Using the owner of `site` is not perfectly reliable, especially given that `site` is stubbed in tests. Instead, we can fetch the owner of the `context` object itself.
Also, Ember sets the owner of an EmberObject based on the first parameter to `create`. This is preferable to a separate `setOwner` call because it means the owner is available during initialization.
Unfortunately we don't have any way to compile raw templates inline during tests, so testing this behavior in core is very difficult. Given that we aim to remove the raw handlebars system in the not-too-distant future, it doesn't make sense to invest lots of time on tooling here. Regressions of this behavior will be detected by theme CI tests [like this](https://github.com/discourse/discourse-topic-thumbnails/pull/21).
This commits introduces a new SiteSetting.enable_new_user_profile_nav_groups
feature flag. When configured, users of the configured groups will see
the new user page navigation links.
As of this commit, only the user activity navigation link has been
converted to the newly proposed dropdown of navigation links.
Mobile support has not been considered.
Previously we were calculating both the minimum and maximum widths for
SK dropdowns using this Popper modifier. The max. width calculation was
causing issues with dropdowns in Firefox and was also sluggish when
rendering.
This switches to using CSS calculations for max. widths. It adds a 600px
global maximum and targeted maximums for the category composer dropdown
and the bookmark list dropdowns.
Highlight.js 11 deprecated the feature to highlight HTML blocks while
keeping the HTML structure, which broke our GitHub onebox syntax
highlight.
This patch adds it back by bringing the maintainers code as a plugin.
See https://github.com/highlightjs/highlight.js/issues/2889
This commit includes various accessibility improvements for the new user menu:
* Add `title` attributes to the user menu tabs
* Properly label lists (by adding `aria-labelledby` to `<ul>` elements) for screen readers
* Change the user menu structure so that the tabs come before the content panel in the DOM, but use CSS to reverse them visually.
Normally, changing the order of elements via CSS is bad for accessibility, but I believe this is one of the rare scenarios where it [makes sense](https://developer.mozilla.org/en-US/docs/Web/CSS/CSS_Flexible_Box_Layout/Ordering_Flex_Items#use_cases_for_order). Prior to this change, if you want to reach the first notification item after you select a tab using the keyboard, you have to hit <kbd>ctrl</kbd>+<kbd>tab</kbd> because the notifications list is before the tabs list. However, with this change, <kbd>tab</kbd> will move you to the first item in the list after you select a tab using your keyboard.
* Aria-hide the unread notifications badge/count on the tabs because the `title` attribute on the tab indicates the unread count.
* Add some tests.
In itself, this change will not cause tests to run in parallel. It just unlocks the ability to use tools like `ember exam` to run tests in parallel. For example:
```
yarn ember exam --load-balance --parallel=3 --random
```
We previously had a system which would generate a 10x10px preview of images and add their URLs in a data-small-upload attribute. The client would then use that as the background-image of the `<img>` element. This works reasonably well on fast connections, but on slower connections it can take a few seconds for the placeholders to appear. The act of loading the placeholders can also break or delay the loading of the 'real' images.
This commit replaces the placeholder logic with a new approach. Instead of a 10x10px preview, we use imagemagick to calculate the average color of an image and store it in the database. The hex color value then added as a `data-dominant-color` attribute on the `<img>` element, and the client can use this as a `background-color` on the element while the real image is loading. That means no extra HTTP request is required, and so the placeholder color can appear instantly.
Dominant color will be calculated:
1. When a new upload is created
2. During a post rebake, if the dominant color is missing from an upload, it will be calculated and stored
3. Every 15 minutes, 25 old upload records are fetched and their dominant color calculated and stored. (part of the existing PeriodicalUpdates job)
Existing posts will continue to use the old 10x10px placeholder system until they are next rebaked
Upgrading to Markdown.it v13 broke empty inline BBCodes. This works around the problem by adding an empty token before a closing token if the previous token was a BBCode token.
It also removes the unused `jump` attribute which was removed in Markdown.it v12.3
The experimental user menu has a tab that displays recent reviewables and at the moment when a new signs up for the site and they need to be approved, admins see a very scary "suspicious user" copy in the reviewables tab in the user menu. We don't need the copy to be very scary because when a user needs to be approved, it's because the site operator has configured the site to force all new users to go through the review queue and it's not some kind of spam detector flagging the user.
