safer Partyline restart handling

We recently improved the way Mediabot v3 handles IRC restarts from the Partyline.

The goal was simple:
when an admin requests a restart, the bot should restart the IRC connection cleanly without killing the whole process. In other words, no brutal die, no unnecessary process loss, and a much better experience for administration and debugging.

What was improved

• safer in-process IRC restart handling
• the Partyline stays alive while the IRC side reconnects
• protection against duplicate restart requests
• cleaner handling of channel nicklist timers
• better behavior during rehash, including preservation of console hooks
• removal of an old duplicate entrypoint that could cause confusion

Why it matters

This makes Mediabot more robust in day-to-day use.
Administrative actions are cleaner, runtime behavior is more predictable, and the bot is less likely to end up in weird states after repeated restart or rehash operations.

Result

The restart logic is now more consistent, easier to maintain, and simply safer.

A small internal improvement, maybe — but the kind that really matters when you want a long-running IRC bot to behave properly.