`Jobs::AutoJoinChannelBatch` was holding a lot of logic which should be in a service. Moreover, this refactoring is the opportunity to address a bug which could cause a duplicate key error.
From now when trying to insert a new membership it won't fail if a membership is already present.
Example error:
```
Job exception: ERROR: duplicate key value violates unique constraint "user_chat_channel_unique_memberships"
DETAIL: Key (user_id, chat_channel_id)=(1, 2) already exists.
Backtrace
rack-mini-profiler-3.1.0/lib/patches/db/pg.rb:110:in `exec'
rack-mini-profiler-3.1.0/lib/patches/db/pg.rb:110:in `async_exec'
(eval):29:in `async_exec'
mini_sql-1.4.0/lib/mini_sql/postgres/connection.rb:209:in `run'
mini_sql-1.4.0/lib/mini_sql/active_record_postgres/connection.rb:38:in `block in run'
mini_sql-1.4.0/lib/mini_sql/active_record_postgres/connection.rb:34:in `block in with_lock'
activesupport-7.0.5.1/lib/active_support/concurrency/load_interlock_aware_monitor.rb:25:in `handle_interrupt'
activesupport-7.0.5.1/lib/active_support/concurrency/load_interlock_aware_monitor.rb:25:in `block in synchronize'
activesupport-7.0.5.1/lib/active_support/concurrency/load_interlock_aware_monitor.rb:21:in `handle_interrupt'
activesupport-7.0.5.1/lib/active_support/concurrency/load_interlock_aware_monitor.rb:21:in `synchronize'
mini_sql-1.4.0/lib/mini_sql/active_record_postgres/connection.rb:34:in `with_lock'
mini_sql-1.4.0/lib/mini_sql/active_record_postgres/connection.rb:38:in `run'
mini_sql-1.4.0/lib/mini_sql/postgres/connection.rb:64:in `query_single'
/var/www/discourse/plugins/chat/app/jobs/regular/chat/auto_join_channel_batch.rb:38:in `execute'
```
Note this commit is also using main branch of `shoulda-matchers` as the gem has not been released yet.
Co-authored-by: Loïc Guitaut <5648+Flink@users.noreply.github.com>
Prior to this commit we were loading a large number of thread messages without any pagination. This commit attempts to fix this and also improves the following points:
- code sharing between channels and threads:
Attempts to reuse/share the code use in channels for threads. To make it possible part of this code has been extracted in dedicated helpers or has been improved to reduce the duplication needed.
Examples of extracted helpers:
- `stackingContextFix`: the ios hack for rendering bug when momentum scrolling is interrupted
- `scrollListToMessage`, `scrollListToTop`, `scrollListToBottom`: a series of helper to correctly scroll to a specific position in the list of messages
- better general performance of listing messages:
One of the main changes which has been made is to remove the computation of visible message during scroll, it will only happen when needed (update last read for example). This constant recomputation of `message.visible` on intersection observer event while scrolling was consuming a lot of CPU time.
Followup to f5e8e73.
This switches the placeholder label to the existing string "optional
tags" and only shows it if there are no items picked.
Co-authored-by: Jarek Radosz <jradosz@gmail.com>
I was only able to get one failure out of 100 tries, this failure didn't get me more info. My best guess ATM is that sometimes, the first session was still loading while receiving the reaction and created some unexpected situation.
The commit attempts to start the "check" session before the session making the reaction hoping that will be enough to prevent this case, if this is the issue.
Sharing a link in chat will create a onebox embed with a source that includes a site icon and title.
This change prevents loading the site icon into lightbox.
This is extracted from #22390.
This patch simplifies a little how we handle uploads in chat, relying on
ActiveRecord mechanisms instead of calling custom methods.
This also makes `Chat::Message#validate_message` a “real” AR validation,
meaning it will run automatically when `#valid?` is called.
The gjs/gts formats are a new pattern for authoring Ember components. This commit introduces support for these patterns to our build pipeline for core/plugins, and converts a handful of components to use the new format. It also introduces relevant updates to our linting config, and to our sample vscode configuration.
Co-authored-by: Godfrey Chan <godfreykfc@gmail.com>
Co-authored-by: Krystan HuffMenne <kmenne+github@gmail.com>
This only moves code around and doesn't change any behavior. This does two things:
1. Extracts the `channel.joined_by?` methods
2. Uses term "members" instead of "participants" for chat members
This is extracted from #22390.
This patch adds a new `optional` option to the `model` step. This
means if an optional model returns something blank (`nil` or an empty
collection) then the service won’t fail and will execute the next step.
However if a model is properly returned, the step will try to check if
it is valid or not (if it responds to `#invalid?`). If the model isn’t
valid, then the step will fail (so no change here).
Turns out making a html4 fragment and then operating on parts of it using html5 fragments is a bad idea. ;)
This seems to fix the issue with occasionally missing GH icons in oneboxes.
Someone who cannot chat is also not able to join chat channels,
so we may not check all the time user.can_chat? && user.can_join_chat_channel?
and just call user.can_join_chat_channel? instead.
Followup to 07c3782e51
The above incorrectly removed the channel unread count in
the mobile/drawer channel list when the user has mentions
(meaning the unreads are urgent). This commit adds it
back and refactors system specs a little.
Define new concept of panels in sidebar. Panels are wrappers around sidebar sections. In the future, it allows creating full focus mode by switching between panels.
A new API method called addSidebarPanel was added. Default main panel is already registered and by default all API sections are mounted to main.
Why this change?
The test being changed in question has been flaky on our CI. However, we
are unable to view the screenshot of why it failed because
ActionDispatch will only take a screenshot of the default session upon
failure. At the same time, taking screenshot of all sessions
automatically upon failure is not possible via the official Capybara or
Rails APIs at the moment. Therefore, we're changing this system test to
avoid using two custom session and instead have the main assertion use
the default session such that any failures will provide us with a
screenshot.
This commit fixes two issues with the thread list:
1. All threads were being shown regardless of whether the user had
a membership in the thread. This was happening because the list
and the channel share the same thread store, so if the channel
had OMs with threads we would load them and they showed in the list.
2. Threads created by the user from a staged thread would double up.
This is because the _cache in the channel threadsManager would use
the staged thread ID even after we'd replaced the object's ID with
the actual thread from the DB. The answer to this is to remove and
re-add the thread to the local cache with the actual ID.
* DEV: Fix and re-enable chat flakys
The early return in JS was added to prevent an error
from channel being null, and it's better to use known
users for the message fabrications in the specs.
* DEV: Use travel_to in drawer spec for thread tracking
Sometimes in the system test the datetime that is last
viewed for the channel for the user and the datetime for
the last message created_at is only microseconds of difference,
and we do not provide that level of fidelity in the MessageBus
serializer, so unreadThreadsCountSinceLastViewed is not
accurate.
Better to just utilize travel_to and move forward 1 minute in
time before sending the new message to easily differentiate.
When we have subscriptions for new messages in a channel,
we also have special handling for messages in a thread. For
cases like DM channels where threads are made in the background
but not used in the UI, this is causing JS errors because we
are trying to fetch the thread but it returns 404.
We only want to do things with messages in threads if the
channel actually has threading enabled.
We need a nice way to only return some hashtag data
sources based on various site settings. This commit
adds an enabled? method that every hashtag data source
must implement. If this returns false the data source
will not be used at all for hashtag lookups or search.
On tablets like iPad where we allow channel and thread to be on the same screen, it was not possible to resize the panels due to code being thought for mouse events. This commit should now correctly allow for this.
The "resizer" has also been made larger to simplify touching.
No test as it's hard to test on iPad and dragging events are also complex.
On iOS we have a hack to prevent the viewport to move when focusing an input, however this code was targeting the textarea node through a global selector which is working fine on iOS as we only show one composer at a time but was failing on iPad as we show both channel and thread on the same screen. As a result `document.querySelector(".chat-composer__input")` was always targeting the first textarea on the screen which was the channel's composer, making it impossible to focus the thread's one.
This commit makes it so that when the user has unread threads
for a channel we show a blue dot in the sidebar (or channel index
for mobile/drawer).
This blue dot is slightly different from the channel unread messages:
1. It will only show if the new thread messages were created since
the user last viewed the channel
2. It will be cleared when the user views the channel, but the threads
are still considered unread because we want the user to click into
the thread list to view them
This necessitates a change to the current user serializer to also
include the unread thread overview, which is all unread threads
across all channels and their last reply date + time.
Why this change?
The test being changed in question has been flaky on our CI. However, we
are unable to view the screenshot of why it failed because
ActionDispatch will only take a screenshot of the default session upon
failure. At the same time, taking screenshot of all sessions
automatically upon failure is not possible via the official Capybara or
Rails APIs at the moment. Therefore, we're changing this system test to
avoid using two custom session and instead have the main assertion use
the default session such that any failures will provide us with a
screenshot.
