It used to fail without displaying an error message if no upload
extensions were authorized. This also disables the button in the
first place to avoid displaying an error to the user (the error
will be displayed only when drag and dropping a file).
In the past, the result of template compilation would be stored directly in `Ember.TEMPLATES`. Following the move to more modern ember-cli-based compilation, templates are now compiled to es6 modules. To handle forward/backwards compatibility during these changes we had logic in `discourse-boot` which would extract templates from the es6 modules and store them into the legacy-style `Ember.TEMPLATES` object.
This commit removes that shim, and updates our resolver to fetch templates directly from es6 modules. This is closer to how 'vanilla' Ember handles template resolution. We still have a lot of discourse-specific logic, but now it is centralised in one location and should be easier to understand and normalize in future.
This commit should not introduce any behaviour change.
This change only affects the redesign user page navigation menu. The
dismiss button was incorrectly positioned below the user notifications
stream. A side effect of this is that it affected our infinite loading
code for the notifications stream.
No tests have been added in this commit as we have yet to quite figure
out how we can reliabily test for behaviours which requires us to scroll
the page.
Back when we introduced hashtag autocomplete in
c1dbf5c1c4 we had to
disallow triggering it using # at the start of the
line because our old markdown engine rendered headers
with `#abc`, but now our new engine does `# abc` so
it is safe to allow hashtag autocompletion straight
away.
When a client "reads" a post, we do no immediately send the data of the
post for processing on the server. Instead, read posts data is batched
together and sent to the server for processing at regular intervals. On
the server side, processing of read posts data is done in the
background. As such, there is a small window of delay before a post is
marked as read by a user on the server side.
If a client reads a topic and loads the messages topic list before the
server has processed the read post, the unread posts count for the topic
which the client just read will appear to be incorrect/outdated.
As part of tracking a post as read, we are already tracking the highest
read post number for the last read topic by the client. Therefore, we
can use this information to correct the highest post read number in the
scenario that was described above. This solution is the same as what
we've been doing for the regular topics list.
We were re-patching the deprecated bootbox methods during every app initialization, which would lead to hundreds of deprecation messages being printed for a single invocation.
When sidebar was enabled before going to narrow screen, it should be brought back when screen is enlarged.
Also, narrow screen value is changed to 1000px instead of 1100px.
- better handling of drawer state using chat state manager
- removes various float and topic occurrences to use drawer
- ensures user can chat before doing a lot of chat setup
- fixes a bug which was creating presence errors in tests
- removes dead code
The `decodedMap` prop comes from https://github.com/terser/terser/pull/1190
> This also exposes a new `decodedMap` property on the result object. Decoded maps are free to create (it's a shallow clone of the `GenMapping` instance), and passing them to `@jridgewell/trace-mapping` is copy-free. With Babel [recently](https://github.com/babel/babel/pull/14497) adding a `decodedMap` field, a dev could pass from the Babel transpilation to Terser without any added memory use for sourcemaps.
Opening of links in a new tab was difficult because it used a hack to
remove the 'href' attribute and adding it back to prevent the event
taking place instead of calling preventDefault.
This hack is no longer necessary because event handling has been
normalized since it has been implemented (see commit
0221855ba7).
Some plugins store their connectors under `{plugin}/assets/javascripts/templates/connectors`, which is read as `templates/connectors` relative to the base of the JS directory. Our connector checking logic was looking for strings including the leading slash (`/templates`), which not be the case here. Instead we can split on `/` and take the last element. This matches the logic we have for themes in https://github.com/discourse/discourse/blob/1dadf4381f/lib/theme_javascript_compiler.rb#L111
This wasn't actually breaking anything, so this is just a housekeeping commit.
In 1279966f we started namespacing modules based on the plugin's defined name rather than the directory name. This commit updates the argument name to match what we're passing in. This it just a readability change - there is no change in behaviour.
Adds the description as a title="" attribute on the hashtag
autocomplete search items for tags, categories, and channels.
These descriptions can be seen by the user since they are
able to see the results that are returned by the search via
Guardian checks.
This makes it so the new hashtags are not tracked,
same as the old ones. Also slight commenting in click-track
to explain mention clicks rejection mechanics.
