* Add hashtag filter to profiles
GET /@:username/tagged/:hashtag
GET /api/v1/accounts/:id/statuses?tagged=:hashtag
* Display featured hashtags on public profile
* Use separate model for featured tags
* Update featured hashtag counters on-write
* Limit featured tags to 10
If the first link to be verified contains a rel=me link with a SSL
error, the VerifyAccountLinksWorker will fail and not try the following
links. This rescues the SSL error when fetching the link, avoiding this
issue.
* Add test for not persisting status when attaching media to scheduled toot
* Prevent status used for validation from being persisted to the database
Fixes#9893
Thanks to tateisu for the help investigating this.
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
* Fix undefined method error in sidekiq
Body can be not nil but still be empty, which causes a
`NoMethodError: undefined method `[]' for nil:NilClass` further in the
code. This checks for an empty body to avoid the issue.
* Fix codeclimate issue
* 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
* Use Contact User as Relay, Report, Subscribe.
* Use Account.representative to fetch contact user.
* Use find_local.
* No reason to use Account.representative in subscribe_service.
* Don't required representative!
* Fallback is included in Account.representative method.
* 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
* Add REST API for creating an account
The method is available to apps with a token obtained via the client
credentials grant. It creates a user and account records, as well as
an access token for the app that initiated the request. The user is
unconfirmed, and an e-mail is sent as usual.
The method returns the access token, which the app should save for
later. The REST API is not available to users with unconfirmed
accounts, so the app must be smart to wait for the user to click a
link in their e-mail inbox.
The method is rate-limited by IP to 5 requests per 30 minutes.
* Redirect users back to app from confirmation if they were created with an app
* Add tests
* Return 403 on the method if registrations are not open
* Require agreement param to be true in the API when creating an account
* 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
- Some associations were missing from the clean-up
- Some attributes were not reset on suspension
- Skip federation and streaming deletes when purging a dead domain
- Move account association definitions to concern
* Eliminate extra accounts select query from FollowService
* Optimistically update follow state in web UI and hide loading bar
Fix#6205
* Asynchronize NotifyService in FollowService
And fix failing test
* Skip Webfinger resolve routine when called from FollowService if possible
If an account is ActivityPub, then webfinger re-resolving is not necessary
when called from FollowService. Improve options of ResolveAccountService
* Nascent tag menu on frontend
* Hook up frontend to search
* Tag intersection backend first pass
* Update yarnlock
* WIP
* Fix for tags not searching correctly
* Make radio buttons function
* Simplify radio buttons with modeOption
* Better naming
* Rearrange options
* Add all/any/none functionality on backend
* Small PR cleanup
* Move to service from scope
* Small cleanup, add proper service tests
* Don't use send with user input :D
* Set appropriate column header
* Handle auto updating timeline
* Fix up toggle function
* Use tag value correctly
* A bit more correct to use 'self' rather than 'all' in status scope
* Fix some style issues
* Fix more code style issues
* Style select dropdown more better
* Only use to_id'ed value to ensure no SQL injection
* Revamp frontend to allow for multiple selects
* Update backend / col header to account for more flexible tagging
* Update brakeman ignore
* Codeclimate suggestions
* Fix presenter tag_url
* Implement initial PR feedback
* Handle additional tag streaming
* CodeClimate tweak
* Add profile to json+ld in Accept
It's required by the ActivityPub spec
* Use headers['Content-type'] instead of mime_type
mime_type strips the profile from the content type, but it's still available raw in the headers hash
* Add test for ld+json with profile
* 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
* Add locality check to ActivityPub::FetchRemoteAccountService
Fix#8643
Because there are a few places where it is called, it is difficult
to confirm if they all previously checked it for locality. It's better
to make sure within the service.