Prior to this change you might end up in a loop where removing a channel would redirect you to this channel and as we auto-follow opened direct message channels, you could never unfollow this last direct message channel.
Followup to d7ef7b9c03,
this adds a spec to test the case where old threads are
still unread for the user and should show at the top regardless
of pagination, and fixes some issues/makes some slight refactors.
This commit attempts to fix an issue where we are ending
up with bad created_at date formats for last messages, which
is breaking the DM sort order and sometimes causing DM channels
to fall off the list, or show "Invalid date" on mobile.
I have not been able to consistently reproduce these issues
locally, however the serialzier for the channels index uses
MultiJSON.dump() and the Chat::Publisher uses .to_json, both of
which format created_at differently for messages.
The former is `2023-07-05T06:53:25.977Z` (iso8601).
The latter is `2023-07-14 03:59:22 UTC` (.to_s default).
Since we are doing comparison and sorting of these dates on the UI
we need consistent formatting for the JS Date parsers (and moment)
to deal with.
If the issue still occurs after this we can investigate further.
Prior to this commit a long press on the image of a chat message would trigger both the actions menu and the contextual menu. This commit ensures we only show the contextual menu in this case.
No test as it's a quite complex behavior to reproduce (would need android for example).
This was causing this event to cause other touch events down the road. For example click a reaction above the composer when the message action was opened could cause the composer to gain focus after the reaction was made.
It could only occur on message created by the user itself and deleted while the user was looking at the channel.
It more generally fix the trash service which was not correctly setting the author of the delete.
`SiteSetting.enable_public_channels` allows site admin to decide if public channels are available at all. There's no distinction between admins or not as we expect admins to create private category channels if they want to limit usage.
Not sure how this was even working previously, since it's trying
to press the reply button on a thread original message, which doesn't
work, you need to click the indicator to open the thread.
Why this change?
The following test is flaky on our CI:
```
1) Navigation when sidebar is configured as the navigation menu when re-opening full page chat after navigating to a channel opens full page chat on correct channel
Failure/Error: measurement = Benchmark.measure { example.run }
expected "/" to equal "/chat/c/random-9/17"
```
The theory here is that system tests is running too fast that we're not
giving the href for the chat header icon a chance to update before
clicking on it. Therefore, we're adding an additional assertion to
assert that the link has the right href before clicking on it.
Initial migration and changes to models as well as
changing the following services to update last_message_id:
* Chat::MessageCreator
* Chat::RestoreMessage
* Chat::TrashMessage
The data migration will set the `last_message_id` for all existing
threads and channels in the database.
When we query the thread list as well as the channel,
we look at the last message ID for the following:
* Channel - Sorting DM channels, and channel metadata for the list of channels
* Thread - Last reply details for thread indicators and thread list
It is now safe to render the message excerpt as HTML since
it is no longer using text_entities: true in the server
PrettyText.excerpt call when creating the message excerpt
from the cooked HTML.
This will fix the issue of things like mentions showing
HTML code instead of the actual mention when replying,
and cannot be used to inject improper HTML like style tags
via XSS.
This commit makes sure we don't load all data into memory when doing CSV exports.
The most important change here made to the recently introduced export of chat
messages (3ea31f4). We were loading all data into memory in the first version, with
this commit it's not the case anymore.
Speaking of old exports. Some of them already use find_each, and it worked as
expected, without loading all data into memory. And it will proceed working as
expected after this commit.
In general, I made sure this change didn't break other CSV exports, first manually, and
then by writing system specs for them. Sadly, I haven't managed yet to make those
specs stable, they work fine locally, but flaky in GitHub actions, so I've disabled them
for now.
I'll be making more changes to the CSV exports code soon, those system specs will be
very helpful. I'll be running them locally, and I hope I'll manage to make them stable
while doing that work.
* FEATURE: Inline topic summary. Cached version accessible to everyone.
Anons and non-members of the `custom_summarization_allowed_groups_map` groups can see cached summaries for any accessible topic. After the first 12 hours and if the posts to summarize have changed, allowed users clicking on the button will automatically re-generate it.
* Ensure chat summaries work and prevent model hallucinations when there are no messages.
Browser capabilities are inherently unconnected to the lifecycle of our app. Making them formally available outside of the service means that they can safely be used in non-app-linked functions without needing risky hacks like `helperContext()` or `discourse-common/lib/get-owner`.
One example of where the old hacks were problematic is the `translateModKey()` utility function. This is called in the root of the `discourse/components/modal/keyboard-shortcuts-help` es6 module. If anything (e.g. a theme/plugin) caused that es6 module to be `require()`d before the application was booted, a fatal error would occur.
Following this commit, `translateModKey()` can safely import and access `capabilities` without needing to worry about the app lifecycle.
The only potential downside to this approach is that the capabilities data now persists across tests. If any tests need to 'stub' capabilities, they will need to revert their changes at the end of the test (e.g. by using Sinon to stub a property).
This commit also updates some legacy references from `capabilities:main` to `service:capabilities`.
This implementation will need more work in the future. For simplification of tracking and other events (new thread, delete/restore OM...) we used the threads from `threadsManager` which makes pagination more complicated as we already have some results when we start.
Note this commit also simplify `Collection` to only have one `load` method which can be called repeatedly.
Why this change?
The specs are flaky on CI and we've unable to figure out why so we've
decided to skip them only on CI for now. The tests are still ran in our
internal build so we still have some protection in place.
Trying to fix two issues:
1. Sometimes the publish_new! event for update_thread_original_message
finishes running on the UI before the one for thread_created, in this
case we just want to do nothing because thread_created will fetch the
new thread along with its preview from the server if needed
2. Sometimes the thread GET and /read events were erroring because
last_reply on the thread was nil, this was potentially occuring because
the thread_created event was coming through to the UI before the rest
of MessageCreator was done, so we just move that after the big update
to set thread_id for the new and existing messages in the reply
chain
Why is this change being made?
We've decided that the previous "community" section should look more
like a primary section that holds the most important navigation links
for the site and the word "community" doesn't quite fit that
description. Therefore, we've made the decision to drop the
section heading for the community section.
As part of removing the section heading, the following changes are made
as well:
1. Button to customize the section has been moved to the "footer" of the
"More..." section when `navigation_menu` site setting is set to `sidebar`.
When `navigation_menu` is set to `header dropdown`, a button to customize
the section is shown inline.
2. The section will no longer be collapsable.
3. The title of the section is no longer customisable as it is no longer
displayed. As a technical note, we have not dropped any previous
customisations of the section's title previously in case we have to
bring back the header in the future.
4. The new topic button that was previously present in the header has
been removed alongside the header. Admins can add a custom section
link to the `/new-topic` route if there would like to make it easier for
users to create a new topic in the sidebar.
The `message_bus_channels` given to `MessageBus.last_ids(*message_bus_channels)` is not ordered, as a result the expectation of the tests could fail, this test ensures we check the contain of the input instead of content+order.
This commit also standardize the naming pattern of modals: `<Chat::Modal::FooBar />` and changes css class accordingly.
Co-authored-by: David Taylor <david@taylorhq.com>
When the loading slider is enabled, the rendering of `application.hbs` is slightly delayed compared to the old 'spinner' strategy. This means that if a route tried to render a dialog during its `model()` hook, the dialog wrapper element would not be present and an error would occur.
This commit detects that situation and delays rendering the error until the next runloop iteration. If the element is still not found, we print a useful error to the console.
In the long term, we should ideally convert the dialog service to use a pure-ember rendering strategy instead of leaning on a11y-dialog. But for now, this workaround should resolve the problems identified by the chat system specs.
It's way more common to have presence enabled than disabled, so we should have been making it the default from start.
This commit also changes the namespace of `<ChatUserAvatar />` into `<Chat::UserAvatar />` and refactors tests.
Sadly this function is one of the very hard to test codepaths of the app. We could in the future attempt to extract the content of the function to unit-test it.
When a user sends their first message in a thread we
automatically track the thread in the backend, but we
don't reflect this in the UI until the user re-opens
the thread. This commit fixes that by showing the new
tracking level in the UI.
Chat drawer was using the `DiscourseURL` hook `afterRouteComplete`. This hook suffer from a very poor implementation which makes it very unreliable:
```javascript
if (typeof opts.afterRouteComplete === "function") {
schedule("afterRender", opts.afterRouteComplete);
}
```
This commit attempts to return the promise from `handleURL` to directly use it and have a very reliable after transition hook.
In previous changes we prevented creating a channel to also make users follow the channel. We were forcing recipients to follow the channel on message sent but this was not including the creator of the message itself.
This commit fixes it and also write an end-to-end system spec to cover these cases. The message creator service is currently being rewritten and should correctly test and ensure this logic is present.