Also deleted the single acceptance spec
since everything is covered better by the unit spec.
This commit unescapes the :emoji: and expands into
an image within hashtag autocomplete results, and
also makes some style tweaks to make sure the emoji
is not too big.
This fix changes the hashtag-raw hashtags, which are
the ones that do not actually match anything, back
to the old style which does not look like mentions.
The user attributes are not updated between clients and that is a
problem with user tips because the same user tip will be displayed
multiple times, once for every client.
Update `release_notes_link` to the current version.
<!-- 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 will be used by plugins to handle the client side of their custom
post validations without having to overwrite the whole composer save
action as it was done in other plugins.
Co-authored-by: Penar Musaraj <pmusaraj@gmail.com>
Our `triggerAction` backwards-compatibility was firing the action on
`parentView`. In most cases this worked, but it doesn't match the
classic behaviour when the DButton is included inside a 'wrapper'
component. In that case, the action should be triggered on the current
'this' context of the template that called the DButton.
This commit mimics the Ember Classic Component manager's behaviour. It
adds the `createCaller` capability to the custom component manager, and
then uses the `callerSelfRef` for dispatching the action.
We automatically refresh the page 'on the next navigation' whenever a
new version of the JS client is available. If the composer is open when
this happens then it will be closed and you'll have to reopen the draft.
In some circumstances, this refresh can also cause some composer content
to be lost.
This commit updates the auto-refresh logic so that it doesn't trigger
while the composer is open, and adds an acceptance test for the
behaviour.
<!-- 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. -->
Product has decided that the tracked section link provides very little
value at this moment in time so we're removing it.
See https://meta.discourse.org/t/245374 for more context.
The user preferences tracking page is only present when the redesign
user navigation menu is enabled. During the first pass of
implementation, some old bugs were introduced and this commit fixes
that. Regression tests have also been added.
Follow up from d3f02a1270
This commit fixes post quoting so that if the new
hashtag-cooked HTML is selected, we convert back to
a regular plain text #hashtag with the correct type and ref.
Bumps [jsdom](https://github.com/jsdom/jsdom) from 20.0.2 to 20.0.3.
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/jsdom/jsdom/releases">jsdom's
releases</a>.</em></p>
<blockquote>
<h2>Version 20.0.3</h2>
<ul>
<li>Updated dependencies, notably <code>w3c-xmlserializer</code>, which
fixes using <code>DOMParser</code> on XML documents containing
emoji.</li>
</ul>
</blockquote>
</details>
<details>
<summary>Changelog</summary>
<p><em>Sourced from <a
href="https://github.com/jsdom/jsdom/blob/master/Changelog.md">jsdom's
changelog</a>.</em></p>
<blockquote>
<h2>20.0.3</h2>
<ul>
<li>Updated dependencies, notably <code>w3c-xmlserializer</code>, which
fixes using <code>DOMParser</code> on XML documents containing
emoji.</li>
</ul>
</blockquote>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="22f7c3c518"><code>22f7c3c</code></a>
Version 20.0.3</li>
<li><a
href="c540630669"><code>c540630</code></a>
Update dependencies and dev dependencies</li>
<li><a
href="cdf07a1f0e"><code>cdf07a1</code></a>
Slight tweaks to GitHub Actions</li>
<li><a
href="bd77578169"><code>bd77578</code></a>
Try to make the issue template clearer</li>
<li>See full diff in <a
href="https://github.com/jsdom/jsdom/compare/20.0.2...20.0.3">compare
view</a></li>
</ul>
</details>
<br />
[](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)
Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.
[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)
---
<details>
<summary>Dependabot commands and options</summary>
<br />
You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot merge` will merge this PR after your CI passes on it
- `@dependabot squash and merge` will squash and merge this PR after
your CI passes on it
- `@dependabot cancel merge` will cancel a previously requested merge
and block automerging
- `@dependabot reopen` will reopen this PR if it is closed
- `@dependabot close` will close this PR and stop Dependabot recreating
it. You can achieve the same result by closing it manually
- `@dependabot ignore this major version` will close this PR and stop
Dependabot creating any more for this major version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this minor version` will close this PR and stop
Dependabot creating any more for this minor version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this dependency` will close this PR and stop
Dependabot creating any more for this dependency (unless you reopen the
PR or upgrade to it yourself)
</details>
Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
This commit fleshes out and adds functionality for the new `#hashtag` search and
lookup system, still hidden behind the `enable_experimental_hashtag_autocomplete`
feature flag.