The key fix in this commit is that it removes `this.replaceState(path)` for anchor-only URLs. We still intercept those routing changes to properly calculate the scroll position of the anchor via `jumpToElement`, but we no longer use the Ember router to override the browser's history. This fixes the subfolder issue and also lets the browser maintain its history correctly.
The commit also includes a small refactor to the `jumpToElement` helper to facilitate stubbing in tests.
Secure media requests go through the app. In topics with many images, this makes it very easy to hit rate limiters. Skipping the low-res placeholders reduces the chance of this problem occuring.
Reverts #18241 and fixes issues with the original PR:
1. Remove an extraneous `margin-left: auto` from a grid cell (this was causing the buggy behavior in webkit)
2. Add `grid-area` name to `.extra-info-wrapper`
3. Account for `.wrap` padding
4. Remove unused css (`.header-row` and inner styles)
This PR restores a small feature which was present in the old menu and allowed users to click on the active tab in the menu to navigate to some page that showed the same items in the menu but with more details.
For example, if you switch to the PMs tab and then click on it again, currently nothing happens. However, with this change, clicking on the tab again will take you to your messages page at `/my/messages`.
Note: plugins that register custom tabs in the menu can provide a `linkWhenActive` property for their tab if they wish to mimic core's tabs, but it's optional; if they don't provide one, the tab will do nothing if the user clicks on it again.
Internal topic: t/73349.
Because Discourse is a single-page application, clicks on the majority of `<a>` elements in the app need to be intercepted by JavaScript to prevent browsers' default action (full page reload). Links in the user menu - which include notifications, reviewables, bookmarks etc. - are no exception to this rule and currently clicks on these items are handled by the global [click-interceptor](1fa21ed415/app/assets/javascripts/discourse/app/lib/intercept-click.js (L20)) which calls the `preventDefault` function on the click event object and uses the `DiscourseURL.routeTo` function to route the user to the page they request.
However, for links in the user menu, there's an extra step which is to let the header know that it should close the user menu after clicking an item in the menu, but the global interceptor doesn't know that because the step is specific to links in the user menu. This can cause a bug on mobile devices where the menu remains open after clicking on a notification which results in the user having to close the menu to see the page that the notification takes them to.
This commit adds a click handler to user menu items that ensures the menu is closed when an item is clicked and navigates the user to wherever the item links to. There's a small downside to this change which is that user menu items now have their own click interceptor instead of relying on the global interceptor, i.e. duplicated logic, but since it's only a couple of lines, I think we can live with it for a while.
I did try to make the click handler of the user menu items only close the menu (call the `closeUserMenu` function), but for some reasons it caused a full page reload to happen when clicking a notification item due to some weird interactions between the header widget and the user menu. I didn't debug this thoroughly because we have plans to change the header implementation from widgets/virtual-dom to Glimmer component, which will likely resolve that weird full page reload issue and we'll be able to make the click handler just close the menu and let the global interceptor prevents the default action and do the routing.
Internal topic: t/71911/118.
This PR changes the icon for `posted` notification types (these are the notifications that you receive when someone posts in a topic you're watching) from `reply` to `discourse-bell-exclamation`. We're doing this to visually distinguish between the `posted` notifications and `replied` notifications which are the notifications that you receive when someone replies to you directly.
Internal topic: t72835.
In Safari, `img.complete` is sometimes true even before the image is loaded. Checking for the presence of `img.naturalHeight` seems to be more reliable. It is very difficult to write a test for this behaviour due to the dependence on network conditions, scroll location, etc.
`img.naturalHeight` is supported by all our target browsers, and I have verified the functionality of this commit in Chrome, Safari and Firefox.
This updates the ember-cli plugin detection logic to match the logic in `Plugin::Instance.find_all`. It will now ignore plugin directories which do not have a `plugin.rb` file.
This commit adds to the experimental user menu a new "other notifications" tab that's very similar to the "all notifications" tab, but with the main difference being that it doesn't show notification types that do have dedicated tabs in the menu (e.g. mentions, likes, replies etc.).
The rationale behind this is that the notification types that do have dedicated tabs tend to dominate the "all notifications" tab, leaving very small chances for the user to notice rarer or infrequent notification types. Adding a tab for all the other types gives the user a way to review those infrequent notification types.