This commit also makes changes on the frontend to instantly follow a DM when you open it, this change prevents a green dot to appear for a split second when you send a message in a channel you were previously not following. Only recipients will see the green dot.
Since we create threads in the background regardless of whether
threading is enabled for a channel, we get the unexpected behaviour
of everyone having a lot of unread threads when threading is enabled
for the channel.
To counteract this, when the admin enables threads for a channel
we can just run a high priority background job to mark all threads
as read in the channel for all users, so they are essentially
starting from a clean slate.
Followup to 802fb3b194
We should not hide the replies count if there is only 1 participant
for a thread, because this makes it look like the last reply is the
only reply.
This introduces a PLATFORM_KEY_MODIFIER const that
can be used both client and server side, to determine
whether we should be using the Meta or Ctrl key based
on whether the user is on Windows/Linux or Mac.
Why this change?
Before this commit, there is a chance that we will transition the user
to a different route if the chat thread component has been destroyed
prior to the request for fetching messasges in a chat thread returning.
This commit makes it such that we simply ignore the request if the chat
thread component has been destroyed.
We believe this is the cause of the flaky system tests in plugins/chat/spec/system/navigation_spec.rb
which we've been seeing on CI.
Why this change?
`Faker::Lorem.paragraph` generates a differrent length of string
every time. When a string happens to be long, it can change the UI
across system test runs making it harder to reason about our system
tests across multiple runs since the state is never really consistent.
We will just generate a paragraph with a fixed length going forward so
that the UI remains consistent. This should make certain tests which
relies on the UI being in a certain state to become less flaky.
Why this change?
This change ensures that we scroll to the top of the message when
hovering over a message to ensure that the message actions container
that appears on hover is not hidden in the chat drawer when the content
of the chat message is long.
This commit includes several fixes and improvements to thread
original message handling:
1. When a thread's original message is deleted, the thread no longer
counts as unread for a user
2. When a thread original message is deleted and the user is looking
at the thread list, it will be removed from the list
3. When a thread original message is restored and the user is looking
at the thread list, it will be added back to the list if it was
previously loaded
In specific conditions (generally a small drawer, with a long message) it is possible to have the message’s actions menu to be displayed hover the drawer's header.
This is particularly hard to fix correctly using popper due to our positioning which is slightly at the limit of the container.
The proposed fix targets mostly the specs by ensuring the messages actions will be hidden before attempting to click any header's button.
Why this change?
In CI, we know we're clicking a link to a chat channel's threads list.
However, the threads list is not loaded and we want to add more
assertions here to try and figure out why. By asserting for the current
URL, we will at least know that the transition to the URL is successful.
- Presence needs to be explicitly set on the component now
- We were not checking and testing correctly the presence of the unread indicator in the menu
This commit replaces two existing screens:
- draft
- channel selection modal
Main features compared to existing solutions
- features are now combined, meaning you can for example create multi users DM
- it will show users with chat disabled
- it shows unread state
- hopefully a better look/feel
- lots of small details and fixes...
Other noticeable fixes
- starting a DM with a user, even from the user card and clicking <kbd>Chat</kbd> will not show a green dot for the target user (or even the channel) until a message is actually sent
- it should almost never do a full page reload anymore
---------
Co-authored-by: Martin Brennan <mjrbrennan@gmail.com>
Co-authored-by: Jordan Vidrine <30537603+jordanvidrine@users.noreply.github.com>
Co-authored-by: chapoi <101828855+chapoi@users.noreply.github.com>
Co-authored-by: Mark VanLandingham <markvanlan@gmail.com>
Without this fix, the following error is raised:
```
ActiveRecord::StatementInvalid:
PG::SyntaxError: ERROR: syntax error at or near ")"
LINE 4: WHERE thread_id IN ()
```
Why this change?
Chat system tests that opens the message actions on mobile have been
flaky on our CI. Those system test usually fails when the message
actions do not show up as expected causing subsequent actions to fail.
In the case of the `Reply to message - channel - mobile when the message has an existing thread replies to the existing thread`
system test, failure screenshot shows that we ended up navigating to the
thread instead of opening the message actions button. To understand why
this happens, we first need to understand that by default Capybara clicks
on the centre of an element. Also, we need to note that the HTML structure of
a chat message is like so:
```
<div class="chat-message-container">
<div class="chat-message">
<div class="chat-message-avatar" />
<div class="chat-message-content" />
<div class="chat-message-thread-indicator" />
</div>
</div>
```
Since `PageObjects::Pages::ChatChannel#expand_message_actions_mobile`
attempts to click on the `.chat-message-contaier`, there is a
possibility that the center of that element is the
`.chat-message-thread-indicator` element which would explain why we
navigated to the thread list instead of opening up the message actions.
This is possible because the content of the original chat message as
well as the message excerpt in the thread is randomly generated where the
length of the message and how the text wraps on mobile can affect the
height of the `.chat-message-content` element as thus its position in
the `.chat-message-container` element. In most cases, the middle of the
`.chat-message-container` happens to be the `.chat-message-content`
which is why this test "flakes" sometimes.
What is the solution?
Instead of clicking on the `.chat-message-container`, we be more
specific and click on the `.chat-message-content` element instead.
* UX: make timestamp font size smaller
* UX: participants use copy instead of avatar
* FIX: Move thread participant count into i18n
---------
Co-authored-by: Martin Brennan <martin@discourse.org>
* DEV: Fix flaky thread nav spec
When we transitioned from the chat thread panel under some conditions
the request for the thread would come back and realise the component
was destroyed, which was trying to do a transition to the channel
itself.
Now we check for the previous route here too and transition to the
correct route.
* DEV: Fix chat transcript spec relying on animation
The on-animation-end modifier is not reliable in system specs
because it fires instantly (we have disabled capybara animations)
so the showCopySuccess boolean can be mutated back to false straight
away.
Better to have a separate boolean tracked with a data-attr that we
can reliably inspect in the system spec.
Why this change?
By ensuring the reset happens in an `ensure` code block, we ensure that
the code will always be run even if code fails or an error is raised.
This helps to prevent leaking custom network condition states and
improves the stability of our system tests.
- Inline mentions on posts
- Inline mentions on chat messages
- The user autocomplete for the composer
- The user autocomplete for chat
- The chat section of the sidebar
Ember 4.x will be removing the 'named outlet' feature, which were previously relying on to render modal 'controllers' and their associated templates. This commit updates the modal.show API to accept a component class, and also introduces a declarative API which can be used by including the <DModal component directly in your template.
For more information on the API design, and conversion instructions from the current API, see these Meta topics:
DModal API: https://meta.discourse.org/t/268304
Conversion: https://meta.discourse.org/t/268057
This fixes a longstanding TODO to move the contents of the
UpdateUserCountsForChannels job to the ensure_consistency!
method of Chat::Channel, which runs every 15 mins as part of
periodical updates.
This commit also addresses the performance issue of the original,
where we would fetch all channels and do an individual query to
get the count and update the count of each one. Now we do it all
in one query, and only publish the changed channels to the UI.
Followup to 3f1024de76
The ActiveModel::Types.register(:array) call for chat was
called too late in the Zeitwerk load order in production,
causing this error:
> `lookup': Unknown type :array (ArgumentError)
> raise ArgumentError, "Unknown type #{symbol.inspect}"
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
We need to load the type and register it manually before the rest
of the chat files are loaded via the engine and Zeitwerk.
This will be used when we move the channel creation for DMs
to happen when we first send a message in a DM channel to avoid
a double-request. For now we can just have a new API endpoint
for creating this that the existing frontend code can use,
that uses the new service pattern.
This also uses the new policy pattern for services where the policy
can be defined in a class so a more dynamic reason for the policy
failing can be sent to the controller.
Co-authored-by: Loïc Guitaut <loic@discourse.org>
This has been flagged by our internal system as well and has been
failing on CI. Skip this for now to improve the stability of our system
test runs while we figure out why it is flaky.
Enabling/Disabling threading has been possible through command line until now. This commit introduces two new UIs:
- When creating a channel, it will be available once the category has been selected
- On the settings page of a channel for admins
Whenever a user opens a channel or marks it read, we now
update the last_viewed_at datetime for that channel membership
record. This is so we will be able to show thread unread indicators
in the channel sidebar that clear independently of the main thread
unread indicators. This unread functionality will follow in another
PR.
Addressing TODO about using chatApi in the ChatChannel model,
but since it's a model we cannot easily use the chatApi service.
The model function is only called in one place so we may as well
just move the call there since the component can use chatApi
Followup to c6b43ce68b
We can just use the rich excerpt everywhere since we know
we don't need text_entities -- that introduced security issues
just to fix a spec.