**Serverside**
We have two plugin API registration methods that are used to define data sources
(`register_hashtag_data_source`) and hashtag result type priorities depending on
the context (`register_hashtag_type_in_context`). Reading the comments in plugin.rb
should make it clear what these are doing. Reading the `HashtagAutocompleteService`
in full will likely help a lot as well.
Each data source is responsible for providing its own **lookup** and **search**
method that returns hashtag results based on the arguments provided. For example,
the category hashtag data source has to take into account parent categories and
how they relate, and each data source has to define their own icon to use for the
hashtag, and so on.
The `Site` serializer has two new attributes that source data from `HashtagAutocompleteService`.
There is `hashtag_icons` that is just a simple array of all the different icons that
can be used for allowlisting in our markdown pipeline, and there is `hashtag_context_configurations`
that is used to store the type priority orders for each registered context.
When sending emails, we cannot render the SVG icons for hashtags, so
we need to change the HTML hashtags to the normal `#hashtag` text.
**Markdown**
The `hashtag-autocomplete.js` file is where I have added the new `hashtag-autocomplete`
markdown rule, and like all of our rules this is used to cook the raw text on both the clientside
and on the serverside using MiniRacer. Only on the server side do we actually reach out to
the database with the `hashtagLookup` function, on the clientside we just render a plainer
version of the hashtag HTML. Only in the composer preview do we do further lookups based
on this.
This rule is the first one (that I can find) that uses the `currentUser` based on a passed
in `user_id` for guardian checks in markdown rendering code. This is the `last_editor_id`
for both the post and chat message. In some cases we need to cook without a user present,
so the `Discourse.system_user` is used in this case.
**Chat Channels**
This also contains the changes required for chat so that chat channels can be used
as a data source for hashtag searches and lookups. This data source will only be
used when `enable_experimental_hashtag_autocomplete` is `true`, so we don't have
to worry about channel results suddenly turning up.
------
**Known Rough Edges**
- Onebox excerpts will not render the icon svg/use tags, I plan to address that in a follow up PR
- Selecting a hashtag + pressing the Quote button will result in weird behaviour, I plan to address that in a follow up PR
- Mixed hashtag contexts for hashtags without a type suffix will not work correctly, e.g. #ux which is both a category and a channel slug will resolve to a category when used inside a post or within a [chat] transcript in that post. Users can get around this manually by adding the correct suffix, for example ::channel. We may get to this at some point in future
- Icons will not show for the hashtags in emails since SVG support is so terrible in email (this is not likely to be resolved, but still noting for posterity)
- Additional refinements and review fixes wil
Uses `module()` instead of `discourseModule()`, native getters instead of `.get()`, fixes some assertions, uses the store instead of creating models directly
Long story short - bumping ember-qunit upgrades webpack, and sinon is not compatible with webpack v5. It uses node's `util` module and that is no longer polyfilled by webpack by default.
This PR adds the `util` polyfill (😑, but what can you do) and also injects a stub of `process` (so polyfill + sinon actually work)
Trying out changes to reduce the number of nav items in the experimental horizontal user nav. These changes should only appear with the redesigned_user_page_nav_enabled feature flag.
1. Created a new "Tracking" route. This combines some tracking-related settings from Notifications and Category and Tag tracking (which were separate tabs previously). Don't love the layout yet, but it's something that we can work on.
2. Moved some user-related settings out of Notifications and to the
Users tab. These seem more user-related to me, and it's nice that we can
associate enabling messages with the setting to limit who can send
messages.
3. Moved the App tab (lists app permissions) to be within the Security tab. It's very similar to Recently Used Devices.
* FIX: allow tl4 to bulk select
- Also allows tl4 to perform batch tagging
---
Long term this needs to be rewritten to account for "bulk action" permission
given from the server.
