Print warning messages before cleanup runs.

[fletcherm]

Note: this code was written largely by ChatGPT 4, so it may look a little different than the surrounding code. It was, of course, still inspected by a human. :)

What appears to be working:

• A warning message is sent 5 minutes before cleanup.

What appears to be missing:

• Even though the code is setup to send more messages every minute, we're only seeing that initial 5 minute warning message. This needs further investigation.

Next steps:

• Ensure a warning message is sent every minute before cleanup.
• (Possibly) adding some tests for this. This file doesn't already have a test suite so not clear to me if the juice is worth the squeeze.
• Compare the color of the cleanup warning text to MBBS proper. Right now we're using the same red color as the final cleanup message, but not sure if this is correct.
• Compare the literal text of the final cleanup message. My spidey sense is that does not match MBBS proper.

[fletcherm]

@enusbaum @paladine This is a draft PR and not ready to merge. But I'd appreciate if you can take a look and at least see if you can identify when I'm only seeing the first warning message and not subsequent messages.

[fletcherm]

Attached is a screenshot of what the cleanup warnings look like on WG.

[fletcherm]

@enusbaum @paladine Alright, would love if you guys can take another look at this.

Only significant change since last round of reviews is that I changed how the cleanup time and grace period work, as we discussed in discord. (grace period comes after cleanup time, for whatever reason) This should better match WG.

My bigger ask is that someone test & confirm the behavior that cleanup and the cleanup warning run more than once, and not just after the initial setup. Normally that would be once a day, of course, but feel free to tweak values to make it happen more frequently.

I did do some experiments to test this myself, and things look good... but I had to hack enough lines of code to make cleanup more frequent that I kinda worry I changed something else incidentally and tainted the experiment. That is why I'd really appreciate another person testing it as well.