Ed Kellett

I'm not certain whether #191 is really a bug, but setting that aside, I don't think we should reveal iline-spoofed IPs here.

I think that'd be ideal, yeah.

Hmm. It'd be nice if setting umode -R when you're _already_ -R would mark R as a changed umode, but clearly harder to do that with how user_mode works now.

This currently sets people -Z when the default umodes change

I'm generally in favour but the impression I get from the code is that people have tried quite hard to fix UID leaks, and I'd like to know if there's...

I'm not thrilled about making things netwide snotes that are propagated globally anyway.

No, but making them netwide feels kind of bad because connects are already propagated globally

unless it's substantially more than just connection class that we need to propagate I'm inclined to say we should do so. other things that report classes probably deserve to be...

Thanks for the PR. So there's two reasons I suggested changing how `get_channel_access` is called. The first is that it's a step on the way to allowing override to work...

I think `ON servername` is the best way. While I don't think it should matter which server you're on, it's occasionally nice to have a shortcut for the local server.