* Remove faux-remote duplicates of local accounts
* Add silent column to mentions
* Save silent mentions in ActivityPub Create handler and optimize it
Move networking calls out of the database transaction
* Add "limited" visibility level masked as "private" in the API
Unlike DMs, limited statuses are pushed into home feeds. The access
control rules between direct and limited statuses is almost the same,
except for counter and conversation logic
* Ensure silent column is non-null, add spec
* Ensure filters don't check silent mentions for blocks/mutes
As those are "this person is also allowed to see" rather than "this
person is involved", therefore does not warrant filtering
* Clean up code
* Use Status#active_mentions to limit returned mentions
* Fix code style issues
* Use Status#active_mentions in Notification
And remove stream_entry eager-loading from Notification
* Add conversations API
* Add web UI for conversations
* Add test for conversations API
* Add tests for ConversationAccount
* Improve web UI
* Rename ConversationAccount to AccountConversation
* Remove conversations on block and mute
* Change last_status_id to be a denormalization of status_ids
* Add optimistic locking
UX-wise, people expect that saving the profile will re-check links even without changing fields content. Bug-wise, `@account` was undefined.
Regression from #8703
* Verify link ownership with rel="me"
* Add explanation about verification to UI
* Perform link verifications
* Add click-to-copy widget for verification HTML
* Redesign edit profile page
* Redesign forms
* Improve responsive design of settings pages
* Restore landing page sign-up form
* Fix typo
* Support <link> tags, add spec
* Fix links not being verified on first discovery and passive updates
* If an Update is signed with known key, skip re-following procedure
Because it means the remote actor did *not* lose their database
* Add CLI method for rotating keys
bin/tootctl accounts rotate [USERNAME]
Generates a new RSA key per account and sends out an Update activity
signed with the old key.
* Key rotation: Space out Update fan-outs every 5 minutes per 1000 accounts
* Skip suspended accounts in key rotation
* Allow accessing local private/DM messages by URL
(Provided the user pasting the URL is authorized to see the toot, obviously)
* Fix SearchServiceSpec tests
* Re-add follow recommendations API
GET /api/v1/suggestions
Removed in 8efa081f21 due to Neo4J
dependency. The algorithm uses triadic closures, takes into account
suspensions, blocks, mutes, domain blocks, excludes locked and moved
accounts, and prefers more recently updated accounts.
* Track interactions with people you don't follow
Replying to, favouriting and reblogging someone you're not following
will make them show up in follow recommendations. The interactions
have different weights:
- Replying is 1
- Favouriting is 10 (decidedly positive interaction, but private)
- Reblogging is 20
Following them, muting or blocking will remove them from the list,
obviously.
* Remove triadic closures, ensure potential friendships are trimmed
Ref: cfa9b6e13a
This breaks compatibility with pre-2.3.0 Mastodon and older
software, but at the time of writing the network is >80% above
that version.
Compatibility broken only for toots with no text.
* Send rejections to followers when user hides domain they're on
* Use account domain blocks for "authorized followers" action
Replace soft-blocking (block & unblock) behaviour with follow rejection
* Split sync and async work of account domain blocking
Do not create domain block when removing followers by domain, that
is probably unexpected from the user's perspective.
* Adjust confirmation message for domain block
* yarn manage:translations
Old statuses and statuses from Pawoo, which runs a modified version of
Mastodon, may not have been marked sensitive even if spoiler text is
present.
Such statuses are still not marked sensitve if they are local or
arrived before version upgrade. Marking recently fetched remote status
sensitive contradicts the behavior.
Considering what people expected when they authored such statuses, this
change removes the sensitivity enforcement.
Do not touch statuses_count on accounts table when mass-destroying
statuses to reduce load when removing accounts, same for
reblogs_count and favourites_count
Do not count statuses with direct visibility in statuses_count
Fix#828
* Track trending tags
- Half-life of 1 day
- Historical usage in daily buckets (last 7 days stored)
- GET /api/v1/trends
Fix#271
* Add trends to web UI
* Don't render compose form on search route, adjust search results header
* Disqualify tag from trends if it's in disallowed hashtags setting
* Count distinct accounts using tag, ignore silenced accounts
Updates account `uri` field on each call to `update_account` instead of
only once during `create_account` to mirror the same behavior in OStatus
`ResolveAccountService` class [0].
ActivityPub accounts are identified using `@username` and `@domain` pair
instead of URI since #6842.
This fixes#7479: a bug when the account identified by `@username` and
`@domain` changes its URI.
[0]:
03b69ebc45/app/services/resolve_account_service.rb (L121)
When an ActivityPub Announce is processed and the boosted toot is not known,
fetch it on behalf of one of the booster's followers. This is to allow
fetching self-boosts of previously-unknown private toots.
