Mastodon expects remote servers to remove follow relationships upon receiving
a Block. However, the spec only evokes Block activities in a C2S context, never
in a S2S context.
This PR, in addition to federating the Block, explicitly sends a Reject for any
affected follow relationship, which makes a bit more sense with regards to the
spec.
* Add Tombstone model to remember object deletion
* Do not recreate a status if it has been deleted
* Record Tombstone for remote deleted items
Also, only record deleted items from same-host actors
* Clear an user's tombstones when their key change
Currently, middle-clicking on a hashtag search result will open a new
instance of the WebUI, which is inconsistent with middle-clicking on
an account result, or a hashtag in a toot.
* Refactor signature verification a bit
* Rescue signature verification if recorded public key is invalid
Fixes#8822
* Always re-fetch AP signing key when HTTP Signature verification fails
But when the account is not marked as stale, avoid fetching collections and
media, and avoid webfinger round-trip.
* Apply stoplight to key/account update as well as initial key retrieval
* Allow to show a specific status in the admin interface
* Let the front-end know the current account is a moderator
* Add admin links to status and account menus
If the current logged-in user is an admin, add quick links to the admin
interface in account and toot dropdown menu. Suggestion by @ashkitten
* Use @statuses.first instead of @statuses[0]
* Ensure blocked user unfollows blocker if Block/Undo Block are processed out of order
* Add specs for Block causing unfollow and for out-of-order Block + Undo
* Do not LDS-sign Follow, Accept, Reject, Undo, Block
* Do not use LDS for Create activities of private toots
* Minor cleanup
* Ignore unsigned activities instead of misattributing them
* Use status.distributable? instead of querying visibility directly
The modal-handling code gives back focus to the element focused when the
modal opened. However, in the case of reply confirmation, it would do so
*after* the composer code itself requested focus.
* Add failing test for windows-1251 link cards
* Ignore low-confidence CharlockHolmes guesses
Fixes#9466
* Fix no method error when charlock holmes cannot detect charset
There is no reason to disable the composer textarea when some media metadata
is being modified, nor is there any reason to focus the textarea when some
media metadata has been modified (prevents clicking one image's description
field right after having modified another).
* Do not animate account header art if user's GIF autoplay setting is off
Fixes#9472
* Honour currently logged-in user's GIF autoplay setting for account avatars
Fixes#9467
* Fix avatar display on some public pages when data is stored on a different host
* Use ternary operator instead of if/else for avatar/header URL
The `hasMore` property of timelines in redux store was set whenever an API
request returned only one page of results, *even* if the query only requested
newer statuses (using `since_id`), causing `hasMore` to be incorrectly set to
false whenever fetching new toots in a timeline, which happens each time
an account's timeline or media gallery is visited.
* Always use the scrollable class for the ScrollList component
Fixes#9499 as well as empty timelines' scrollbar “blinking”
between loading state and empty state.
* Do not display empty message when the list is known to have more elements
Fixes#9500
* Fix LoadMore usage in account media gallery
(Even though that codepath is probably never actually used…)
* Make account media gallery more consistent with account timeline
Fixes#9498
Display “load more” more consistently, add a loading indicator on first load.
* Fix “load more” button when no data has been fetched
* Add instance-wide setting to disable profile directory
Fixes#9496
When the profile directory is disabled:
- The “discoverable” setting is hidden from users
- The “profile directory” link is not shown on public pages
- /explore returns 404
* Move Setting.profile_directory check to a before_action filter
* Add setting to not aggregate reblogs
Fixes#9222
* Handle cases where user is nil in add_to_home and add_to_list
* Add hint for setting_aggregate_reblogs option
* Reword setting_aggregate_reblogs label
* Use `update_column` instead of `update_attribute` in callback
`update_attribute` would normally cause callbacks to be called.
Called from a callback, it seems to stop further callbacks from executing.
`update_column` does the same work, but without calling callbacks or
preventing other callbacks from executing.
* Fix deadlocks by moving account stats update outside transaction
Fixes#9309
This only allows hyphens in the middle of a username, much like dots,
although I don't have a compelling reason to do so other than keeping
the changes minimal.
* Do not hide boost notifications from followed people with hidden boosts
Not displaying boosts from a followed user in the Home timeline and not
having notifications when they reblog your own content are two very
separate concerns, tying them together seem counter-intuitive and unwanted.
* Update specs accordingly
e.g., typing “blob_cat_p” used to search for “blob” and “cat”, but not
“blob_cat_p”, which means “blob_cat_patpat” is very unlikely to show up,
although it is likely what the user wanted to type in the first place.
Webpacker (Ruby wrapper to webpack) uses RAILS_ENV-based environments while
the javascript configuration for webpack re-reads this configuration file using
the NODE_ENV environment variable. This means that when RAILS_ENV=test, running
“assets:precompile” compiled the production packs in “public/packs” while
webpacker expects them in “public/packs-test”. This causes Ruby to recompile
them on-the-fly, possibly leading to race conditions in parallel_tests.
This changes:
- Disables on-the-fly compilation in test environment
- Changes the javascript part to read the correct environment
* Fix uncaching worker
* Revert to using Paperclip's filesystem backend instead of fog-local
fog-local has lots of concurrency issues, causing failure to delete files,
dangling file records, and spurious errors UncacheMediaWorker
* Add “bundle clean” suggestion from CircleCI doc
Cf. https://circleci.com/docs/2.0/caching/#bundler-ruby
* Use workspaces instead of caches for ruby gems
Cache are not guaranteed to be available, while the test jobs *require* their
depencies to run. Workspaces are thus more suitable.
One downside is that the order of workspace layer additions need to be
deterministic, which is why install-ruby{2.3,2.4} now depend on
install-ruby2.5.
ThibGandEugen Rochko
