tried to obtain the web lock from a thread other than the main thread [..]? Am I just un/lucky?

Also, the view is only released when one clicks ‘Send’ rather than ‘Cancel’ but I see someone already added an errata comment for this.

Best regards,
Hans

Is it because we don’t try to edit instance members but just call loadRequest and it handles the dispatch to the main thread’s queue for us?

[OT] If I changed the view property from nonatomic to atomic: would one still be faced by synchronization issues? I expect it would break all sorts of other conventions and hurt performance but would that work?