If fetching on behalf of a user fails, try fetching it anonymously: the
selected follower of a boosting user may be banned by the boosted toot's
author.
- POST /api/v1/push/subscription
- PUT /api/v1/push/subscription
- DELETE /api/v1/push/subscription
- New OAuth scope: "push" (required for the above methods)
Offload creation of local notifications to a worker. Remove two
redundant SQL queries from ProcessMentionsService, remove n+1
XML/JSON serialization via memoization
* No need to re-require sidekiq plugins, they are required via Gemfile
* Add derailed_benchmarks tool, no need to require TTY gems in Gemfile
* Replace ruby-oembed with FetchOEmbedService
Reduce startup by 45382 allocated objects
* Remove preloaded JSON-LD in favour of caching HTTP responses
Reduce boot RAM by about 6 MiB
* Fix tests
* Fix test suite by stubbing out JSON-LD contexts
* Remove most behaviour disparities between blocks and mutes
The only differences between block and mute should be:
- Mutes can optionally NOT affect notifications
- Mutes should not be visible to the muted
Fix#7230Fix#5713
* Do not allow boosting someone you blocked
Fix#7248
* Do not allow favouriting someone you blocked
* Fix nil error in StatusPolicy
* Add equals_or_includes_any? helper in JsonLdHelper
* Support arrays in JSON-LD type fields for actors/tags/objects.
* Spec for resolving accounts with extension types
* Style tweaks for codeclimate
* Added a timeline for Direct statuses
* Lists all Direct statuses you've sent and received
* Displayed in Getting Started
* Streaming server support for direct TL
* Changes to match other timelines in 2.0
* Add bio fields
- Fix#3211
- Fix#232
- Fix#121
* Display bio fields in web UI
* Fix output of links and missing fields
* Federate bio fields over ActivityPub as PropertyValue
* Improve how the fields are stored, add to Edit profile form
* Add rel=me to links in fields
Fix#121
to_s method of HTTP::Response keeps blocking while it receives the whole
content, no matter how it is big. This means it may waste time to receive
unacceptably large files. It may also consume memory and disk in the
process. This solves the inefficency by checking response length while
receiving.
HTTP connections must be explicitly closed in many cases, and letting
perform method close connections makes its callers less redundant and
prevent them from forgetting to close connections.
Media attachments are part of the association cache of statuses,
since they are presumed to be immutable. Unless this cache is
cleared manually, the statuses will continue to look like they
have media embedded.
* Fix#201: Account archive download
* Export actor and private key in the archive
* Optimize BackupService
- Add conversation to cached associations of status, because
somehow it was forgotten and is source of N+1 queries
- Explicitly call GC between batches of records being fetched
(Model class allocations are the worst offender)
- Stream media files into the tar in 1MB chunks
(Do not allocate media file (up to 8MB) as string into memory)
- Use #bytesize instead of #size to calculate file size for JSON
(Fix FileOverflow error)
- Segment media into subfolders by status ID because apparently
GIF-to-MP4 media are all named "media.mp4" for some reason
* Keep uniquely generated filename in Paperclip::GifTranscoder
* Ensure dumped files do not overwrite each other by maintaing directory partitions
* Give tar archives a good name
* Add scheduler to remove week-old backups
* Fix code style issue
* Add full-text search for authorized statuses
- Search API will return statuses that match the query
- Only for logged in users
- Only if you are author of the status,
- Or you were mentioned in it
- Or you favourited or reblogged it
- Configuration over `ES_ENABLED`, `ES_HOST`, `ES_PORT`, `ES_PREFIX`
- Run `rails chewy:deploy` to create & populate index
Fix#5880Fix#4293Fix#1152
* Add commented out docker-compose configuration for ES container
* Optimize index import, filter search results
* Add basic normalization to the index
* Add better stemming and normalization to the index
* Skip webfinger request if search query includes both @ and a space
* Fix code style
* Visually separate search result sections
* Fix code style issues
* When must_be_following_dm is on, only notify if recipient dm'ed user
Currently, when must_be_following_dm is on, if a user sends a direct
message replying to any status from the recipient, the recipient gets a
notification. This should not be the case, as if the recipient posted
something publicly this can be used to spam their notifications.
