Feels like a thing you could sugar away by having ...
# developmenta
average-vr-56795
09/29/2017, 5:07 PMFeels like a thing you could sugar away by having the and_then(|value|...) closure have a pointer to an anonymous temporary variable if you wanted, but they don't, so I need to pull out a named variable
w
witty-crayon-22786
09/29/2017, 5:27 PMMm, right. Alternatively, I think if you don't unwrap in that case what passes through the chain is the MutexGuard, which keeps the reference alive...?
witty-crayon-22786
09/29/2017, 5:28 PM Copy code
foo.and_then(|arcmutex| arcmutex.deref().lock()).and_then(|guard| do_something_with(guard.unwrap().get(|hashmap| hashmap.get(bar)))witty-crayon-22786
09/29/2017, 5:28 PM? ...but with regard to locks, i'm ok with being slightly more explicit
witty-crayon-22786
09/29/2017, 5:28 PMi've returned the guard out of functions before, and it feels like a filthy abstraction leak, heh.
a
average-vr-56795
09/29/2017, 5:29 PMguard still needs to be in scope everywhere that the ouptut from the hashmap is used, I think
average-vr-56795
09/29/2017, 5:29 PMWhich means forcing extra nesting 😞
w
witty-crayon-22786
09/29/2017, 5:32 PMah, yes. unless you copy.
a
average-vr-56795
09/29/2017, 5:33 PMAnd I'm also vaguely ok with that for Mutexes, but I've run into other cases where it was just a straight ownership thing which was kind of annoying 🙂
w
witty-crayon-22786
09/29/2017, 5:34 PMyep
a
average-vr-56795
09/29/2017, 5:34 PMI generally like these to be equivalent:
foo.bar;
let tmp = foo;
tmp.bar;
average-vr-56795
09/29/2017, 5:34 PMAnd I keep just bumping into places where they're not 🙂