Internal ticket: t72978.
Co-authored-by: OsamaSayegh <asooomaasoooma90@gmail.com>
We were using the directory name rather than the plugin's defined `name`. This was an unintended change in behaviour from the old sprockets implementation. This commit makes the ember-cli naming logic match the old sprockets logic.
If the server is configured to use a view like `categories_and_latest_topics`, it will preload data for the topic list. If the client doesn't actually use it (e.g. on mobile), then that preloaded data would remain cached and be used for the next loaded topic list. This commit ensures it's always removed, and adds a test for the behaviour.
https://meta.discourse.org/t/237126/35
Previously we were relying on a highly-customized version of the unmaintained Barber gem for theme template compilation. This commit switches us to use our own DiscourseJsProcessor, which makes use of more modern patterns and will be easier to maintain going forward.
In summary:
- Refactors DiscourseJsProcessor to move multiline JS heredocs into a companion `discourse-js-processor.js` file
- Use MiniRacer's `.call` method to avoid manually escaping JS strings
- Move Theme template AST transformers into DiscourseJsProcessor, and formalise interface for extending RawHandlebars AST transformations
- Update Ember template compilation to use a babel-based approach, just like Ember CLI. This gives each template its own ES6 module rather than directly assigning `Ember.TEMPLATES` values
- Improve testing of template compilation (and move some tests from `theme_javascript_compiler_spec.rb` to `discourse_js_processor_spec.rb`
The generate RSA key and import theme routes worked separate from each
other. The RSA key returned both the public and private key and it was
the frontend which posted the private key back to the server. With this
commit, only the public key is necessary as the server keeps a map of
public and private keys that is used to get the private key back from
a public key.
It used to return the next URL anyway which lead to an additional
request. On the frontend, if the result set was empty, it kept retrying
until at least one result was returned. This bug is fixed in this commit
too.
`popupAjaxError` is used by the controller to handle errors when saving the badge model. That only works properly when it gets the original exception from the ajax call. This manipulation of the error in the badge model was breaking that behavior.
The admin-badges controller is the only place which saves this model, so this shouldn't cause any issues elsewhere in the app.
This lets us use all our normal JS tooling like prettier, esline and babel on the splash screen JS. At runtime the JS file is read and inlined into the HTML. This commit also switches us to use a CSP hash rather than a nonce for the splash screen.
* FIX: hide welcome topic banner as soon as the welcome topic is edited
This commit adds a message bus listener on client to hide the welcome
topic banner as soon as the welcome topic is edited.
* update test
* only subscribe when show_welcome_topic_banner is true
* Do not lookup for messageBus service if it's not required
* Remove unneeded code
* Cache result for Site.show_welcome_topic_banner
* Update tests per latest changes
* Changes per PR review
Each new user menu notifications should have their own count. Therefore, we need to include all types to serializer and not only `grouped_unread_high_priority_notifications`
Additional PR will be created for chat and assign plugin, as they will have to switch to `grouped_unread_notifications` as well.
When installing private themes and theme components, the public key does
not show until the administrator types a valid Git repo URL. The regular
expression that checked the URL was too strict and it required the URL
to end with ".git".
In eb12daa7f8 when adding community
section support for anonymous users, we changed the `sectionLinks`
property into a getter method. This meant that if the getter method was
called again after the community section has been rendered, we would end
up reintializing the section links classes. As part of the
initialisation, some section links would setup a TopicTrackingState
onStateChange listener. However, the listener is only removed when the
entire community section is removed which resulted in us leaking the
onStateChange listeners.
This commit reverts the `sectionLinks` from being defined as a getter
method into a property which is only set once when the community section
is being constructor. Also, we changed it such that the community
section will register the listener instead of each section link since it
makes cleaning up much easier to reason about.
No tests have been added for this commit because the original bug is
not possible after this change and we already have an existing tests
ensuring that TopicTrackingState change listeners are cleaned up when
the community section is destroyed.
Internal ref: /t/73224
When you receive a new notification, Discourse prepends a small count `(n)` to the tab title (i.e. `document.title`) if the tab is in the background to alert the user that they have a new notification. The count that's shown in the tab title should reflect the numbers shown on the notification bubbles above the user's avatar. Prior to the experimental user menu, there were 2 bubbles: a blue one which was removed once the user opened the menu and a green one that indicated high priority notifications and it was only removed when the user read all of their high priority notifications.