* Refactor replied_to_status_is_direct_message?
Following suggestion in PR
* Fix regeneration marker not being removed after completion
* Return HTTP 206 from /api/v1/timelines/home if regeneration in progress
Prioritize RegenerationWorker by putting it into default queue
* Display loading indicator and poll home timeline while it regenerates
* Add graphic to regeneration message
* Make "not found" indicator consistent with home regeneration
* Fix actors accepting invalid URI schemes or different host between URI and URL
* Fix statuses accepting invalid URI scheme or different host to actor
* Adjust tests to new requirements
* Improve readability of mismatching_origin?/invalid_origin? methods
without them, such as is the case with GNU social
Fixes the ability to find GNU social accounts via URL in search and
when using remote follow function
Avoid error when the service returns a mostly valid oembed, but has no
url in it, causing a MethodError: undefined method `url'
for #<OEmbed::Response::Photo:0x000056505def9620>
* Change conditional to avoid nil into string error in sidekiq
When obtaining information about users with mastodon in a different
subdomain, sidekiq was giving out a 'no implicit conversion of nil into String'
* Use presence instead of blank? with ternary.
Following suggestion on PR
* Add list of lists component to web UI
* Add list adding
* Add list removing
* List editor modal
* Add API account search limited by following=true relation
* Rework list editor modal
* Remove mandatory pagination of GET /api/v1/lists/:id/accounts
* Adjust search input placeholder
* Fix rspec (#5890)
* i18n: (zh-CN) Add missing translations for #5811 (#5891)
* i18n: (zh-CN) yarn manage:translations -- zh-CN
* i18n: (zh-CN) Add missing translations for #5811
* Fix some issues
- Display loading/missing state for list timelines
- Order lists alphabetically in overview
- Fix async list editor reset
- Redirect to /lists after deleting unpinned list
- Redirect to / after pinning a list
* Remove dead list columns when a list is deleted or fetch returns 404
* Add semi-support for Video/Image objects in ActivityPub
Video and Image objects will create corresponding status records
with manually crafted text contents (title + URL)
* Extract html-url-finding logic into JsonLdHelper
* Fallback to id when url missing, extract supported object types
* Avoid sending explicit Undo->Announce when original deleted
* Do not forward a reply back to the server that sent it
* Deduplicate inboxes of rebloggers' followers for delete forwarding
* Adjust test
* Fix wrong class, bad SQL, wrong variable, outdated comment
* Allow hiding of reblogs from followed users
This adds a new entry to the account menu to allow users to hide
future reblogs from a user (and then if they've done that, to show
future reblogs instead).
This does not remove or add historical reblogs from/to the user's
timeline; it only affects new statuses.
The API for this operates by sending a "reblogs" key to the follow
endpoint. If this is sent when starting a new follow, it will be
respected from the beginning of the follow relationship (even if
the follow request must be approved by the followee). If this is
sent when a follow relationship already exists, it will simply
update the existing follow relationship. As with the notification
muting, this will now return an object ({reblogs: [true|false]}) or
false for each follow relationship when requesting relationship
information for an account. This should cause few issues due to an
object being truthy in many languages, but some modifications may
need to be made in pickier languages.
Database changes: adds a show_reblogs column (default true,
non-nullable) to the follows and follow_requests tables. Because
these are non-nullable, we use the existing MigrationHelpers to
perform this change without locking those tables, although the
tables are likely to be small anyway.
Tests included.
See also <https://github.com/glitch-soc/mastodon/pull/212>.
* Rubocop fixes
* Code review changes
* Test fixes
This patchset closes#648 and resolves#3271.
* Rubocop fix
* Revert reblogs defaulting in argument, fix tests
It turns out we needed this for the same reason we needed it in muting:
if nil gets passed in somehow (most usually by an API client not passing
any value), we need to detect and handle it.
We could specify a default in the parameter and then also catch nil, but
there's no great reason to duplicate the default value.