Introduced in cec68b3e2c,
this is flaky because if you click the back button before
the route is fully transitioned to the loaded thread,
we end up going to the history _before_ the thread list,
which ends up being the channel.
We need to make sure that everything is loaded for the
thread first, meaning the skeleton is not there.
Also exclude some noise from the capybara logs (image load failures)
Followup to 1526d1f97d
This commit fixes an N1 for mentions/user status
when querying chat threads. This only happened if
any of the thread OMs had mentions.
* FEATURE: Sort thread list by unread threads first
This commit changes the thread list to show the threads that
have unread messages at the top of the list sorted by the
last reply date + time, then all other threads sorted by
last reply date + time.
This also fixes some issues by removing the last_reply
relationship on the thread, which did not work for complex
querying scenarios because its order would be discarded.
* FIX: Various fixes for thread list loading
* Use the channel.threadsManager and find the channel first rather
than use activeChannel in the threads manager, otherwise we may
be looking at differenct channels.
* Look at threadsManager directly instead of storing result for threads
list otherwise it can get out of sync because of replace: true in
other places we are loading threads into the store.
* Fix sorting for thread.last_reply, needed a resort.
When clicking back from a thread, we want to either go back to the
channel if the thread was opened from an indicator, or to the thread
list if we opened it from there. Since ember doesn't give a nice way
to get the previous route, we need to store this ourselves. We only
do this on mobile, on desktop we just follow existing behaviour.
Also implements a chat router history.
---------
Co-authored-by: Joffrey JAFFEUX <j.jaffeux@gmail.com>
Updates the interface for implementing summarization strategies and adds a cache layer to summarize topics once.
The cache stores the final summary and each chunk used to build it, which will be useful when we have to extend or rebuild it.
Cater for polls that can have multiple votes per user.
This fixes an older UserMerge and migration which was intended to
de-duplicate poll votes but did not account for "multiple" type polls.
This small patch registers a new `ActiveModel` type: `array`.
It will split a string on `,` to create a new array. If the value is
already an array, nothing will happen and for all other types, it will
wrap the value in an array.
Here’s an example on an existing contract:
```ruby
attribute :target_usernames
before_validation do
self.target_usernames =
(
if target_usernames.is_a?(String)
target_usernames.split(",")
else
target_usernames
end
)
end
# can be rewritten as:
attribute :target_usernames, :array
```
- gridified the thread message indicator, alleviating some problems with positioning and overflow
participant avatars will overlap/smush on smaller size and mobile
- the excerpt went from 3 > 2 lines of wrapping on smaller size, still 1 line on large size
- dropped the copy of "last reply"
- fixed wrong line height
- moved the "x replies" over to the right near the participants, as that makes more sense
- using a bubble to indicate other participants, instead of copy
This PR introduces the @container query, which is experimental. Nothing will break when it's being viewed in a not-supported browser, but it will be less elegant.
To export chat messages, go to `/admin/plugins/chat` and click the Create export
button in the _Export chat messages_ section. You'll receive a direct message
when the export is finished.
Currently, this exports all messages from the last 6 months, but not more than
10000 messages.
This exports all chat messages, including messages from private channels and
users' direct conversations. This also exports messages that were deleted.
This should prevent the message to show as active on mobile when making a touch to start scrolling.
This commit also makes naming of touch lifecycle functions coherent.
We had a bug in this code recently, sometimes users saw weird notifications
like:
User mentioned all_mentioned_user_ids in the help chat channel
We fixed that bug in b85d057.
This refactoring is a follow-up to that fix. As that bug showed, it’s quite easy
to introduce a key that may end up being sent to the `NotifyMentioned` job,
which can lead to such weird notifications. This refactoring makes sure that
the `to_notify` hash contains only IDs of users that should be notified about
mentions.
This PR adds a new parameter to fetch chat messages: `target_date`.
It can be used to fetch messages by a specific date string. Note that it does not need to be the `created_at` date of an existing message, it can be any date. Similar to `target_message_id`, it retrieves an array of past and future messages following the query limits.
This commit adds an aria-label attribute to cooked hashtags using
the post/chat message decorateCooked functionality. I have just used
the inner content of the hashtag (the tag/category/channel name) for
the label -- we can reexamine at some point if we want something
different like "Link to dev category" or something, but from what I
can tell things like Twitter don't even have aria-labels for hashtags
so the text would be read out directly.
This commit also refactors any ruby specs checking the HTML of hashtags
to use rspec-html-matchers which is far clearer than having to maintain
the HTML structure in a HEREDOC for comparison, and gives better spec
failures.
c.f. https://meta.discourse.org/t/hashtags-are-getting-a-makeover/248866/23?u=martin
https://meta.discourse.org/t/markdown-preview-and-result-differ/263878
The result of this markdown had different results in the composer preview and the post. This is solved by updating Loofah to the latest version and using html5 fragments like our user had reported. While the change was only needed in cooked_post_processor.rb for this fix, other areas also had to be updated due to various side effects.
- Moves `<ChatMessageInfo />` to `<Chat::Message::Info />`
- Moves `<ChatMessageAvatar />` to `<Chat::Message::Avatar />`
- Moves `<ChatMessageLeftGutter />` to `<Chat::Message::LeftGutter />`, adds tests
- Creates `<Chat::Message::Error />`
- Creates `<Chat::Message::MentionWarning />`, adds tests and a styleguide
- Creates a model for ChatMessageMentionWarning, adds fabricator for it
- Keeps the enter/leave viewport logic inside the `<ChatMessage />` component instead of bubbling it to the channel and thread components
- Adds a scale animation when clicking a reaction
- Creates `chat/later-fn` modifier which accepts a function and a delay. It allows to call a function Xms after a component has been inserted, it's useful for animations.
- Moves css code out of chat-message into relevant files
- Deletes unused code
<!-- NOTE: All pull requests should have tests (rspec in Ruby, qunit in JavaScript). If your code does not include test coverage, please include an explanation of why it was omitted. -->
Previously this was defining methods like `UserOption#never?`, `UserOption#all_new?`, `UserOption#dm_and_mentions?`. Now they will be prefixed like `UserOption#chat_header_indicator_never?`
The layout was broken for messages replying to another message in non threaded channels.
This commit also refactors the chat-message-test to use fabricators.
Unfortunately, Discourse's `UserOption` model is not currently autoloaded. That means that modifying it via a `reloadable_patch` will try to apply the changes repeatedly. Normally this doesn't matter since the changes are idempotent. However, introducing ActiveRecord enums is not idempotent - they raise an error if the same enum already exists on the model. This commit adds a check to avoid hitting this 'duplicate definition' error.
Reproduced
```
rails runner 'Rails.application.reloader.reload!'
```
- FIX: improves reactions and thread indicator touch event on mobile
These "buttons" are located inside a scroll list which makes them very specific. The general idea is to ensure these events are passive and are not bubbling to the parent.
- DEV: moves state on top level message node
- FIX: ensures popover arrow has the correct border
- FIX: makes a message expanded by default
- FIX applies the same ios scroll fix on thread and channel
- UI: better active/hover state for thread indicator
- UI: attempts to follow more closely our BEM naming scheme
- FIX: reduces bottom padding on message with thread indicator and user info hidden
- UI: add padding for first message in thread
- FIX: prevents actions backdrop to open thread
- UI: makes thread indicator resizable
This method is a huge footgun in production, since it calls
the Redis KEYS command. From the Redis documentation at
https://redis.io/commands/keys/:
> Warning: consider KEYS as a command that should only be used in
production environments with extreme care. It may ruin performance when
it is executed against large databases. This command is intended for
debugging and special operations, such as changing your keyspace layout.
Don't use KEYS in your regular application code.
Since we were only using `delete_prefixed` in specs (now that we
removed the usage in production in 24ec06ff85)
we can remove this and instead rely on `use_redis_snapshotting` on the
particular tests that need this kind of clearing functionality.
This commit adds a tracking dropdown to each individual thread, similar to topics,
that allows the user to change the notification level for a thread manually. Previously
the user had to reply to a thread to track it and see unread indicators.
Since the user can now manually track threads, the thread index has also been changed
to only show threads that the user is a member of, rather than threads that they had sent
messages in.
Unread indicators also respect the notification level -- Normal level thread tracking
will not show unread indicators in the UI when new messages are sent in the thread.
Why is this change required?
When a site is newly setup and a user has just been created, the
categories and tags sections are hidden from the user. This happens
because the admin has not configured the `default_navigation_menu_categories` or
`default_navigation_menu_tags` site settings. When the categories and tags
sections are hidden from the user, the sidebar looks extremely bare and
does not create a good experience.
What is being change?
In this commit, we're changing the logic such that the site's top
categories and tags are displayed if the user does not have any
categories/tags configured in each respective section. The only
regression introduced in this change is that the categories and tags
section can no longer be hidden as a result. However, we have plans to
address this in the future by allowing sidebar sections to be configured
to be hidden by each individual user.