In the new experimental user menu, we no longer have the green bubble; everything is now combined (including flags/reviewables) into the blue one with no change to its behavior (i.e. it's removed once the user opens the menu). However, the logic that is responsible for updating the tab title hasn't been updated and still updates the tab title to include the count of the old green bubble. This commit updates the logic for the tab title count so that it only reflects the number on the blue bubble when the experimental user menu is enabled.
This commit makes a number of improvements to the DiscourseJsProcessor:
1. Remove dependence on the out-of-date Ember template compiler from the ember-rails gem; switch to modern template compiler
2. Refactor to make use of a proper module system with `define`/`require`
3. Introduce `babel-plugin-ember-template-compilation` to enable inline hbs compilation
The `mini-loader` is upgraded to support relative lookup and `require.has`, so that these new JS packages work correctly.
We were already compiling the markdown bundle via ember-cli, but that version was only being used in the test environment. This commit improves the implementation, and updates the filename so it's also used in production.
This commit also
- Removes the vendored copy of `markdown-it.js` and fetches from node_modules instead
- Updates `pretty_text.rb` to remove the custom sprockets-manifest-parsing
- Removes `pretty-text-bundle.js`, which was only being used by `pretty_text.rb`
This adds a new framework for accessible dialogs that will eventually replace bootbox. Under the hood, it uses the a11y-dialog package and an in-repo Ember addon. See PR for usage details.
The old logic did not make sense and hid the selector from regular users
even if they could tag PMs or showed selector for admins even if they
could not tag PMs.
The preload key was changed in e7a84948b9 but this location was missed. This caused an extra AJAX request and left the cached topic list in the PreloadStore, which would then be accidentally used when navigating to the next topic-list route.
This commit extends the plugin API introduced in 40fd82e2d1 to the `Bookmark` and `Notification` models. It also refactors the code that's responsible for loading items in the experimental menu to use `async`...`await` instead of `Promise`s.
This API allows plugins to transform a list of model objects before they're rendered in the UI. At the moment, this API is limited to items/lists of the experimental user menu, but it may be extended in the future to other parts of the app.
Additional context can be found in https://github.com/discourse/discourse/pull/18046.
The previous sprockets implementation was including admin-specific JS in the plugin's main JS file, which would be served to all users regardless of admin status. This commit achieves the same result under the ember-cli plugin asset compiler with one difference: the admin js is compiled into a separate file. That means that in future, we'll be able to make it loaded only for admins. For now though, it's loaded for everyone, just like before.
When preloading topic_list data we were giving it a 'preload key' which was loosely based on the parameters of the list. However, it did not include all parameters, and mismatches between client/server-side logic would cause the preloaded data to be ignored.
This commit simplifies things by using a single key for all topic_list preloading. This works on the assumption that "The first topic_list the JS app will load is the one which was preloaded". That assumption also existed to some extent in the old design, so we don't expect any regressions here.
Omitting the flag from optional-features enables the runtime deprecation notice.
Also introduces `ember-jquery-legacy` which can be used to migrate to the new behaviour early. Details at https://deprecations.emberjs.com/v3.x/#toc_jquery-event
Core does not appear to make use of `originalEvent` in Ember event handlers. When searching for `originalEvent` there are some matches which relate to our pan-events mixin, but this is our own implementation and not affected by this deprecation.
Plugins often change core behavior, and thereby cause core's tests to fail. In CI, we work around this problem by running core CI without any plugins loaded.
In development, the only option to safely run the core tests is to uninstall all plugins, which is clearly a bad developer experience. This commit aims to improve that experience.
The `qunit_skip_plugins=1` flag would previously prevent the plugin **tests** from running. This commit extends that flag to also affect the plugin's application JS.
Default sidebar tags for not authenticated users can be defined in admin panel. Otherwise, top 5 categories and tags are taken.
Optionally, if categories are set up in permanent order, then the first 5 categories are taken.
`#main` in the test environment is replaced with `#ember-testing`, so this code would break. It never did only because we don't test these code paths 👀
The logic was added in commit ec8306835d,
to show the like action even if the user could not like the post. It is
not necessary for this logic to be implemented on the server side.
