* Add files via upload
* Update predicates.py
Changed sender from a discord.Member object to ctx.
Added a channel check.
Combined the same method and channel method into a validator and applied through-out.
valid_role and has_role methods now check for either an id or a name.
contained now uses string.lower() when testing for membership in a collection.
Signed-off-by: Redjumpman <redjumpman@users.noreply.github.com>
API changes:
- Cogs must now inherit from `commands.Cog` (see #2151 for discussion and more details)
- All functions which are not decorators in the `redbot.core.checks` module are now deprecated in favour of their counterparts in `redbot.core.utils.mod`. This is to make this module more consistent and end the confusing naming convention.
- `redbot.core.checks.check_overrides` function is now gone, overrideable checks can now be created with the `@commands.permissions_check` decorator
- Command, Group, Cog and Context have some new attributes and methods, but they are for internal use so shouldn't concern cog creators (unless they're making a permissions cog!).
- `__permissions_check_before` and `__permissions_check_after` have been replaced: A cog method named `__permissions_hook` will be evaluated as permissions hooks in the same way `__permissions_check_before` previously was. Permissions hooks can also be added/removed/verified through the new `*_permissions_hook()` methods on the bot object, and they will be verified even when permissions is unloaded.
- New utility method `redbot.core.utils.chat_formatting.humanize_list`
- New dependency [`schema`](https://github.com/keleshev/schema)
User-facing changes:
- When a `@bot_has_permissions` check fails, the bot will respond saying what permissions were actually missing.
- All YAML-related `[p]permissions` subcommands now reside under the `[p]permissions acl` sub-group (tbh I still think the whole cog has too many top-level commands)
- The YAML schema for these commands has been changed
- A rule cannot be set as allow and deny at the same time (previously this would just default to allow)
Documentation:
- New documentation for `redbot.core.commands.requires` and `redbot.core.checks` modules
- Renewed documentation for the permissions cog
- `sphinx.ext.doctest` is now enabled
Note: standard discord.py checks will still behave exactly the same way, in fact they are checked before `Requires` is looked at, so they are not overrideable.
Signed-off-by: Toby Harradine <tobyharradine@gmail.com>
Use `async with Lock` instead of deprecated `with await lock` usage.
Forces a file fsync prior and a directory fsync (where available) after rename to prevent issues with left behind temp files.
Also should clarify: this is not threadsafe. Comments were clarified, function names remain misleading.
What's changed:
- Fixed issues mentioned on #2031
- Fuzzy help displays more like help manual
- Fuzzy help is easier and more flexible to use
- Fuzzy help string-matching ratio lowered to 80
- Help formatter is more extendable
- Help command has been optimized, cleaned up and better incorporates fuzzy help
- Added async_filter and async_enumerate utility functions because I was using them for this PR, then no longer needed them, but decided they might be useful anyway.
- Added `Context.me` property which is a shortcut to `Context.guild.me` or `Context.bot.user`, depending on the channel type.
* more filters
* note to future people touching this
* chan mentions were almost missed...
* Swap strategies as the previous escaped the mention, while still pinging the user/role
* [Trivia] Move trivia lists back home
Removes red-trivia as a dependency.
Signed-off-by: Toby Harradine <tobyharradine@gmail.com>
* Include package data in distribution
Signed-off-by: Toby Harradine <tobyharradine@gmail.com>
* Add test from red-trivia repo, and fix package data setup
* The distribution will now include all files under any data/ sub-directory of a package, as well as all *.po files under any locales/ sub-directory (as it should have been before).
* MANIFEST.in has been simplified to comply with these changes and redbot/cogs/audio/application.yml has been moved to the data/ sub-directory to maintain consistency in how we declare package data.