* Add test for temporary account resolving failures in ProcessMentionsService
* Fix processing of mentions to already-known remote accounts on temporary failures
* Serialize moved accounts into REST and ActivityPub APIs
* Parse federated moved accounts from ActivityPub
* Add note about moved accounts to public profiles
* Add moved account message to web UI
* Fix code style issues
* Add structure for lists
* Add list timeline streaming API
* Add list APIs, bind list-account relation to follow relation
* Add API for adding/removing accounts from lists
* Add pagination to lists API
* Add pagination to list accounts API
* Adjust scopes for new APIs
- Creating and modifying lists merely requires "write" scope
- Fetching information about lists merely requires "read" scope
* Add test for wrong user context on list timeline
* Clean up tests
* Add a hide_notifications column to mutes
* Add muting_notifications? and a notifications argument to mute!
* block notifications in notify_service from hard muted accounts
* Add specs for how mute! interacts with muting_notifications?
* specs testing that hide_notifications in mutes actually hides notifications
* Add support for muting notifications in MuteService
* API support for muting notifications (and specs)
* Less gross passing of notifications flag
* Break out a separate mute modal with a hide-notifications checkbox.
* Convert profile header mute to use mute modal
* Satisfy eslint.
* specs for MuteService notifications params
* add trailing newlines to files for Pork :)
* Put the label for the hide notifications checkbox in a label element.
* Add a /api/v1/mutes/details route that just returns the array of mutes.
* Define a serializer for /api/v1/mutes/details
* Add more specs for the /api/v1/mutes/details endpoint
* Expose whether a mute hides notifications in the api/v1/relationships endpoint
* Show whether muted users' notifications are muted in account lists
* Allow modifying the hide_notifications of a mute with the /api/v1/accounts/:id/mute endpoint
* make the hide/unhide notifications buttons work
* satisfy eslint
* In probably dead code, replace a dispatch of muteAccount that was skipping the modal with launching the mute modal.
* fix a missing import
* add an explanatory comment to AccountInteractions
* Refactor handling of default params for muting to make code cleaner
* minor code style fixes oops
* Fixed a typo that was breaking the account mute API endpoint
* Apply white-space: nowrap to account relationships icons
* Fix code style issues
* Remove superfluous blank line
* Rename /api/v1/mutes/details -> /api/v2/mutes
* Don't serialize "account" in MuteSerializer
Doing so is somewhat unnecessary since it's always the current user's account.
* Fix wrong variable name in api/v2/mutes
* Use Toggle in place of checkbox in the mute modal.
* Make the Toggle in the mute modal look better
* Code style changes in specs and removed an extra space
* Code review suggestions from akihikodaki
Also fixed a syntax error in tests for AccountInteractions.
* Make AddHideNotificationsToMute Concurrent
It's not clear how much this will benefit instances in practice, as the
number of mutes tends to be pretty small, but this should prevent any
blocking migrations nonetheless.
* Fix up migration things
* Remove /api/v2/mutes
* Add option to block direct messages from people you don't follow
Fix#5326
* If the DM responds to a toot by recipient, allow it through
* i18n: Update Polish translation (for #5669) (#5673)
* Work around Twidere and Tootdon bug
Tootdon and Twidere construct @user@domain handles from mentions in toots based
solely on the mention text and account URI's domain without performing any
webfinger call or retrieving account info from the Mastodon server.
As a result, when a remote user has WEB_DOMAIN ≠ LOCAL_DOMAIN, Twidere and
Tootdon will construct the mention as @user@WEB_DOMAIN. Now, this will usually
resolve to the correct account (since the recommended configuration is to have
WEB_DOMAIN perform webfinger redirections to LOCAL_DOMAIN) when processing
mentions, but won't do so when displaying them (as it does not go through the
whole account resolution at that time).
This change rewrites mentions to the resolved account, so that displaying the
mentions will work.
* Use lookbehind instead of non-capturing group in MENTION_RE
Indeed, substitutions with the previous regexp would erroneously eat any
preceding whitespace, which would lead to concatenated mentions in the
previous commit.
Note that users will “lose” up to one character space per mention for their
toots, as that regexp is also used to remove the domain-part of mentioned
users for character counting purposes, and it also erroneously removed the
preceding character if it was a space.
* Resolve remote accounts when mentioned even if they are already known
This commit reduces the risk of not having up-to-date public key or protocol
information for a remote account, which is required to deliver toots
(especially direct messages).