Co-authored-by: Martin Brennan <martin@discourse.org>
TrackedObject allows us to reference SiteSettings in autotracking contexts (e.g. JS getters referenced from a Glimmer template) without the need for EmberObject's `get()` function. TrackedObject is backwards-compatible with Ember's legacy reactivity model, so it can be referenced in things like computed properties.
Co-authored-by: David Taylor <david@taylorhq.com>
This change is intended to be backwards-compatible with all the previous arguments to `DButton`.
A deprecation warning will be triggered when a string is passed to the `@action` argument. This kind of action bubbling has been deprecated in Ember for some time, and should be updated to use closure actions.
Co-authored-by: Dan Gebhardt <dan@cerebris.com>
Co-authored-by: David Taylor <david@taylorhq.com>
When opening the invite acceptance page when the user
was already logged in, we were still showing the Accept
Invitation prompt even if the user had already redeemed
the invitation and was present in the `InvitedUser` table.
This would lead to errors when the user clicked on the button.
This commit fixes the issue by hiding the Accept Invitation
button and showing an error message instead indicating that
the user had already redeemed the invitation. This only applies
to multi-use invite links.
* FIX: Update user options only once
Performing actions that updated list of seen popups used to update user
options everytime instead of checking if the change has any effect.
* FIX: Load updated user data from response
node-fetch is now a ES module, so it has to either imported with `import/from` syntax (which can't be used in addon's index.js) or using the dynamic `import()`
Previously we were trying to handle both async and sync use cases in a single function, but it was confusing to read and led to subtle race conditions. This commit separates the async version into a separate function.
Previously we had a combination of a computed property and `this.set`. This was triggering the `computed-property.override` deprecation. This commit moves everything into the `dir` property, makes it a native getter, and adds a test to verify the reactive behavior.
Some locations in the app were `.set`-ing these computed properties. This would trigger the `computed-property.override` Ember deprecation, and also lead to inconsistency between the `..._categories_ids` property and the `...Categories` property.
This commit updates these properties to have getters/setters, with all state being stored in the `..._ids` property. The `@dependentKeyCompat` decorator is used to ensure these 'autotracking' getters can still be used as dependent keys in other computed properties.
- Count deprecations and print them to the console following QUnit runs
- In GitHub actions, write the same information as a job summary
- Add documentation to `discourse-common/lib/deprecated`
- Introduce `id` and `url` options to `deprecated`
- Introduce `withSilencedDeprecations` helper to allow testing deprecated code paths without making noise in the logs
This was previously reverted in 47035693b7.
* DEV: Add utility to hide all user tips
* DEV: Add UserTip Glimmer component
* DEV: Add tests for existing user tips
* FEATURE: Add user tip for post menu
* FEATURE: Add user tip for topic notification level
* FEATURE: Add user tip for suggested topics
* FEATURE: Hide new popups for existing users
This reverts commit 8c48285145. This introduced a bug which could cause sites to break when certain deprecations are hit. We'll re-introduce a fixed version of this change in a future commit.
* FEATURE: Default Composer Category Site Setting
- Create the default_composer_category site setting
- Replace general_category_id logic for auto selecting the composer
category
- Prevent Uncategorized from being selected if not allowed
- Add default_composer_category option to seeded categories
- Create a migration to populate the default_composer_category site
setting if there is a general_category_id populated
- Added some tests
* Add missing translation for the new site setting
* fix some js tests
* Just check that the header value is null
- Count deprecations and print them to the console following QUnit runs
- In GitHub actions, write the same information as a job summary
- Add documentation to `discourse-common/lib/deprecated`
- Introduce `id` and `url` options to `deprecated`
- Introduce `withSilencedDeprecations` helper to allow testing deprecated code paths without making noise in the logs
Previously `this.chatService.appEvents.on(
"chat:user-tracking-state-changed"...)` was registered on constructor and disabled on `willDestroy`. Constructor is evaluated only once, so when the section was collapsed and collapsed then the events were not observed anymore.
didInsert allows evaluating code each time a component is rendered.
Allows quick inline replies in chat push notifications. This will allow users
in compatible platforms (Windows 10+ / Chrome OS / Android N+) to reply
directly from the notification UI.
