| Commit message (Collapse) | Author | Lines |
|
It now explains that new users can only see a limited amount of public
channels, and that there will be more once they verify.
Co-authored-by: Sebastiaan Zeeff <[email protected]>
|
|
|
|
|
|
|
|
We will begin pinging users in the verification channel, prompting
them to join. This can cause a surge of activity that may trigger
the filter.
A better solution would involve allowing per-filter channel config,
but after internal discussion this is seen as unnecessary for now.
|
|
|
|
|
|
|
|
This reverts commit 59c62162
|
|
If a RedisCache instance was being accessed before bot has created
the `redis_cache` instance, the `_redis` alias was being set to
None, causing AttributeErrors in lookups.
See: #1090
|
|
|
|
We're making good use of d.py's tasks framework. RedisCache is used to
persist the reminder message ids, which can conveniently be converted
into timestamps.
It is therefore trivial to determine the time to sleep before the first
ping. After that, the bot simply pings every n hours.
|
|
Using the channel's category isn't reliable since it may take Discord a
while to actually move the channel once it's received a request from the
bot. I suppose using redis technically has the same problem, but it
should be much faster and less susceptible to lag than Discord.
Fixes #1074
|
|
The code is identical to the else block and there's no reason for
help commands to have an explicit check.
|
|
Commands, cogs, etc. created via internal eval won't have a source file
associated with them, making source retrieval impossible.
Fixes #1083
Fixes BOT-7K
|
|
When cogs reload, they used new Scheduler instances, which aren't aware
of previously scheduled tasks. This led to duplicate scheduled tasks
when cogs re-scheduled tasks upon initialisation.
Fixes #1080
Fixes BOT-7H
|
|
|
|
|
|
|
|
|
|
Allows referencing the constants within the message bodies.
|
|
|
|
Turns out that it's necessary to cancel the task manually. Otherwise,
duplicate tasks can be running concurrently should the extension
be reloaded.
|
|
Cog is getting large so let's allow collapsing related bits.
|
|
|
|
Let's only use this function to check on the guild status. It can be
exposed via a command in the future.
Name adjusted to be more accurate w.r.t. Discord terminology.
|
|
This will be used to guard the call to `_kick_members`.
|
|
|
|
|
|
See docstring for details. The coroutine will be registered as a task
at a later point.
|
|
|
|
Let's access these via the qualified name. The amount of imported names
was starting to get unwieldy.
|
|
|
|
This reverts commit be14db91b1c70993773e67cfa663fef0cfa85666.
|
|
This will allow mods to use '!whitelist get guild_invites'
in addition to '!whitelist get guild_invite'
This is just a naive implementation which works if the plural
form is a simple s at the end of the word. It's implemented
into the converter.
|
|
Previously, this would not provide any feedback at all, which is really
terrible UX. Sorry about that.
This also adds error handling in case the API call fails.
|
|
|
|
|
|
|
|
Thanks @Den4200!
|
|
|
|
|
|
Let's give it a better name so that it's clear when this message is
sent. The initial words are adjusted to avoid repetition after the
on join message.
|
|
This message will be sent via direct message to each user who joins
the guild.
|
|
This adds a little bit of logic to the Help Channel `init_available`
coroutine, which runs when the cog loads. This ensures that if there are
more help channels in available than there should be, we remove the
superfluos ones.
Previously, if the bot started with too many channels, it would maintain
and defend that excessive amount. This is because we never actually
count the number of channels before adding in new available channels
whenever one disappears.
If we ever get too many available channels in the future, this can be
solved by simply reloading this cog.
|
|
requested by lemon
|
|
|
|
|
|
|
|
|