expose UID in (cliconn,cliexit,nick_change) snotes

[jesopo]

this might not be a good idea, but i'd like some discussion!

if a bot that keeps track of which users are on the network has to reboot, it has to immediately dispose of that state because it doesn't know if users have reconnected since. this could be best-effort solved by being able to compare the UID you used to know them by with the UID they currently have.

would also maybe be nice to expose this in WHOIS

[edk0]

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 a particular technical reason for that I'm not thinking of.

[jesopo]

tried quite hard to fix UID leaks

except in nick collisions which is arguably the most well exposed

[dwfreed]

I'm not sure what the value of hiding UID is. It's exposed by SAVE, and it's not difficult to tell the UID has changed (or not) through other means (WHOIS connect time). It's not particularly meaningful data other than a shortcut to identifying that the person is the same person they were before, assuming the ircd itself wasn't restarted in the intervening time as well.

[jesopo]

tangential but i wonder if i should instead take this as an opportunity to make more bot-parse-friendly information-dense forms of snotes instead of trying to shoehorn UID in to cliconns/etc