Probable follow ups include:
- inline replies for posts
- handling failure of reply
- fallback to draft creation if business logic error
- store and try again later if connectivity error
- sent inline replies lack the in_reply_to param
- i18n of inline reply action text and placeholder
This implementation attempts to be more resilient to background tab.
Notes:
- adds support for immediate arg in @debounce decorators
- fixes a bug in discourseDebounce which was not supporting immediate arg in tests
- chat-audio-manager has no tests as audio requires real user interaction and is hard to test reliably
Connector actions are already added as properties of the generated component, but they were not bound. Using them like `{{on "click" this.someAction"}}` and trying to access `this` would not work as expected. This commit binds all actions to the component generated component instance.
Currently, we have available three 2fa methods:
- Token-Based Authenticators
- Physical Security Keys
- Two-Factor Backup Codes
If the first two are deleted, user lose visibility of their backup codes, which suggests that 2fa is disabled.
However, when they try to authenticate, the account is locked, and they have to ask admin to fix that problem.
This PR is fixing the issue. User still sees backup codes in their panel and can use them to authenticate.
In next PR, I will improve UI to clearly notify the user when 2fa is fully disabled and when it is still active.
- allows to scroll while hovering the menu
- correctly changes message background color while hovering menu
- prevents a bug where it would sometimes close the menu while moving from menu to the 3 dots expanded dropdown. This was caused by the gap between header/body of the 3 dots dropdown, which would sometimes allow to create a mouseover event on a possible different underlying message
- removes recent/favorite reactions on drawer mode
- grayscale reactions until hover
- boxshadow on msgactions container
- removes useless code
This reverts commit 136174e0ee.
That worked only 50% of the time (at best), as it was conflicting with other topic-view scrolling code. The reverted feature will eventually be restored as I continue to fix scroll-related issues.
Our method of loading a subset of client settings into tests via
tests/helpers/site-settings.js can be improved upon. Currently we have a
hardcoded subset of the client settings, which may get out of date and not have
the correct defaults. As well as this plugins do not get their settings into the
tests, so whenever you need a setting from a plugin, even if it has a default,
you have to do needs.setting({ ... }) which is inconvenient.
This commit introduces an ember CLI build step to take the site_settings.yml and
all the plugin settings.yml files, pull out the client settings, and dump them
into a variable in a single JS file we can load in our tests, so we have the
correct selection of settings and default values in our JS tests. It also fixes
many, many tests that were operating under incorrect assumptions or old
settings.
Co-authored-by: Joffrey JAFFEUX <j.jaffeux@gmail.com>
Even with the `chunkFilename` change, the sourcemaps are non-deterministic because they include references to the broccoli cache directory which has a different name for each build. This commit disables auto-import sourcemaps in production to improve caching performance.
Followup to 3673d3359c
Previously if a specific plugin route was not available (e.g. there was an error loading the plugin's JS due to an ad blocker), the entire page would fail to load. This commit updates the behavior to catch this kind of issue and display a user-friendly message at the top of the screen.
Previously, we didn't have a site-wide setting to set the default behavior for user profile visibility and user presence features. But we already have a user preference for that.
Repro steps:
- enable permanent deletes (via hidden site setting)
- set `min_topic_views_for_delete_confirm` to 0
When permanently deleting, the delete confirm modal is shown (for a
second time) and it doesn't pass the `force_destroy` parameter to the
request and the action results in a 422 error (i.e. can't perma-delete).
This change skips showing the confirm modal when perma-deleting given
that it has already been show on the first delete action.
Passing a string action name to `DButton` causes it to use `sendAction`, which is deprecated and will be removed in Ember 4.x. The action helper converts a string to a closure action.
This also fixes compatibility with https://github.com/discourse/discourse/pull/17767
There are two possible ordering for categories shown in sidebar with
this commit.
When the `fixed_category_positions` site setting is enabled, the
categories are ordered based on `Category#position` which is a configurable
option by the user. When said site setting is disabled, the categories
are ordered based on `Category#name`.
The categories in Sidebar are also sorted in such a way where child
categories are always ordered right after their parents. When multiple
child categories are present, the child categories are ordered based on
the ordering described above.
