Brujo Benavides

Same thing happens if there are too many changes in a single file, as you can see in https://github.com/esl/MongooseIM/pull/1259/files#diff-31008e4e10b2f4d7c420a7eaa2996e57

Not a bad idea. We'll likely accept a PR that allows numbers in atoms as long as the atom doesn't _just_ contain numbers.

Would you be so kind as to send a pull request with that change, please?

Wait a second, @sugus-seu … that queue was started with [`apns_queue:start_link`](https://github.com/inaka/apns4erl/blob/master/src/apns_connection.erl#L86), therefore it's linked to the `apns_connection` process that created it. By the time it reaches `terminate`, or soon after...

Good point, @sugus-seu.

@sugus-seu ^ I created a new PR for you, pointing to the main repo (this one) instead of yours. Your changes do not compile and @elvisci is not happy with...

@sugus-seu any updates on this?

Just in case, @uwiger … the error that travis reports is: > The code in the following (LINE, COL) locations has the same structure: (237, 3), (251, 3). It's an...

I guess we can safely ignore codecov… right, @ferigis?

What if we somehow make the retries _internal_? We can queue up send requests in, say… an ets… from which the `gen_server` pulls messages to send. Before/instead `gen_server:cast`, we insert...