|  | Commit message (Collapse) | Author | Lines | 
|---|
|  | Co-authored-by: MarkKoz <[email protected]> | 
|  | This reverts commit ad90978f | 
|  | The parsing may take up to a few hundred ms depending on the amount
of work it has to do | 
|  | The strainer now forces the text attribute to be None, simplifying
the check on strings and falls back to the superclass' method on non
string elements | 
|  | The html we parse frequently ends up with trailing and sometimes leading
newlines which get stripped out by discord anyway, we have no reason
to keep those around when sending the Markdown over to redis | 
|  |  | 
|  | The symbol is also no longer sent back to the user, as it is not
necessary and we can skip the cleanup on it | 
|  | This allows the caller to work with the message further | 
|  | Co-authored-by: MarkKoz <[email protected]> | 
|  | With only two strings, the addition is a bit clearer than
constructing and joining a tuple
Co-authored-by: MarkKoz <[email protected]> | 
|  |  | 
|  | Co-authored-by: MarkKoz <[email protected]> | 
|  |  | 
|  | Co-authored-by: MarkKoz <[email protected]> | 
|  | Co-authored-by: MarkKoz <[email protected]> | 
|  | The else is a bit clearer than the early return | 
|  | All commands that refresh the inventories in some way are now locked to
prevent various race conditions that may have occurred in the unlikely
scenario that they got triggered together, the fetching part of the
get command now also has to wait for the running inventory refresh to
finish before proceeding to fetch and parse the html | 
|  | discord.py uses the globals of functions to resolve forward refs
in commands, previously decorators applied before commands
broke the bot with forwardrefs to names that weren't in the namespace
of the module where they were defined, the new function takes care of
merging the globals in a new function to mitigate this issue.
closes: #1323 | 
|  | While technically correct, always sending success could be misleading
in case of a typo on the package | 
|  | The finally will make sure we reset the task and log it no matter
what happens, additionally the clearing of the variable is now only
done in one place as the finally also executes when the coro is
cancelled | 
|  | Previously in case get_markdown for an item ran twice, the one
that ran second would overwrite the future created by the first one,
potentially causing the coro to wait for it infinitely as _parse_queue
would only be able to set the last future | 
|  | We no longer need to keep the items around since everything is in redis
and the costs of always going through redis is fairly small | 
|  |  | 
|  | Previously we used packages as the top level keys and fields
contained the url and the symbol id, however if we want to store
all symbols from fetched pages instead of only the ones that were
fetched by the users this comes worse off than using the page url
in the field and setting EXPIREs for them instead of doing it manually
in python.
The new implementation uses package:url as the redis key and only
the symbol id for field names, with the expire being set to a week
on the key, this means we have to pattern match the keys when deleting
the cache for a package but that's being done far less than the expire
checking done previously. | 
|  | GitHub Actions has an implicit status condition, `success()`, that
is added whenever an `if` condition lacks a status function check
of its own. In this case, while the upload step did check for the
outcome of the previous "always" step, it did not have an actual
status check and, thus, only ran on success.
Since we always want to upload the artifact, even if other steps
failed, I've added the "always" status function now. | 
|  | This commit introduces the same custom status embed as is already being
used for Sir Lancebot. The default embeds GitHub sends are disabled, as
they were causing slight issues with rate limits from time to time.
It works like this:
- The Lint & Test workflow stores an artifact with PR information, if we
  are linting/testing a PR.
- Whenever we reach the end of a workflow run sequence, a status embed
  is send with the conclusion status.
Signed-off-by: Sebastiaan Zeeff <[email protected]> | 
|  |  | 
|  | There wasn't any reason the command needed its own cog, so the exact
same functionality is now in the Verification cog. | 
|  | `!verify` command allows moderators to apply the Developer role to a
user. `!verify` is therefore removed as an alias for `!accept`. | 
|  | The result of _split_parameters is only iterated over, so a list is not
needed. Making it lazy may also save some time in cases where we don't
use all parameters | 
|  | The two variables were initialized and cleared together and contained
related information | 
|  | Previously the code assumed ' and " can be used interchangeably,
and strings that were inside of brackets were ignored for depth but
their contents weren't causing strings like "ab[cd" to increase the
depth | 
|  |  | 
|  | "Superstarified!"
Per internal staff discussion, we'll keep the wording of the message. | 
|  |  | 
|  | Whereas one of my previous commits makes the message the user gets and the infraction that gets recorded the same, the recorded infraction is now shorter, but the message the user gets is more similar to the embed posted in the public channel. We also softened the language of the user-facing message a bit. |