The events leading to this mistake are unclear but we decided few months ago to make direct messages NOT flaggable and even wrote a spec for this, when we actually support flagging of direct messages.
This commit ensures it will show for direct messages channels and inverses the existing spec.
https://meta.discourse.org/t/updating-our-initializer-naming-patterns/241919
For historical reasons, Discourse has different initializers conventions than standard Ember:
```
| Ember | Discourse | |
| initializers | pre-initializers | runs once per app load |
| instance-initializers | (api-)initializers | runs once per app boot |
```
In addition, the arguments to the initialize function is different – Ember initializers get either the `Application` or `ApplicationInstance` as the only argument, but the "Discourse style" gets an extra container argument preceding that.
This is confusing, but it also causes problems with Ember addons, which expects the standard naming and argument conventions:
1. Typically, V1 addons will define their (app, instance) initializers in the `addon/(instance-)initializers/*`, which appears as `ember-some-addon-package-name/(instance-)initializers/*` in the require registry.
2. Just having those modules defined isn't supposed to do anything, so typically they also re-export them in `app/(instance-)initializers/*`, which gets merged into `discourse/(instance-)initializers/*` in the require registry.
3. The `ember-cli-load-initializers` package supplies a function called `loadInitializers`, which typically gets called in `app.js` to load the initializers according to the conventions above. Since we don't follow the same conventions, we can't use this function and instead have custom code in `app.js`, loosely based on official version but attempts to account for the different conventions.
The custom code that loads initializers is written with Discourse core and plug-ins/themes in mind, but does not take into account the fact that addons can also bring initializers, which causes the following problems:
* It does not check for the `discourse/` module prefix, so initializers in the `addon/` folders (point 1 above) get picked up as well. This means the initializer code is probably registered twice (once from the `addon/` folder, once from the `app/` re-export). This either causes a dev mode assertion (if they have the same name) or causes the code to run twice (if they have different names somehow).
* In modern Ember blueprints, it is customary to omit the `"name"` of the initializer since `ember-cli-load-initializers` can infer it from the module name. Our custom code does not do this and causes a dev mode assertion instead.
* It runs what then addon intends to be application initializers as instance initializers due to the naming difference. There is at least one known case of this where the `ember-export-application-global` application initialize is currently incorrectly registered as an instance initializer. (It happens to not use the `/addon` folder convention and explicitly names the initializer, so it does not trigger the previous error scenarios.)
* It runs the initializers with the wrong arguments. If all the addon initializer does is lookup stuff from the container, it happens to work, otherwise... ???
* It does not check for the `/instance-initializers/` module path so any instance initializers introduced by addons are silently ignored.
These issues were discovered when trying to install an addon that brings an application initializer in #22023.
To resolve these issues, this commit:
* Migrates Discourse core to use the standard Ember conventions – both in the naming and the arguments of the initialize function
* Updates the custom code for loading initializers:
* For Discourse core, it essentially does the same thing as `ember-cli-load-initializers`
* For plugins and themes, it preserves the existing Discourse conventions and semantics (to be revisited at a later time)
This ensures that going forward, Ember addons will function correctly.
This commit fixes the selection of message in threads and also applies various refactorings
- improves specs and especially page objects/components
- makes the channel/thread panes responsible of the state
- adds an animationend modifier
- continues to follow the logic of "state" should be displayed as data attributes on component by having a new `data-selected` attribute on chat messages
When merging users, polls may error out if the source and target users have both voted on the same poll before. 😢
There is no constraint on the `poll_votes` table either to support this. Ideally a composite primary key can be used `(poll_id, user_id)`, but alas there is no support yet, which is probably why it wasn't created in the first place.
This fix ensures that merging is successful by only keeping the target poll votes if duplicates exist.
This fix also runs a migration on older poll votes where failed merges would have caused a single user to have voted twice on a single poll. e.g. this weird edge case
This commit adds the initial part of thread indicator improvements:
* Show the reply count, last reply date and excerpt,
and the participants of the thread's avatars and
count of additional participants
* Add a participants component for the thread that
can be reused for the list
* Add a query class to get the thread participants
* Live update the thread indicator more consistently
with the last reply and participant details
image image
In subsequent PRs we will cache the participants since
they do not change often, and improve the thread list
further with participants.
This commit also adds a showPresence boolean (default
true) to ChatUserAvatar, since we don't want to show the
online indicator for thread participants.
---------
Co-authored-by: chapoi <charlie@discourse.org>
Since we created user_chat_thread_memberships in
cc2570f we haven't
yet backfilled it for users who previously sent a message in
in threads -- this migration creates the UserChatThreadMemberships
needed for those threads, making sure the last read message id
is accurate for those participants.
* FEATURE: Content custom summarization strategies.
This PR establishes a pattern for plugins to register alternative ways of summarizing content by extending a class that defines an interface.
Core controls which strategy we'll use and who has access to it through the `summarization_strategy` and `custom_summarization_allowed_groups`. It also defines the UI for summarizing topics.
Other plugins can access this summarization mechanism and implement their features, removing cross-plugin customizations, as it currently happens between chat and the discourse-ai plugin.
* Group membership validation and rate limiting
* Work with objects instead of classes
* Port summarization feature from discourse-ai to chat
* Rename available summaries to 'Top Replies' and 'Summary'
In early 2015, the poll plugin was writing its data to custom fields on
the post containing the poll. It was later changed to have dedicated SQL
tables and the polls were migrated but we forgot to clean the existing
data.
* move the chat unread indicator to top to match the profile avatar indicator
* add white border to profile avatar indicator (badge notification) to match chat indicator and userstatus styling
* change `.urgent` to BEM
* congregate all styling into mixin
* update chat index to use mixin
* update thread indicator to use mixin
* update header indicator to use mixin
---------
Co-authored-by: Joffrey JAFFEUX <j.jaffeux@gmail.com>
Co-authored-by: Martin Brennan <martin@discourse.org>
This is actually making things more sluggish than necessary. If any perf issue happen out of this they should be handled in the consequences of the resizing, not the resizing itself.
Currently navigating a long topic and then opening chat would cause the view to be scrolled to the bottom. Using `scrollTop` here ensures we correctly scroll to top.
This had been incorrectly moved into `deactivate` during another change.
* FIX: increases resize observer throttle delay
25ms is not necessary and was sometimes causing jankyness.
* FIX: removes ios momentum fix delay
Instead of a 50ms, simply use next+schedule("afterRender") to attempt to have the shortest delay possible.
* FIX: backdrop event propagation
Prevents backdrop touch to propagate to underlying channel/thread.
* UX: adds is-active class to container of active message
This change allows to keep the background on the active message while the actions menu is displayed.
* FIX: prevents skip-link to be selected on press
* UX: allows to close actions menu instantly
The backdrop should always receive events, we don't need to wait for the menu to be fully displayed.
* UI: adds spacing between last message and composer
* UI: makes backdrop less dark
* FIX: makes events passive on long-press modifier
We have been struggling a lot on this lately as it's almost impossible to write a decent test for this.
The important things which need to happen:
- fetch the unread/mention state and last message bus channel ids of each chat channels
- stop all subscriptions
- restart global chat subscriptions
- update channels with new state and ensure the message bus ids are updated
- restart subscriptions of each chat channel
As a followup we need to start implementing a standard way to query for a resource state. Something similar to: `/channels/tracking` and `/channels/:id/tracking`
Each of these endpoints would return a state similar to:
```json
{
tracking: { ... },
message_bus_ids: { ... }
}
Removing a reaction could start a long press at the same time and put the screen in a stuck state.
This commit ensures we give an opportunity to the reaction to capture the event first and not propagate further.
These spec are flaky only in CI, not locally and not in GitHub actions.
The previous attempt was in 44eabde, but actually the failure happens
a bit earlier. This is another attempt to fix these specs. Quite a lot of
async logic is happening in emulateAutocomplete(), a call to settled()
in the end should help make it more reliable.
This commit attempts to have a bullet proof solution to the following case:
- long press on message (finger is still pressed)
- menu appears
- a button is now at finger location
- user releases finger
- a click is triggered on the button
Classic event canceling solution won't work here for performance reasons as we need the event to be passive in a scroll list.
In some cases, plugins may want to hide some of these actions
at all times, overriding the rules for canX with hiding these
buttons. To achieve this, a plugin can call the API
`removeChatComposerSecondaryButtons` and pass the list of button
IDs that should be removed as argument, like the example below:
```
withPluginApi("1.2.0", (api) => {
api.removeChatComposerSecondaryActions("copyLink", "select");
});
```
---------
Co-authored-by: Martin Brennan <martin@discourse.org>
These specs were skipped in d6d5eae1. They sometimes failed, and only on CI,
not in GitHub Actions.