* FIX: Do not allow to remove like if topic is archived
* FIX: Always show like button
The like button used to be hidden if the topic was archived and it had
no likes. This commit changes that to always show the like button, but
with a not-allowed cursor if the topic is archived.
When `EMBER_CLI_PLUGIN_ASSETS=1`, plugin application JS will be compiled via Ember CLI. In this mode, the existing `register_asset` API will cause any registered JS files to be made available in `/plugins/{plugin-name}_extra.js`. These 'extra' files will be loaded immediately after the plugin app JS file, so this should not affect functionality.
Plugin compilation in Ember CLI is implemented as an addon, similar to the existing 'admin' addon. We bypass the normal Ember CLI compilation process (which would add the JS to the main app bundle), and reroute the addon Broccoli tree into a separate JS file per-plugin. Previously, Sprockets would add compiled templates directly to `Ember.TEMPLATES`. Under Ember CLI, they are compiled into es6 modules. Some new logic in `discourse-boot.js` takes care of remapping the new module names into the old-style `Ember.TEMPLATES`.
This change has been designed to be a like-for-like replacement of the old plugin compilation system, so we do not expect any breakage. Even so, the environment variable flag will allow us to test this in a range of environments before enabling it by default.
A manual silence implementation is added for the build-time `ember-glimmer.link-to.positional-arguments` deprecation while we work on a better story for plugins.
When there is not enough window width to display the admin menu on the
right, we display it on the left instead. Behavior is reversed on RTL
layout.
This commit also removes jQUery usage.
This commit adds the profile tab to the experimental user menu. We're adding it to the user menu because it contains links/buttons that are not available anywhere else. We may remove the tab again if we find better places for those links/buttons, but for now it'll stay.
For more context on the experimental user menu, see https://github.com/discourse/discourse/pull/17379.
This regressed when we switched these bundles to be compiled by ember-cli. Before/after sizes:
```
before after
admin.js: 1370 KB (190 KB gz) --> 782 KB (125 KB gzipped)
wizard.js: 127 KB ( 22 KB gz) --> 68 KB ( 18 KB gzipped)
```
The logic to determine what post excerpt to show for
a topic-level bookmark based on the last unread post
was complex and slow, so we decided to remove it and
always just use the first post excerpt.
This commit also fixes an issue where a couple of
instances of for_topic were missed when doing the
Bookmarkable refactors, so:
1. Clicking the topic bookmark link was not taking
the user to the last unread post
2. When replying to a topic where there was a topic
level bookmark with the auto delete preference
of "on owner reply", we were not removing the
bookmark from the UI correctly.
A test has been added for the former, the latter would
be quite time-consuming to test and not really worth
it considering it's quite an edge case UI bug.
Ever opened `/tests`, immediately tried to change the config, and got frustrated that focus keeps getting stolen by the running tests? Worry no more - this commit will make the tests auto-abort when you click any of the input/dropdowns in the QUnit header. It's not perfect - abort will wait until the end of the currently running test, so sometimes focus will be stolen one more time. But it's still a lot better than having to manually find and click the abort button.
The following changes are made in this commit:
1. Move caret icon in sidebar section header to the right.
1. Each row in sidebar takes the full width which enables us to do a
full width highlight on hover and when sidebar link is active.
1. Ensure each row in Sidebar is of the same height.
Internal refs: /t/70546, /t/72196, /t/71820
These packages are not published anywhere - they only exist for organizational purposes within this repository. Even so, it makes sense for their metadata to match that of the top-level discourse/discourse repository.
By default, in CI environments, Ember CLI does not output anything between "building..." and "cleaning up". Depending on configuration and hardware, Discourse asset builds can take upwards of 60s, and so this lack of output can make the build feel 'stuck'.
This commit introduces an addon which checks for CI mode, and then outputs status information periodically. The logic is very similar to Ember CLI's non-CI progress output implementation (https://github.com/ember-cli/ember-cli/blob/04a38fda2c/lib/models/builder.js#L183-L185).
Without this filter, a binary file (e.g. `.DS_Store`) in one of these addon directories will be concatenated into the output JS. Note that this filter is applied at the end of the build pipeline, so any hbs files have already been transpiled into `.js`.
This mirrors the filtering performed on the main application bundle by Ember CLI: https://github.com/ember-cli/ember-cli/blob/04a38fda2c/lib/broccoli/default-packager.js#L1255