Previously these were set to expire after 9999 days (27 years). This commit updates them to last 1 year, and to automatically be extended on every user visit.
This updates the behavior of the list destination setting for links in the sidebar.
By default, new/unread content will show a dot like chat, rather than the count of new/unread topics.
If a user chooses to link to new/unread in the sidebar, we'll show the count.
The goal here is to find a simple default for typical users (new/unread indication, no counts, default links) while providing a different workflow for power users (showing new/unread counts, and linking directly to new/unread).
Internal Ref: /t/82626
`siteSettings` is now a service which means there should only be one
state for `siteSettings` during the life time of the application. This
also helps to maintain parity with production where the `site` model
relies on the `siteSettings` service and not a clone of the attributes.
Previously when a topic had e.g. 10 posts and you read them all, the link to the "first unread" would be `/11`, even when we knew there are only 10. (the topic route/controller would then fix that in the location bar after a second if you followed that URL)
This commit fixes a bug on the client site where we would include the
`regular_category_ids` field when trying to update the notification levels of
categories for a user. The `regulary_category_ids` field should only be
included when the `mute_all_categories_by_default` is enabled
Our dialog service doesn't accept HTML by default and we shouldn't include HTML in the error message string. And given that the Ajax error handler is called in multiple contexts, it's tricky to properly support line breaks via either HTML or `\n` so we are opting for plain text in AJAX error messages.
A callback that's provided as a string, such as `{{action "doSomething"}}`, may target the method `doSomething` on the context OR the context's `action` hash (if it exists).
Having the group name in the `class` attribute can cause a clash with 'real' CSS classes. Putting it in a data attribute is much safer, and can still be targetted via CSS if desired.
The previous sidebar default tags and categories implementation did not
allow for a user to configure their sidebar to have no categories or
tags. This commit changes how the defaults are applied. When a user is being created,
we create the SidebarSectionLink records based on the `default_sidebar_categories` and
`default_sidebar_tags` site settings. SidebarSectionLink records are
only created for categories and tags which the user has visibility on at
the point of user creation.
With this change, we're also adding the ability for admins to apply
changes to the `default_sidebar_categories` and `default_sidebar_tags`
site settings historically when changing their site setting. When a new
category/tag has been added to the default, the new category/tag will be
added to the sidebar for all users if the admin elects to apply the changes historically.
Like wise when a tag/category is removed, the tag/category will be
removed from the sidebar for all users if the admin elects to apply the
changes historically.
Internal Ref: /t/73500
Classic Ember components (i.e. "@ember/component") rely upon "event
delegation" to listen for events at the application root and then dispatch
those events to any event handlers defined on individual Classic components.
This coordination is handled by Ember's EventDispatcher.
In contrast, Glimmer components (i.e. "@glimmer/component") expect event
listeners to be added to elements using modifiers (such as `{{on "click"}}`).
These event listeners are added directly to DOM elements using
`addEventListener`. There is no need for an event dispatcher.
Issues may arise when using Classic and Glimmer components together, since it
requires reconciling the two event handling approaches. For instance, event
propagation may not work as expected when a Classic component is nested
inside a Glimmer component.
`normalizeEmberEventHandling` helps an application standardize upon the
Glimmer event handling approach by eliminating usage of event delegation and
instead rewiring Classic components to directly use `addEventListener`.
Specifically, it performs the following:
- Invokes `eliminateClassicEventDelegation()` to remove all events associated
with Ember's EventDispatcher to reduce its runtime overhead and ensure that
it is effectively not in use.
- Invokes `rewireClassicComponentEvents(app)` to rewire each Classic
component to add its own event listeners for standard event handlers (e.g.
`click`, `mouseDown`, `submit`, etc.).
- Configures an instance initializer that invokes
`rewireActionModifier(appInstance)` to redefine the `action` modifier with
a substitute that uses `addEventListener`.
Additional changes include:
* d-button: only preventDefault / stopPropagation for handled actions
This allows unhandled events to propagate as expected.
* d-editor: avoid adding duplicate event listener for tests
This extra event listener causes duplicate paste events in tests.