I wasn't able to reproduce failures locally, but I expect clicking the send button
in chat composer should be more reliable than emulating pressing <kbd>Enter</kbd>.
When editing a message, we call `message.cook()` in the beginning of
`#sendEditMessage` methods, but when sending a new message,
the call to `message.cook()` is hidden in the `stageMessage` method.
We can just call `message.cook()` before sending the message, no matter
whether this is a new message or an edited message.
No test as this is very much a hack while lightbox is being revamped. We currently have no good/easy way AFAIK to stop event propagation on escape in magnificpopup.
This behavior has been possible for a long time and has been made more in recent commits. On PWA iOS when release the touch after the mobile actions menu has been shown, if the finger was over one of the buttons of the menu, it would trigger a click.
This commit should now correctly trap and cancel events.
The following case would create the perception of a broken back button on desktop:
- open discourse on home page
- click chat button in header
- hit back button
- observes that we are still on the channel didn't navigate to homepage as we would have expected
The back button is actually working but it's in a loop. We were doing a `transitionTo` after finding the ideal channel to show, so the browser history would look something like this:
- home
- chat index
- channel page
When hitting back, we would go to chat index which would run the same logic and transition us to channel page.
This change will use `replaceWith` to replace the chat index step by the channel step, this way our history will now look like this:
- home
- channel page
Hitting back will now correctly bring us to home.
This commit attempts to refactor our long press logic to make it more resilient and precise.
With this improvement two very UX/UI changes have been made:
- scale animation on long press
- prevents click on reaction to propagate to the message which would cause the active state of the message to trigger
This fixes an issue where a user could send an empty
string as a chat message .e.g ' ' and the message would
be posted. We don't want this, we need to strip the message
first before validating for length etc.
Followup to e6c6c342d9,
we missed one part of this refactor which was to give
the correct composer element reference to ChatComposerUploads.
Without this the pasteEventListener for uploads was not
bound so uploading via paste did not work.
I tried to test this, and though I can write binary/text to
the clipboard and paste it into the composer, it does not
seem to be possible to end up with a paste event that
has clipboardData.files filled in, which is what is used
for the uploads. I think this is a restriction of JS
generally, and there doesn't seem to be a way to work around
it, so unfortunately we have to have no test for this still.
Usually, when a user is promoted to TL2 two messages are sent. The
first one is a system message 'tl2_promotion_message' which triggers a
'system_message_sent' Discourse event.
When the event is fired and if Discourse Narrative Bot is enabled, then
a second message is sent to the recipient of the first message. The
recipients was determined by looking at the list of users that can
access that topic and pick the last one. This method does not work if
'site_contact_group_name' site setting is set because it adds the group
in the list of recipients.
A solution to this problem would have been to select the last user in
the list of 'topic_allowed_users', but an even better solution is to
pass the name of the recipients when the 'system_message_sent'
Discourse event is fired.
This commit contains multiple changes to improve the composer behavior especially in the context of a thread:
- Generally rename anything of the form `chatChannelThread...` to `chatThread...``
- Moves the textarea interactor instance inside the composer server
- Improves the focus state and closing of panel related to the use of the Escape shortcut
- Creates `Chat::ThreadList` as a component instead of having `Chat::Thread::ListItem` and others which could imply they were children of a the `Chat::Thread` component
Previously, there was an issue where closing the message actions menu on mobile would unintentionally trigger a click event on an element below it, such as a thread indicator or a reaction. With the recent fix, this problem has been resolved. Now, when you close the menu, it will no longer interfere with or activate any elements positioned underneath it.
One user can create a post or chat message with a hashtag they
have permission to use, but then when other users look at that
post they will see an empty space next to the hashtag because they
do not have the permission to load the colors in CSS classes for
the related category.
This fixes the issue by adding a default color with a special
CSS class if the user doesn't have permission to see the linked
channel/category on the hashtag.
Few weeks ago we implemented `onPresenceChangeCallback` to re-sync chat channels state when going back to a long time inactive tab. This codepath however contained a bug as we were reseting all subscriptions but only restarting global subscriptions and not per channel subscriptions.
This commit should correctly ensure we correctly do so. It's sadly very hard to test time related changes in system specs.
Co-authored-by: Martin Brennan <mjrbrennan@gmail.com>
What is the problem?
We were calling out to methods that calls `has_css?` or `has_selector?`
which returns a boolean. Since we are not using the return value, it
means the methods can be deemed unnecessary. However, we do want those
checks and this commit adds the necessarily assertions to make use of
the return values.
When a user type a message with mentions, the autocomplete popup
may suggest users or groups. We were adding all these object to
the `currentMessage.mentionedUsers` collection, while we should
have been adding only users. A group added to that collection led to
the error later when trying to update user status on mentions.
This will make it simpler to work with this code. This also can make this code more stable and increase stability of our test suite.
Cooked message now will be available immediately after cooking, it wasn't the case before:
await message.cook();
const cooked = message.cooked;
This also removes a call to `message.cook()` from message fabricator. Alternatively we may leave the call there and make the fabricator function async, but I fill it's better this way. If someone needs to test something related to cooked message, they can either pass cooked text to fabricator:
message = fabricators.message({ cooked: "<p>cooked</p>" });
or call `message.cook()` after fabrication:
message = fabricators.message({ message: "raw message" });
await message.cook()
This reverts commit ddf4ecba04.
Causing a flaky test to appear:
```
main $ LOAD_PLUGINS=1 rspec plugins/chat/spec/system/chat/composer/shortcuts/channel_spec.rb
Randomized with seed 17765
.....F..
Failures:
1) Chat | composer | shortcuts | channel when using ArrowUp when last message is staged does not edit a message
Failure/Error: channel_page.send_message
expected `#<PageObjects::Components::Chat::Messages:0x00007fe823ac1710 @context=".chat-channel">.has_message?({:persisted=>true, :text=>"2"})` to be truthy, got false
[Screenshot Image]: /home/tgxworld/work/discourse/tmp/capybara/failures_r_spec_example_groups_chat_composer_shortcuts_channel_when_using_arrow_up_when_last_message_is_staged_does_not_edit_a_message_148.png
```
What is the problem?
We were calling out to methods that calls `has_css?` or `has_selector?`
which returns a boolean. Since we are not using the return value, it
means the methods can be deemed unnecessary. However, we do want those
checks and this commit adds the necessarily assertions to make use of
the return values.
If we're asserting that something is missing, we want to use
`has_no_css?` instead of `!has_css?` since `has_css?` will wait the full
capybara default wait time before return if the selector is not present.
* FEATURE: reduce avatar sizes to 6 from 20
This PR introduces 3 changes:
1. SiteSetting.avatar_sizes, now does what is says on the tin.
previously it would introduce a large number of extra sizes, to allow for
various DPIs. Instead we now trust the admin with the size list.
2. When `avatar_sizes` changes, we ensure consistency and remove resized
avatars that are not longer allowed per site setting. This happens on the
12 hourly job and limited out of the box to 20k cleanups per cycle, given
this may reach out to AWS 20k times to remove things.
3.Our default avatar sizes are now "24|48|72|96|144|288" these sizes were
very specifically picked to limit amount of bluriness introduced by webkit.
Our avatars are already blurry due to 1px border, so this corrects old blur.
This change heavily reduces storage required by forums which simplifies
site moves and more.
Co-authored-by: David Taylor <david@taylorhq.com>
- ensures buttons are aligned to the bottom
- makes the emoji icon tertiary as initially intended
- correctly sets the icon scale of the sending button
- Made the emoji btn blue when composer is focused
- Moved everything chat-composer-button to its own file and BEM-ified it and making the choice to only work with our own is-disabled definition instead of with the attribute :disabled, for consistency
This should fix this failure:
```
Failures:
1) Thread list in side panel | full page when there are no threads that the user is participating in shows a message
Failure/Error: measurement = Benchmark.measure { example.run }
expected to find text "You are not participating in any threads in this channel." in "Community\nEverything\nMy Posts\nMore\nMessages\nInbox\nChannels\nRandom 25\nPersonal chat\nRandom 25\nShowing all messages\nOngoing discussions"
```
The screenshot failure was clearly showing the spinner still being present.
If the drawer receives an unexpected route, attempt to show the index. This is probably a more serious issue with subfolder but should limit the effects.
Rescuing them still makes timing-out tests fail but doesn't break `after` spec cleanup (which could trigger more errors) Using custom error class to avoid any other possible timeout-catching code.
Also:
* remove an unnecessary `.select { |x| x.size > 0 }`
* fix a typo in a test title
We were calling reset without the proper params which was causing errors in the console. This commit does the following changes:
- ensures `composer.cancel()` is the only way to cancel editing/reply
- adds a `draftSaved` property to chat message to allow for better tests
- writes a spec to ensure the flow is correct
- adds more page objects for better tests
- homogenize the default state of objects on chat message
Co-authored-by: Martin Brennan <martin@discourse.org>
Editing a message to an empty string and sending it, will delete it.
