Mobile app can capture event and launch a separate login flow. Should
help resolve issues with passkeys (which aren't available in webviews)
and non-local login methods.
While using `OpenStruct` is nice, it’s generally not a very good idea as
it usually leads to performance problems.
The `OpenStruct` source code even says basically to avoid it.
Since the context object is crucial in our services, this patch replaces
`OpenStruct` with a custom implementation instead.
This commit attempts to improve the mobile experience for
admin page header and subheader by automatically collapsing
all action buttons in these components into a DMenu when viewing
mobile.
This is done by using different "list" wrapper components and a
DMenu trigger and a DropdownMenu on mobile only, and uses has-block
to determine whether to render the DMenu trigger at all.
This also removes the `PluginOutlet` in `AdminPluginConfigPage`, it
was too inflexible for this `DropdownMenu` case, and since the `:actions`
were always rendering we couldn't rely on `has-block`. A new plugin API,
`registerPluginHeaderActionComponent`, has been introduced instead to
replace it.
Because of unreliability, the spec was temporarily disabled. However, it is ensuring that the custom flags system is working correctly. Therefore it would be great to enable it again.
I made a few fixes to try to mitigate this situation:
- Reduced amount of Redis calls;
- When deleting, ensure that the modal is closed before checking the result;
- Moved duplicated name tests to a separate block;
- Increased wait time to 3 times the default because I noticed that sometimes it gets stuck for a moment. Most of the time it is fast, but sometimes when I run tests in a loop 50 times I see slowness.
Dismissing admin notices is an admin-only action. This is enforced on the back-end both by a routing constraint and a policy in the relevant service.
However, we still unconditionally display the "Dismiss" button to anyone with access to the admin dashboard. When clicked, it results in a 404 modal (due to the routing constraint.)
With this change we only render the dismiss button for admins.
If you have the admin dashboard open, and one of the admin notices listed has already been dismissed (e.g. in another tab, or by another admin) we would show an ugly "FAILED" modal.
This change makes the admin dismiss endpoint idempotent. If the admin notice is already destroyed, then respond with 200. This will also correctly remove it from the list in the front-end.
When a post has some replies, and the user click on the button to show them, we would load ALL the replies. This could lead to DoS if there were a very large number of replies.
This adds support for pagination to these post replies.
Internal ref t/129773
FIX: Duplicated parent posts
DEV: Query refactor
Recently we updated the icon library from Font Awesome `5` to `6.6.0`. Since we were running Font Awesome 5 for a long time while 6 had already been released, we often specified in the codebase with the text _"FontAwesome 5"_. However, now that we are in the latest version, there is no need for our API's/comments to keep specifying for version 5. This PR updates all instances of FontAwesome 5 or FA5 and removes the version number to be the more generic: "FontAwesome"
This PR limits this feature:
On all devices:
- Browsers with OffScreenCanvas support
- Browsers with createImageBitmap
On Apple Safari
- At least version 18
It also adds a routine that terminates the worker after 5 uses on all devices to handle any WASM memory leaks. All this together fixes crashes that could occur on iPhones.
It still leaves the feature disabled by default on iOS, which will be revisited after testing this changes.
Currently in services, the `contract` step is only used to define where
the contract will be called in the execution flow. Then, a `Contract`
class has to be defined with validations in it.
This patch allows the `contract` step to take a block containing
validations, attributes, etc. directly. No need to then open a
`Contract` class later in the service.
It also has a nice side effect, as it’s now easy to define multiples
contracts inside the same service. Before, we had the `class_name:`
option, but it wasn’t really useful as you had to redefine a complete
new contract class.
Now, when using a name for the contract other than `default`, a new
contract will be created automatically using the provided name.
Example:
```ruby
contract(:user) do
attribute :user_id, :integer
validates :user_id, presence: true
end
```
This will create a `UserContract` class and use it, also putting the
resulting contract in `context[:user_contract]`.
We are going to start making section landing pages
for admin for each sidebar section. This lays the framework
with routes and simple components that can be further
refined by a designer, but I have taken the base CSS from
AI which Kris made.
The initial section landing items will be used in AI to replace
the placeholders added in this commit b8b3c61451
* Fixes big gap at the left of the plugins list.
* Fixes plugin settings list padding, the yellow overridden
dot was cut off on mobile.