* group-manage-email-settings: Monitor `input` instead of `change` event for checkboxes
If there's an error loading a topic, and later you load it successfully - the the stale error message would still briefly show up every time you navigate to that topic
This reverts commit 28be5d3037 and fcb4675415
This caused qunit timeouts in our internal CI environments. Not sure of the exact cause yet, but we're reverting for now while we investigate.
In test mode we reinitialize the Application for every test. We only want to apply the class reopens once to avoid performance regressions and memory leaks in the test suite.
Classic Ember components (i.e. "@ember/component") rely upon "event
delegation" to listen for events at the application root and then dispatch
those events to any event handlers defined on individual Classic components.
This coordination is handled by Ember's EventDispatcher.
In contrast, Glimmer components (i.e. "@glimmer/component") expect event
listeners to be added to elements using modifiers (such as `{{on "click"}}`).
These event listeners are added directly to DOM elements using
`addEventListener`. There is no need for an event dispatcher.
Issues may arise when using Classic and Glimmer components together, since it
requires reconciling the two event handling approaches. For instance, event
propagation may not work as expected when a Classic component is nested
inside a Glimmer component.
`normalizeEmberEventHandling` helps an application standardize upon the
Glimmer event handling approach by eliminating usage of event delegation and
instead rewiring Classic components to directly use `addEventListener`.
Specifically, it performs the following:
- Invokes `eliminateClassicEventDelegation()` to remove all events associated
with Ember's EventDispatcher to reduce its runtime overhead and ensure that
it is effectively not in use.
- Invokes `rewireClassicComponentEvents(app)` to rewire each Classic
component to add its own event listeners for standard event handlers (e.g.
`click`, `mouseDown`, `submit`, etc.).
- Configures an instance initializer that invokes
`rewireActionModifier(appInstance)` to redefine the `action` modifier with
a substitute that uses `addEventListener`.
Additional changes include:
* d-button: only preventDefault / stopPropagation for handled actions
This allows unhandled events to propagate as expected.
* d-editor: avoid adding duplicate event listener for tests
This extra event listener causes duplicate paste events in tests.
* group-manage-email-settings: Monitor `input` instead of `change` event for checkboxes
The complex regex-based detection was based on Ember CLI's implementation, which is necessarily generic and needs to auto-detect the name. In our case, we know the name of the plugin so we can just pass it in - no need for dynamic detection. This resolves issues when there are other files in the `discourse/plugins/{name}` directory which are sorted before `discourse/`
Lone hbs files in the `/components` are automatically assumed to be template-only Glimmer components. The `templateOnly()` stub is only required when templates are in the `/templates/components` directory.
The clientside allowPersonalMessages function introduced
in e62e93f83a sometimes did not
work correctly, because the currentUser.groups property
only contained **visible** groups for the current user, which
could exclude auto groups that had their permissions set to
be owner-only visible.
It was unnecessary to add this anyway since we already have
can_send_private_messages on the CurrentUserSerializer. It's
better the backend does this calculation anyway. Use that
in the clientside code instead and get rid of allowPersonalMessages
Given that the category structure is generally speaking the backbone of most Discourse instances, it makes sense to show the edit button for the category even when the user is in a category/tag intersection route.
Debouncing inline anonymous functions does not work.
This fixes all instances of that error by extracting the function or using the new `@debounce(delay)` decorator
Our theme system injects a magical `settings` object at the top of themes JS modules to allow theme authors to access the settings as configured by admins in the UI. Within this `settings` object, there are a couple of special objects `theme_uploads` and `theme_uploads_local` that contain URLs for all the assets/uploads that the theme has.
For test modules/files, the theme system also injects a `settings` object at the top of tests modules, but it's not the same object as the object that's injected in non-test files. The difference is that in tests we want the settings to have their default values as opposed to any custom values that may exist in the site's database. This ensures that test results are consistent no matter the site that runs them.
However, the `settings` object in tests files currently doesn't have the special objects `theme_uploads` and `theme_uploads_local` which means that if a theme includes an asset that's lazy-loaded, it's not possible to write tests for anything that depends on the lazy-loaded asset because the theme will not be able to load the asset during the tests since `theme_uploads_local` and `theme_uploads` don't exist. This PR adds these special objects inside the `settings` object for test files.