* Do not add mentions in private messages for remote users we cannot deliver to
Mastodon does not deliver private and direct toots to OStatus users, as there
is no guarantee the remote software understands the toot's privacy. However,
users currently do not get any feedback on it (Mastodon won't attempt delivery,
but the toot will be displayed exactly the same way to the user).
This change introduces *some* feedback by not processing mentions that are
not going to be delivered. A long-term solution is still needed to have
delivery receipts or at least some better indication of what is going on, but
at least an user can see *something* is up.
This commit reduces the risk of not having up-to-date public key or protocol
information for a remote account, which is required to deliver toots
(especially direct messages).
Specifically, this fixes status length calculation to be same as JS side.
BTW, since this pattern used in not only preview card fetching, we
should extract it (with twitter-regex?) and write tests I think.
Remote ActivityPub users that have never been known as OStatus users
(whether or not they support it) will not have a “remote_url” attribute
set. In case they reside on an instance with WEB_DOMAIN ≠ LOCAL_DOMAIN,
the current check did rely on “remote_url” to verify the user's domain.
* If OEmbed response doesn't have a required property `type`, ignore it.
e.g. `NoMethodError: undefined method 'type' for ...`
* If we failed to detect encoding, fallback to default behavior of Nokogiri.
e.g. `KeyError: key not found: :encoding`
- Rename Mastodon::TimestampIds into Mastodon::Snowflake for clarity
- Skip for statuses coming from inbox, aka delivered in real-time
- Skip for statuses that claim to be from the future
* Improve error handling on LinkCrawlWorker
* Ignore TimeoutError and InvalidURIError too
* Record errors to debug log
* Enable dead job queue on LinkCrawlWorker
Since most of acceptable errors were already ignored, only our side issue should go to dead job queue.
* Ignore all http gem errors
The whole point of verified_webfinger? is to check the WebFinger-discoverable
URI maps back to the known author URI. This was not actually verified if the
first Webfinger request was not a redirection.
* Use non-serial IDs
This change makes a number of nontrivial tweaks to the data model in
Mastodon:
* All IDs are now 8 byte integers (rather than mixed 4- and 8-byte)
* IDs are now assigned as:
* Top 6 bytes: millisecond-resolution time from epoch
* Bottom 2 bytes: serial (within the millisecond) sequence number
* See /lib/tasks/db.rake's `define_timestamp_id` for details, but
note that the purpose of these changes is to make it difficult to
determine the number of objects in a table from the ID of any
object.
* The Redis sorted set used for the feed will have values used to look
up toots, rather than scores. This is almost always the same as the
existing behavior, except in the case of boosted toots. This change
was made because Redis stores scores as double-precision floats,
which cannot store the new ID format exactly. Note that this doesn't
cause problems with sorting/pagination, because ZREVRANGEBYSCORE
sorts lexicographically when scores are tied. (This will still cause
sorting issues when the ID gains a new significant digit, but that's
extraordinarily uncommon.)
Note a couple of tradeoffs have been made in this commit:
* lib/tasks/db.rake is used to enforce many/most column constraints,
because this commit seems likely to take a while to bring upstream.
Enforcing a post-migrate hook is an easier way to maintain the code
in the interim.
* Boosted toots will appear in the timeline as many times as they have
been boosted. This is a tradeoff due to the way the feed is saved in
Redis at the moment, but will be handled by a future commit.
This would effectively close Mastodon's #1059, as it is a
snowflake-like system of generating IDs. However, given how involved
the changes were simply within Mastodon, it may have unexpected
interactions with some clients, if they store IDs as doubles
(or as 4-byte integers). This was a problem that Twitter ran into with
their "snowflake" transition, particularly in JavaScript clients that
treated IDs as JS integers, rather than strings. It therefore would be
useful to test these changes at least in the web interface and popular
clients before pushing them to all users.
* Fix JavaScript interface with long IDs
Somewhat predictably, the JS interface handled IDs as numbers, which in
JS are IEEE double-precision floats. This loses some precision when
working with numbers as large as those generated by the new ID scheme,
so we instead handle them here as strings. This is relatively simple,
and doesn't appear to have caused any problems, but should definitely
be tested more thoroughly than the built-in tests. Several days of use
appear to support this working properly.