Signed-off-by: Toby Harradine <tobyharradine@gmail.com>
* [i18n] Use redgettext over pygettext
* Clear out autogenerated `messages.pot` files
* Remove redundant `regen_messages.py` files
* Refactor `generate_strings.py` to use redgettext
* Install redgettext in Travis Crowdin job
* Clean up some problematic usages of gettext function
* Reformat
* Replace generate_strings.py with Makefile argument
* Update to redgettext 2.1, use exclusion pattern
Signed-off-by: Toby Harradine <tobyharradine@gmail.com>
* Make calling groups useful
This makes config.Group.__call__ effectively an alias for Group.all(),
with the added bonus of becoming a context manager.
get_raw has been updated as well to reflect the new behaviour of
__call__.
* Fix unintended side-effects of new behaviour
* Add tests
* Add test for get_raw mixing in defaults
* Another cleanup for relying on old behaviour internally
* Fix bank relying on old behaviour
* Reformat
* Add output sanitization defaults to context.send
Add some common regex filters in redbot.core.utils.common_filters
Add a wrapper for ease of use in bot.send_filtered
Sanitize ModLog Case's user field (other's considered trusted as moderator input)
Sanitize Usernames/Nicks in userinfo command.
Santize Usernames in closing of tunnels.
* Add documentation
* Handle missing cogs correctly, add some helpful algorithms
For cog loading, only show "cog not found" if the module in question was the one
that failed to import. ImportErrors within cogs will show an error as they should.
- deduplicator, benchmarked to be the fastest
- bounded gather and bounded async as_completed
- tests for all additions
* Requested changes + wrap as_completed instead
So I went source diving and realized as_completed works the way I want it to,
and I don't need to reinvent the wheel for cancelling tasks that remain
if the generator is `break`ed out of. So there's that.
* [V3] Stop `tmp` dir showing up
* [V3] Remove requirements.txt and declare in install_requires
* Remove requirements.txt from tox.ini
* Update and pin all dependencies and sub-dependencies
* Update for breaking changes
* Reformat
* Update docs/requirements.txt and tox.ini requirements
* Add 3.7 to identifiers and travis/tox builds
* Attempt at fixing the travis build matrix
* Attempt #2
* Attempt 3
* aiohttp.ClientSession.close() -> detach() in sync code
* Add raven-aiohttp to requirements
* Fix stuff in setup.py
- Added discord.py back into requirements list
- Fix typo in alabaster extra requirement
Also in the Pipfile:
- Removed allow_prereleases and explicitly pinned black, since this is the only dep we want a prerelease for.
* Update to Rapptz/discord.py@8ccb98d395
* Add proper 3.7 build in Travis
See travis-ci/travis-ci#9815
* Which version of 3.6 does Xenial install then?
* Maybe we should stop pipenv installing useless stuff
* Nevermind, back to specific minor version
* Remove lots of WET dependency stuff
* Fix egg fragment for dependency link
* enforce commands as ours
* clearer user feedback
* No more 'one more tweak' commits without verifying anyway
* more detailed error with docs link + docs update
Resolves#1972.
* [CogManager] Correctly manage core and install paths
This keeps the core and install paths separate from those set in the
config.
It also displays the core path separately in `[p]paths`
Resolves#1982.
* [CogManager] Fix old reference to method removed in previous commit
Also did a bit of a general cleanup of cog_manager.py's code
* Make the core path a class attribute
* [CogManager] Paths should default to a list
Fixes#1943
This just skips cogs when `__module__` is None. Also:
- backported rapptz/discord.py#621
- moved RPC handler unregister to remove_cog()
- raise an exception when a load would overwrite an existing extension
* basic caching layer
* bit more work, now with an upper size to the cache
* cache fix
* smarter cache invalidation
* One more cache case
* Put in a bare skeleton of something else still needed
* more logic handling improvements
* more work, still not finished
* mass-resolve is done in theory, but needs testing
* small bugfixin + comments
* add note about before/after hooks
* LRU-dict fix
* when making comments about optimizations, provide historical context
* fmt pass