Internal topic: t/71825/52.
Multiple things in the app need the height of the header to be correct (for example, scrolling to a post), so we need the header offset calculation. However, we shouldn't be calculating it on scroll, it's too resource intensive and it causes flickering on iPads (and possible other devices too).
This commit removes header offset calculation on scroll and adds a one-time calculation as soon as the header is first rendered. This ensures that users get scrolled to the correct post even if they open it in a new tab.
This commit adds a new `/hashtag/search` endpoint and both
relevant JS and ruby plugin APIs to handle plugins adding their
own data sources and priority orders for types of things to search
when `#` is pressed.
A `context` param is added to `setupHashtagAutocomplete` which
a corresponding chat PR https://github.com/discourse/discourse-chat/pull/1302
will now use.
The UI calls `registerHashtagSearchParam` for each context that will
require a `#` search (e.g. the topic composer), for each type of record that
the context needs to search for, as well as a priority order for that type. Core
uses this call to add the `category` and `tag` data sources to the topic composer.
The `register_hashtag_data_source` ruby plugin API call is for plugins to
add a new data source for the hashtag searching endpoint, e.g. discourse-chat
may add a `channel` data source.
This functionality is hidden behind the `enable_experimental_hashtag_autocomplete`
flag, except for the change to `setupHashtagAutocomplete` since only core and
discourse-chat are using that function. Note this PR does **not** include required
changes for hashtag lookup or new styling.
Before, `sidebar_list_destination` was an attribute on UserOptionSerializer. The problem was that this attribute was added to user model only when the user entered the preferences panel. We want that attribute to be available all the time, therefore it was moved to CurrentUserSerializer.
This view can show multiple posts from the same topic and the aria labels
will now include the post number to more easily differentiate posts in
screen readers.
Theme javascript is now minified using Terser, just like our core/plugin JS bundles. This reduces the amount of data sent over the network.
This commit also introduces sourcemaps for theme JS. Browser developer tools will now be able show each source file separately when browsing, and also in backtraces.
For theme test JS, the sourcemap is inlined for simplicity. Network load is not a concern for tests.
Normally, arguments passed to components are lazily evaluated. `get prefixElementColors` will only be evaluated for `@prefixType="span"`. However, when using the Ember Inspector in development, arguments are eagerly evaluated and their values displayed in the inspector. Therefore we need to make sure that getters can always be evaluated without exceptions being thrown.
This allows plugins to colocate component JS and HBS under `/plugins/{name}/assets/javascripts/discourse/components`.
`discourse-presence` is updated to use this new pattern, which also serves as an integration test for this part of the build pipeline.
Ember's default resolver only looks for components/services/etc. which are namespaced under the app's `modulePrefix` (`discourse`, in our case). To use addon components/services/etc., the addon must re-export them in its `app/` directory.
In order to support plugins, our custom resolver does a 'suffix match'. This has an unintended side-effect of matching things which are not part of the discourse app or themes/plugins. We've come to rely on this for a few in-repo addons like `select-kit`, `admin` and `wizard`.
This unrestricted 'suffix matching' can cause some very unexpected behaviour. For example, the ember-inspector browser extension has a module called `ember_debug/service/session`. When looking up `service:session`, our resolver was choosing that third-party service over our own Session service. This means Discourse fails to boot when the Ember Inspector is open.
This commit restricts the 'suffix matching' to a known set of namespaces. This brings us one step closer to the default Ember Resolver implementation, and reduces the chance of unexpected behaviour like the ember-inspector issue.
This commit also updates the `dialog-holder` addon to export its service under the app directory, so that we don't need to account for it in the resolver. We may want to consider doing the same for things like `select-kit` and `truth-helpers`, but is beyond the scope of this commit.
Displays a sidebar section link to admin users when
`default_sidebar_categories` site setting has not been configured for the
site.
Internal Ref: /t/73500
Tab order acts strangely in Chrome when the last focusable element in a
modal is a radio group: it switches focus to the address bar. This is a
problem, because for keyboard users, it becomes very hard to return to
the previous context.
This PR adds a focusable "Cancel" button, whose mere presence fixes the
issue.
