`conf.supybot.databases()` may be called without any plugin supporting
sqlite3 being loaded yet, which causes `sqlite3` to be missing from
`sys.modules`; so it wouldn't be used by plugins loaded afterward.
`registry.Value.__call__()` is a wrapper around access to
`registry.Value.value`, that checks if the value was set before the latest
call to `registry.open_registry`; and updates the `value` if needed.
When accessing `registry.Value.value` directly, this cache can't be
invalidated, causing the old value to still be used, until the next call
to `registry.Value.__call__()`.
This commit reverts db7ef3f02517f9f2a3c56829a22b9fad3c36e374
(though it keeps the year updates)
After discussion with several people, it seems better to mention
copyright owners explicitly. eg. https://reuse.software/faq/#vcs-copyright
explains the issue of using VCSs to track copyright.
As db7ef3f02517f9f2a3c56829a22b9fad3c36e374 only replaced mentions
of my name with 'The Limnoria Contributors', this commit only needs
to undo that + add one person who contributed to setup.py.
* Fix joins to many channels
If you have enough channels that the 512 byte message limit on the JOIN
message is hit then limnoria was losing the channel that put it over the
limit and not including it in the next JOIN message. This resulted in
losing one channel for every JOIN message that pushed us over 512 bytes.
We fix this by generating the JOIN message immediately after resetting
the channels list to ensure we include the channel that pushed us over
the limit. Then the next time through our JOIN msg construction we'll
add subsequent channels without forgetting the one that pushed us over.
* Add test for channel join lists
This adds a test for the issue that is fixed in the previous commit. We
ensure that when JOINs are split over multiple messages we JOIN to all
channels that were part of the input list and don't forget any of them.
Otherwise, if some IP addresses don't work (eg. all odd ones), the bot will
consecutively fail because it can't connect, then connect + get STS + reconnect,
then fail again, then connect + get STS, etc.
If on an insecure connection: reconnect.
If on a secure connect: store it and do nothing else.
For now, stored STS policies are not read when connecting to an
insecure server.
There's no reason to use it anymore instead of Socket.
It's already missing features compared to Socket, and I don't want to
maintain it anymore so it will keep getting worse.
To prevent infinite recursion on startup, because defaultHttpHeaders
gets the value; which is unset, so 'registry._lastModified > self._lastModified'
is True, so .set() is called, so .setValue() is called again.
03c4529d9e3f6ce13c288fa1aa4bfb933288f27d made irc.state.history network specific, so reconnecting each network separately will update its history buffer's max length.
