Brian Warner

ah, ok, there are *two* connection termination events. Whoever disconnects first causes the `usage` record to be written, but the list of active connections isn't touched until the *second* disconnect...

I'd also be kind of interested in knowing when the two sides used the same externally-visible IP address. That would tell us about network conditions where the two sides are...

Hm, that's an interesting idea. There's not currently any support for that: the helper will accept connections from anybody who wants, and will glue it to the second connection that...

One more comment to migrate from the old ticket: The API I'm thinking of would be like: ```python w = wormhole.create(stuff) w.set_code(code) w.connect() # ... seed = w.get_seed() # later...

Clients should store a mailbox, not a nameplate. We only use nameplates because they're short and easy to dictate/transcribe. As such, they're short lived, and mutable (the `2-` I use...

I have vague memories of a long-ago conversation about this, and one thing that came out of it was: *programs* should perform lockdown, not modules. That is, only an entry-point...

Hello old friends! So.. I've been working on NAT hole-punching support for magic-wormhole, and I have an additional need for SO_REUSEPORT, which probably isn't supported by the linked PR, and...

Ok, so any opinions about an API? I think the contenders are: * @adiroiban 's `ReusableIPv{4,6}Address`, maybe with additional arguments to control just-ADDR or just-PORT or both, maybe leaving room...

> > * @moshez 's `TCP4ServerEndpoint()` 's extra `reusePort=` argument, in [#9101 allow TCP listeners to re-use ports across processes #994](https://github.com/twisted/twisted/pull/994) / [twisted.web.server.Request should not set Content-Type for Content-Length: 0...

Ok, so the constraints are: * other reactors that claim to implement `IReactorTCP` will implement `listenTCP/connectTCP` with exactly the existing arguments, so we can't add new arguments because we'd be...