This commit also refactors a lot of channel/thread composer shortcuts specs.
---
This commit also includes various spec fixes which have been flakey while finishing this pull request.
These specs were disabled in 786f7503. While investigating this, I found out that at some point `:user_membership` got deleted. It's hard to tell why exactly without investing more time, but it seems using `let!` instead of `fab!` solves the issue.
If in the future we decide to investigate why these tests were flaky with `fab!` to reproduce the failure run:
LOAD_PLUGINS=1 rspec --seed 46586 plugins/chat/spec/mailers/user_notifications_spec.rb
This changes the thread header positioning of the
unread indicator to match the designs based on the route:
1. When the channel is open, show the indicator of # unread
threads with the icon
2. When the threads list is open, show no indicator since
you are on the list and can see which threads are unread
3. When a single thread is open, show the unread threads
indicator along with a left < back button, with a label
to show that this goes back to ongoing discussions
Drawer changes to come in another PR.
Why is this change required?
In the `PageObjects::Components::Chat::Messages#has_no_message?` method,
it ended up calling `has_selector` when trying to assert that the
selector is not present. This is an anti-pattern which results in us
waiting the full Capybara default wait time
What is this change required?
In the `chat/spec/system/transcript_spec.rb` test, there is a helper
method that uses `page.has_css?` in a conditional but it do not
specify a wait time and hence the default Capybara default max wait
time is used. However, there is no need for us to be waiting here so
we specify the `wait: 0` option.
Followup 55ef2d0698.
In the cases where the user has no last_read_message_id for
a channel, we want to make sure that a page_size is set for
the ChannelViewBuilder + MessagesQuery, otherwise we end up
loading way more messages than needed (the additional message
loading was fixed in the last commit).
This commit introduces a couple of changes:
1. When editing a chat channel's slug, we were using `this.model.set("title", title)` when the `set`
function does not exist. This was actually throwing the error in the
"can edit slug" system test where the modal was not closed after
saving and was flashing an error.
2. Introduce `PageObjects::Pages::ChatChannelAbout` and
`PageObjects::Modals::ChatChannelEdit` page object to encapsulate
logic better.
When a thread is created / a new message is created in the
thread, we want to make sure that the original message user
has a membership for that thread, otherwise they will not
receive unread indicators for messages in the thread.
This commit attempts to fix the case where the messages loaded initially don't fill the screen. It would prevent user to scroll and as a result to load more.
There are multiple fixes in this commit:
- the main fix is removing this code which was preventing the actual fill:
```javascript
// prevents an edge case where user clicks bottom arrow
// just after scrolling to top
if (loadingPast && this.#isAtBottom()) {
return;
}
```
- ensures we always give a page site to the `chatApi.channel(...)` call if we have one, in the current state when `fetchFromLastRead` was `true` we would not set `args.page_size`
- ensures the `query_paginated_messages` is having a valid page size, which is not nil and not > `MAX_PAGE_SIZE`
- write a spec for the autofill, it was a challenging spec to write but it should give us the confidence we need here
Since 5cce829 and the new
channel view builder, we have no need of these obsolete
routes which have way too much logic in the controller, which
has been superseded by the view builder anyway.
Remove the routes and update the channel message loading to use it.
* Moved the settings cog from thread list to thread and
put it in a new header component
* Remove thread original message component, no longer needed
and the list item and thread indicator styles/content
will be quite different
* Start adding content (unread indicator etc.) to the thread
list item and changing structure to be more like designs
* Serialize the last thread reply when opening the thread index,
show in list and update with message bus
In some cases activeChannel can be null so this will error,
also it is limiting to have this code in chatApi. Instead
move to the threads manager, and also lean on channelsManager.find
to get the channel from the cache instead, which will not error.
The current behavior is to close drawer when pressing escape inside the input.
After this change, first escape will blur the input, and second escape will close the drawer.
This commit also refactors the whole shortcuts for drawer system spec.
Followup to d4a5b79592,
this introduced an N1 because every message in the list
we had to query users for the mentions and then the user's
status too. Instead we can just include both in Chat::MessagesQuery.
#### FIX: Do not use client lastReadMessageId when fetching channel messages
We had an issue where the following happened:
1. User opened channel and saw the last message, and we set the
lastReadMessageId on the server and the client
2. User navigated to another channel
3. Another user deleted the message in the original channel
4. The first user navigated back to the original channel before
the MessageBus event for the deleted message arrived, and got
a 404 error because we were sending the deleted lastReadMessageId as
target_message_id to the channel controller.
Instead of this which is a bit flaky and is hard to cover all
the issues for, instead we can pass a fetch_from_last_read boolean
param to the channels controller, and just get the user's
last_read_message_id straight from the database to use for the
target_message_id. This gets rid of any sources of race conditions
or lack of updates from MessageBus.
#### FIX: Include missing memberships for thread tracking publish
When we publish the channel/message tracking state for a
user and that message was a thread reply the publisher
was erroring because we were not telling Chat::TrackingStateReportQuery
to return missing memberships (which have zeroed out unread counts)
as well, which is what we do for the channel tracking state here.
Also just make sure that the TrackingStateReport does not error
when passed an ID it doesn't have data for.
This flakey has been very visible by the new headless chrome. The problem was that after moving a message we automatically redirect to the channel where the message has been moved to. However, we were not explicitly waiting for this transition and a result it could happen that we attempt to check the presence of the message on the channel page before the redirect actually happened.
The various naming changes are due to an early mistake we made in chat specs to use `chat` as the variable name for the page object which prevents to use the automatic path `chat.channel_path(...)`.
Co-authored-by: Alan Guo Xiang Tan <gxtan1990@gmail.com>
This commit follows up b6c5a2da08
by serializing the user's thread memberships in these cases:
1. When we do the initial channel fetch with messages, we get
all threads and all the user's thread memberships for those
messages.
2. When the thread list is fetched, we get all the user's memberships
in that list.
3. When the single thread is fetched, either from opening it from
the list, an OM indicator, or just from doing .find() on the
manager when a new MessageBus message comes in
This will let us track the lastReadMessageId on the client, and
will also let us fix an issue where the unread indicator in the
channel header was incrementing for every thread that got a
new message, regardless of whether the user was a member.
This patch introduces policy objects to chat services. It allows putting
more complex logic in a dedicated class, which will make services
thinner. It also allows providing a reason why the policy failed.
Some change has been made to the service runner too to use more easily
these new policy objects: when matching a failing policy (or any failing
step actually), the result object is now provided to the block. This
way, instead of having to access the reason why the policy failed by
doing `result["result.policy.policy_name"].reason` inside the block,
this one can be simply written like this:
```ruby
on_failed_policy(:policy_name) { |policy| policy.reason }
```
This commit adds the thread index and individual thread
in the index list unread indicators, and wires up the message
bus events to mark the threads as read/unread when:
1. People send a new message in the thread
2. The user marks a thread as read
There are several hacky parts and TODOs to cover before
this is more functional:
1. We need to flesh out the thread scrolling and message
visibility behaviour. Currently if you scroll to the end
of the thread it will just mark the whole thread read
unconditionally.
2. We need to send down the thread current user membership
along with the last read message ID to the client and
update that with read state.
3. We need to handle the sidebar unread dot for when threads
are unread in the channel and clear it based on when the
channel was last viewed.
4. We need to show some indicator of thread unreads on the
thread indicators on original messages.
5. UI improvements to make the experience nicer and more
like the actual design rather than just placeholders.
But, the basic premise around incrementing/decrementing the
thread overview count and showing which thread is unread
in the list is working as intended.
This commit regroups 3 changes
- serializes channel ID and json draft when calling the debouncer instead of inside the debounced function, it seems very unlikely to happen, but in a case where the debouncing wouldn't be canceled and the new message not set yet, we could save the draft on an invalid channel
- cancel persist draft handler when changing channel
- ensures we exit early when channel is not set
Very fast or specific mouse moves could allow to leave a message actions menu without reseting the active message. This commit should ensure we correctly catch this event.
No test as it's hard and not reliable to reproduce these in a test.
This should also make `message_notifications_with_sidebar_spec.rb` more resilient as we are now checking for `is-persisted` class instead of checking for the absence of `is-staged`.
This commit attempts to correctly change draft when the channel changes. It moves responsibility to the composer instead of the channel.
A new service `chatDraftsManager` is being introduced here to allow finer control and pave the way for future thread draft support.
These changes also now allow an editing message to be stored as a draft.
This PR adds status to mentions in chat and makes those mentions receive live updates.
