Alex Shpilkin

By the way, it’s unclear to me how either PEP 568 or forbidding `yield` inside generators could accomodate creating nurseries inside context managers using `contextlib.asynccontextmanager`, like that [mentioned in #569](https://github.com/python-trio/trio/issues/569#issuecomment-408419260).

Реализовывать через SOCKS (даже прозрачный при [содействии][1] iptables) приятнее тем, что придётся писать только свою реализацию SOCKS (и DNS), а не свою реализацию TCP поверх tun-интерфейса. Обе известные мне реализации...

В любом случае нужно в первую очередь организовать DNS-сервер, выдающий адреса из выделенного диапазона (лучше IPv6), а потом уже можно думать, как на эти адреса получать пакеты. (Ещё одна интересная...

@larsbrinkhoff I vaguely remembered about this before I posted, but I thought that it was quite a while back, and Discussions only entered limited testing in March and general release...

@rdrop-exit: > Just imagine if processor instruction sets used long hyphenated instruction names instead of short mnemonics, rotate-through-carry-right, jump-on-not-zero, convert-word-to-extended-doubleword. It would drive me batty. :-) It’s possible to get...

Quoting Jeffrey Massung (2019-01-10 18:29:37): > A few times I've gone down this road, trying it out for fun, but kept > getting hung-up on in a couple areas: Well,...

@MitchBradley wrote: > You can't solve every problem so pick a problem that you can solve that has a worthwhile payoff, solve it, and iterate. Well, I did warn that...

(I don’t know you, so sorry if that came of as too flippant... the desire to make a cutely arranged argument overcame the effort to make a polite one.)

@MitchBradley wrote: > For some early work on Forth type checking, have a look at [Peter Knaggs' PhD thesis][PK] Thanks for the reference! Somehow I missed until now that Knaggs...

Well, first of all, a good collections library is hard. The Scheme list library (SRFI 1) took a lot of design work, and it only deals with a particular flavour...