On Wed, May 13, 2015 at 12:25:25 +0100, Michael Rogers wrote: > A few quick comments: > > 1. You're calling a blocking method from the UI thread. Don't do that. > Instead, call the blocking method on a background thread and post the > results back to the UI thread when it returns. That's how the code I > posted yesterday works. If you want to collect all the results before > returning them, modify the listener interface to accept a list of > clients, and collect the results on a background thread. You can use a > CountDownLatch to wait for all the results to be returned.
Note that this is intentional and mentioned in the docs. A CountDownLatch works much like calling invokeAll and iterating over the futures, right? With the only difference that you're suggesting I do that on a background thread. The whole point of the blocking method is for it to be easier to use for very basic use cases. If you don't want to block the UI thread, either call this method on a separate thread or use the other method, which doesn't wait for the results to be ready. Perhaps I'm not using the best coding/interface practice for this, but I think the overall idea is not wrong. > 2. You have two methods for getting the reachable clients, but you're > only calling one of them. If you're worried about verbosity, get rid of > the unused method and interface. :-) This is a library with an example app to show and test what it can do. I don't think it is required that the example app use all the library methods :) > 3. You're declaring stuff final all over the place, which suggests you > don't know which things need to be final. Try to work out which things > need to be final and why, it will give you a better understanding of > your code. It's obvious that I'm no java expert. As far as I know final is useful for stuff like attributes so that simple methods can be memoized. The rest I simply don't know and couldn't find proper docs explaining why or why not use final in each case, so I just went with the easy route of using it by default. Even though I'm not fond of this kind of verbosity - I wish that the compiler did these kinds of optimisations without me having to use "final". So suggestions welcome. -- Daniel Martí - mv...@mvdan.cc - http://mvdan.cc/ PGP: A9DA 13CD F7A1 4ACD D3DE E530 F4CA FFDB 4348 041C
pgp8tHwtXlrGA.pgp
Description: PGP signature
_______________________________________________ List info: https://lists.mayfirst.org/mailman/listinfo/guardian-dev To unsubscribe, email: guardian-dev-unsubscr...@lists.mayfirst.org