There are known unfinished part in this implementation: when posting a message, status on mentions on that message appears immediately, but only if a user used autocomplete when typing the message. If user copy and paste a message with mentions into chat composer, those mentions won't have user status on them.
PRs with fixes for both problems are following soon.
Preparations for this PR that were made previously include:
- DEV: correct a relationship – a chat message may have several mentions 0dcfd7ddec
- DEV: extract the logic for extracting and expanding mentions from ChatNotifier 75b81b6854
- DEV: Always create chat mention records fa543cda06
- DEV: better split create_notification! and send_notifications logic e292c45924
- DEV: more tests for mentions when updating chat messages e7292e1682
- DEV: extract updating status on mentions into a lib function e49d338c21
- DEV: Create and update chat message mentions earlier 35a414bb38
- DEV: Create a chat_mention record when self mentioning 2703f2311a
- DEV: When deleting a chat message, do not delete mention records f4fde4e49b
In the ChannelViewBuilder, we introduced a check to see if
the target message exists, which errors if the message has
been trashed. However if the user is the creator of the message
or admin then they are able to see trashed messages, so
we need to take this into account.
Followup to c908eeacc9
Instead of using the latest message ID in the channel, which
could cause issues if you have an earlier last read message ID
that matches the deleted one, instead we use the first non-deleted
message that comes before the deleted message by ID.
Followup ae3231e140, when a
message is trashed we already update the lastReadMessageId of
all users in the channel to the latest non-deleted message on
the server side. However we didn't propagate this to the client,
so in some cases when we did the following:
1. Delete the last message in the channel
2. Switch to another channel
3. Switch back to the original
We would get a 404 error from the target message ID being looked
up still being the old lastReadMessageId (now deleted) for the
user's channel membership.
All we need to do is send the last not-deleted message ID for
the channel (or thread) to all the member users.
Before this commit the following actions would have shown the issue:
- visit a thread
- changes the width of the side panel
- open threads list
- the size has reverted to previous state
This was caused by the width change to not correctly be tracked.
It easier to check for presence in this case that to check for something not present, as depending on performance of the machine running the test this could take sometime to be changed and the test would fail.
This issue was especially visible in tests. the `@debounce(100)` was not cancelled when changing channel which was causing 404s as we were trying to load messages on a channel which was deleted as the channel has been destroyed at the end of the test.
This is still not a perfect solution, as we can only cancel the start of `fetchMessages`, but we can't cancel the actual `chatApi.channel` request which result can potentially happens after channel changed, which we try to mitigate with various checks on to ensure visible channel == loaded messages channel.
This commit also tries to make handler naming and cancelling more consistent.
<!-- NOTE: All pull requests should have tests (rspec in Ruby, qunit in JavaScript). If your code does not include test coverage, please include an explanation of why it was omitted. -->
This commit makes it easier to ensure each button have a similar behavior: hover, disabled, focus...
It also removes some dead code (__inline-button) and fixes a bug where the emoji button didnt have the right un-focused color (it was more visible than it should.
This commit makes some fundamental changes to how hashtag cooking and
icon generation works in the new experimental hashtag autocomplete mode.
Previously we cooked the appropriate SVG icon with the cooked hashtag,
though this has proved inflexible especially for theming purposes.
Instead, we now cook a data-ID attribute with the hashtag and add a new
span as an icon placeholder. This is replaced on the client side with an
icon (or a square span in the case of categories) on the client side via
the decorateCooked API for posts and chat messages.
This client side logic uses the generated hashtag, category, and channel
CSS classes added in a previous commit.
This is missing changes to the sidebar to use the new generated CSS
classes and also colors and the split square for categories in the
hashtag autocomplete menu -- I will tackle this in a separate PR so it
is clearer.
- few improved alignments
- displays emoji picker button inline on desktop
- keeps composer focused when focusing dropdown button
- align buttons to bottom when increasing height of textarea
- max-height of textarea is now linked to the height of the screen
Co-authored-by: chapoi <charlie@discourse.org>
This commit moves message lookup and querying to the
/chat/api/channel/:id endpoint and adds the ability
to query the tracking state overview for threads as well
as the threads and thread tracking state for any thread
original messages found.
This will allow us to get an initial overview of thread
tracking for a user when they first enter a channel, rather
than pre-emptively loading N threads and tracking state
for those across all channels on the current user serializer,
which would be expensive.
This initial overview will be used in subsequent PRs to
flesh out the thread unread indicators in the UI.
This also moves many chunks of code that were in services
to reusable Query classes, since use of services inside
services is discouraged.
Regressed in eec10efc3d. It means that backend plugin spec failures in CI were not failing the spec suite.
Fixes recent regressions and skips two of them - to be handled next week.
---------
Co-authored-by: Andrei Prigorshnev <a.prigorshnev@gmail.com>
Wrap scroll to bottom inside a next block to ensure the message has correctly been added to the array before actually attempt to scroll.
This commit also removes an unnecessary line which was essentially adding the message two times with no real consequences as we are uniq on ID.
- uses current user as user for fabricators, allows for correct avatar image and presence indicator
- uses a non existing channel ID to avoid setting a draft of an existing channel
- attempts to make color toggle more reliable
- Improves styleguide support
- Adds toggle color scheme to styleguide
- Adds properties mutators to styleguide
- Attempts to quit a session as soon as done with it in system specs, this should at least free resources faster
- Refactors fabricators to simplify them
- Adds more fabricators (uploads for example)
- Starts implementing components pattern in system specs
- Uses Chat::Message creator to create messages in system specs, this should help to have more real specs as the side effects should now happen
Followup from 9953a6edd9,
which broke an issue fixed in e8d6277062.
In the refactor we did not update the place where we resync
channel tracking based on onPresenceChange to use the new
tracking object we added to the serializer, and to directly
update channel.tracking
We had safe-area-inset-bottom still set at a wrong place which was causing issues but was also useful to some cases. This commit removes it and ensures the affected cases are corrected.
This is only the first steps of a redesign
- redesigns the buttons to have a larger hitzone
- generally bigger composer
- clicking near textarea focuses the input
- relies on the fact that safe-area-inset-bottom is set globally and doesn’t need to be set in sub components
This moves chat tracking state calculation for channels
and threads into a central Chat::TrackingStateManager service, that
serves a similar purpose to the TopicTrackingState model
in core.
This service calls down to these query classes:
* ThreadUnreadsQuery
* ChannelUnreadsQuery
To get the unread_count and mention_count for the appropriate
channels and threads.
As well as this, this commit refactors the client-side chat
tracking state.
Now, there is a central ChatTrackingStateManager Ember Service
so all tracking is accessible and can be counted from one place,
which can also initialize tracking from an initial payload.
The actual tracking counts are now maintained in a ChatTrackingState
class that is initialized on the `.tracking` property of both channel and
thread objects.
This removes the attributes on UserChatChannelMembership and decoration
of said membership from ChannelFetcher, preferring instead to have an additional
object for tracking in the JSON.
* DEV: add new thread icon
* FIX: Use new thread icon, fix typo in SVG
UX: move the thread list icon to the right of
the collapse button
---------
Co-authored-by: Martin Brennan <martin@discourse.org>
This commit also removes safe-area-inset-bottom when keyboard is displayed to avoid having a taller than needed space between composer and replying indicator.
Before this commit chat was applying a fixed height on everything under the `/chat` route. It's only really needed on the channel page with the composer at the bottom of the page.
This commits makes the following changes:
- moves height limitation from `#main-outlet-wrapper` to `.chat-channel`
- makes browse channel page and members list page full height and rely on main document scrollbar
- adds height computation for draft header and direct message creator block to ensure the height is correct when creating a draft channel
- makes chat index full height to rely on the browser scrollbar. As a result the <kbd> + </kbd> button used on mobile to create a direct message as been moved out of `<ChannelsList>` into the chat index template
- sidebar height was relying on chat setting a max height, as a result the height computation of sidebar has been changed to work correctly, especially with an opened keyboard on mobile or ipad
Hard to write a test for this behavior, this is a micro optimisation which doesn’t change the behavior but only makes it smoother by happening right before async request.
The failure screenshot shows the message is on screen while the error is:
```
Failure/Error: example.run
expected to find text "My favorite message" in "Community\nEverything\nMy Posts\nMore\nMessages\nInbox\nChannels\nPolitics 1\nPersonal chat\nPolitics 1". (However, it was found 1 time including non-visible text.)
```
I expect the arrow element might e slightly hiding the link, but not 100% sure of this.
When using `navigator.virtualKeyboard.overlaysContent = false` we can rely on using only the resize event. Also attempts to no over trigger `setProperty` when value didn't change.
A follow-up to 54b2a85b. That commit didn't fix the issue because the to_notify hash that we return from the notify_edit method isn't used anywhere apart from tests (that's confusing, we're going to fix that soon).