BREAKING CHANGE:
The major(!) change here is that IDs are now returned as strings by the
REST endpoints, rather than as integers. In practice, relatively few
changes were required to make the existing JS UI work with this change,
but it will likely hit API clients pretty hard: it's an entirely
different type to consume. (The one API client I tested, Tusky, handles
this with no problems, however.)
Twitter ran into this issue when introducing Snowflake IDs, and decided
to instead introduce an `id_str` field in JSON responses. I have opted
to *not* do that, and instead force all IDs to 64-bit integers
represented by strings in one go. (I believe Twitter exacerbated their
problem by rolling out the changes three times: once for statuses, once
for DMs, and once for user IDs, as well as by leaving an integer ID
value in JSON. As they said, "If you’re using the `id` field with JSON
in a Javascript-related language, there is a very high likelihood that
the integers will be silently munged by Javascript interpreters. In most
cases, this will result in behavior such as being unable to load or
delete a specific direct message, because the ID you're sending to the
API is different than the actual identifier associated with the
message." [1]) However, given that this is a significant change for API
users, alternatives or a transition time may be appropriate.
1: https://blog.twitter.com/developer/en_us/a/2011/direct-messages-going-snowflake-on-sep-30-2011.html
* Restructure feed pushes/unpushes
This was necessary because the previous behavior used Redis zset scores
to identify statuses, but those are IEEE double-precision floats, so we
can't actually use them to identify all 64-bit IDs. However, it leaves
the code in a much better state for refactoring reblog handling /
coalescing.
Feed-management code has been consolidated in FeedManager, including:
* BatchedRemoveStatusService no longer directly manipulates feed zsets
* RemoveStatusService no longer directly manipulates feed zsets
* PrecomputeFeedService has moved its logic to FeedManager#populate_feed
(PrecomputeFeedService largely made lots of calls to FeedManager, but
didn't follow the normal adding-to-feed process.)
This has the effect of unifying all of the feed push/unpush logic in
FeedManager, making it much more tractable to update it in the future.
Due to some additional checks that must be made during, for example,
batch status removals, some Redis pipelining has been removed. It does
not appear that this should cause significantly increased load, but if
necessary, some optimizations are possible in batch cases. These were
omitted in the pursuit of simplicity, but a batch_push and batch_unpush
would be possible in the future.
Tests were added to verify that pushes happen under expected conditions,
and to verify reblog behavior (both on pushing and unpushing). In the
case of unpushing, this includes testing behavior that currently leads
to confusion such as Mastodon's #2817, but this codifies that the
behavior is currently expected.
* Rubocop fixes
I could swear I made these changes already, but I must have lost them
somewhere along the line.
* Address review comments
This addresses the first two comments from review of this feature:
https://github.com/tootsuite/mastodon/pull/4801#discussion_r139336735https://github.com/tootsuite/mastodon/pull/4801#discussion_r139336931
This adds an optional argument to FeedManager#key, the subtype of feed
key to generate. It also tests to ensure that FeedManager's settings are
such that reblogs won't be tracked forever.
* Hardcode IdToBigints migration columns
This addresses a comment during review:
https://github.com/tootsuite/mastodon/pull/4801#discussion_r139337452
This means we'll need to make sure that all _id columns going forward
are bigints, but that should happen automatically in most cases.
* Additional fixes for stringified IDs in JSON
These should be the last two. These were identified using eslint to try
to identify any plain casts to JavaScript numbers. (Some such casts are
legitimate, but these were not.)
Adding the following to .eslintrc.yml will identify casts to numbers:
~~~
no-restricted-syntax:
- warn
- selector: UnaryExpression[operator='+'] > :not(Literal)
message: Avoid the use of unary +
- selector: CallExpression[callee.name='Number']
message: Casting with Number() may coerce string IDs to numbers
~~~
The remaining three casts appear legitimate: two casts to array indices,
one in a server to turn an environment variable into a number.
* Only implement timestamp IDs for Status IDs
Per discussion in #4801, this is only being merged in for Status IDs at
this point. We do this in a migration, as there is no longer use for
a post-migration hook. We keep the initialization of the timestamp_id
function as a Rake task, as it is also needed after db:schema:load (as
db/schema.rb doesn't store Postgres functions).
* Change internal streaming payloads to stringified IDs as well
This is equivalent to 591a9af356faf2d5c7e66e3ec715502796c875cd from
#5019, with an extra change for the addition to FeedManager#unpush.
* Ensure we have a status_id_seq sequence
Apparently this is not a given when specifying a custom ID function,
so now we ensure it gets created. This uses the generic version of this
function to more easily support adding additional tables with timestamp
IDs in the future, although it would be possible to cut this down to a
less generic version if necessary. It is only run during db:schema:load
or the relevant migration, so the overhead is extraordinarily minimal.
* Transition reblogs to new Redis format
This provides a one-way migration to transition old Redis reblog entries
into the new format, with a separate tracking entry for reblogs.
It is not invertible because doing so could (if timestamp IDs are used)
require a database query for each status in each users' feed, which is
likely to be a significant toll on major instances.
* Address review comments from @akihikodaki
No functional changes.
* Additional review changes
* Heredoc cleanup
* Run db:schema:load hooks for test in development
This matches the behavior in Rails'
ActiveRecord::Tasks::DatabaseTasks.each_current_configuration, which
would otherwise break `rake db:setup` in development.
It also moves some functionality out to a library, which will be a good
place to put additional related functionality in the near future.
Additionally, ActivityPub::FetchRemoteStatusService no longer parses
activities.
OStatus::Activity::Creation no longer delegates to ActivityPub because
the provided ActivityPub representations are not signed while OStatus
representations are.
If the signature could not be verified and the webfinger of the account
was last retrieved longer than the cache period, try re-resolving the
account and then attempting to verify the signature again
- Previously they wouldn't receive it unless they were author's
followers
- Skip unpush from public/hashtag timelines if status wasn't
public in the first place
* Fix JavaScript interface with long IDs
Somewhat predictably, the JS interface handled IDs as numbers, which in
JS are IEEE double-precision floats. This loses some precision when
working with numbers as large as those generated by the new ID scheme,
so we instead handle them here as strings. This is relatively simple,
and doesn't appear to have caused any problems, but should definitely
be tested more thoroughly than the built-in tests. Several days of use
appear to support this working properly.
BREAKING CHANGE:
The major(!) change here is that IDs are now returned as strings by the
REST endpoints, rather than as integers. In practice, relatively few
changes were required to make the existing JS UI work with this change,
but it will likely hit API clients pretty hard: it's an entirely
different type to consume. (The one API client I tested, Tusky, handles
this with no problems, however.)
Twitter ran into this issue when introducing Snowflake IDs, and decided
to instead introduce an `id_str` field in JSON responses. I have opted
to *not* do that, and instead force all IDs to 64-bit integers
represented by strings in one go. (I believe Twitter exacerbated their
problem by rolling out the changes three times: once for statuses, once
for DMs, and once for user IDs, as well as by leaving an integer ID
value in JSON. As they said, "If you’re using the `id` field with JSON
in a Javascript-related language, there is a very high likelihood that
the integers will be silently munged by Javascript interpreters. In most
cases, this will result in behavior such as being unable to load or
delete a specific direct message, because the ID you're sending to the
API is different than the actual identifier associated with the
message." [1]) However, given that this is a significant change for API
users, alternatives or a transition time may be appropriate.
1: https://blog.twitter.com/developer/en_us/a/2011/direct-messages-going-snowflake-on-sep-30-2011.html
* Additional fixes for stringified IDs in JSON
These should be the last two. These were identified using eslint to try
to identify any plain casts to JavaScript numbers. (Some such casts are
legitimate, but these were not.)
Adding the following to .eslintrc.yml will identify casts to numbers:
~~~
no-restricted-syntax:
- warn
- selector: UnaryExpression[operator='+'] > :not(Literal)
message: Avoid the use of unary +
- selector: CallExpression[callee.name='Number']
message: Casting with Number() may coerce string IDs to numbers
~~~
The remaining three casts appear legitimate: two casts to array indices,
one in a server to turn an environment variable into a number.
* Back out RelationshipsController Change
This was made to make a test a bit less flakey, but has nothing to
do with this branch.
* Change internal streaming payloads to stringified IDs as well
Per
https://github.com/tootsuite/mastodon/pull/5019#issuecomment-330736452
we need these changes to send deleted status IDs as strings, not
integers.
Eugen Rochko