* Also increased settings filter input size and settings
sidebar button margin.
After #28603, the options "agree and suspend" and "agree and silence" in the review queue weren't working. This was happening because the optionalService, when used as a decorator, needs a name argument to work properly. We were also lacking tests for this.
In 61c1d35f17 I added a
PluginOutlet to AdminPluginConfigPage. This was intended to be
used as a way to render actions buttons inside the header of
a plugin that has a custom admin UI page. This worked, but
since the outlet was generically named, as soon as one plugin
used it the button would show on all plugins.
This fixes the immediate issue by naming the outlet based
on the plugin, then having each plugin specify their own
outlet to render into. There may be a nicer way to do this,
but for now this stops the bleeding.
This commit introduces a feature that allows an admin to delete a user's
associated account. After deletion, a log will be recorded in staff
actions.
ref=t/136675
Permanently deleting posts that no longer have a user associated was not
working as expected because of UserAction.log which expected user_id to
be present.
This will help to enforce a consistent pattern for creating service
actions.
This patch also namespaces actions and policies, making everything
related to a service available directly in
`app/services/<concept-name>`, making things more consistent at that
level too.
In our test suite, we sometimes see ChunkLoadErrors. This plugin should cause those failed requests to be retried seamlessly. It'll also help clients with flaky internet connections in production.
When running checks, we look to the existing problem check trackers and try to grab their ProblemCheck classes.
In some cases this is no longer in the problem check repository, e.g. when the check was part of a plugin that has been uninstalled.
In the case where the check was scheduled, this would lead to an error in one of the jobs
the radical change in the implementation doesn't stem from the glimmer migration, but rather the fact that previously the component was single-use – changing any of its args didn't (and couldn't) be reflected because hljs was replacing the nodes so all the ember bookkeeping was gone.
Co-authored-by: David Taylor <david@taylorhq.com>
There is a risk of overriding and then deleting a prop of the context in case of a naming clash between localName and that prop, e.g.
```js
class Test {
item = "foo";
items = [1, 2];
}
const template = `
{{#each items as |item|}}
{{item}}
{{/each}}
`;
const compiledTemplate = compile(template);
const object = new Test();
// object.item === "foo"
const output = compiledTemplate(object, RUNTIME_OPTIONS);
// object.item === undefined
```
…but I think we can accept this risk and just be careful.`#each` isn't widely used in hbr anyway (as proven by the other long-standing and recently fixed bug) and hbr is on its way out anyway.
his is a new feature that lets admins dismiss notices from the dashboard. This helps with self-service in cases where a notice is "stuck", while we work on provisions to prevent "sticking" in the first place.
In some very rare cases, the header element doesn't yet have the bg
when this code is run. This PR adds a simple retry mechanism.
No tests because this relies on specific load timing from the browser.
In TopicController, in addition to ensure_can_move_posts!, we also
checked if the topic is private message in this line:
```ruby
raise Discourse::InvalidAccess if params[:archetype] == "private_message" && !guardian.is_staff?
```
However, this was not present in `guardian.can_move_posts?`. As a result,
the frontend topic view got an incorrect serialized result, thinking
that TL4 could move the private message post. In fact, once they tried
to move it, they got the `InvalidAccess` error message.
This commit fixes that TL4 will no longer sees the "move to" option in
the "select post" panel for a private message.
Anonymous users are "shadow" users created when an existing real user desires to post anonymously. This feature is off by default, but it can be enabled via the `allow_anonymous_posting` site setting. Those shadow users shouldn't be included in the users directory (`/u`).
In 14cf8eacf1, we added the
`user_search_similar_results` site setting which when enabled will use
trigram matching for similarity search in `UserSearch`. However, we
noted that adding the `index_users_on_username_lower_trgm` index is
causing the PG planner to not use the `index_users_on_username_lower`
index when the `=` operator is used against the `username_lower` column.
Based on the PG mailing list discussion where support for the `=`
operator in gist_trgm_ops was being considered, it stated that "I also have checked that btree_gist is preferred over pg_trgm gist
index for equality search." This is however quite different from reality
on our own PG clusters where the btree index is not preferred leading to
significantly slower queries when the `=` operator is used.
Since the pg_trgm gist index is only used for queries when the `user_search_similar_results` site setting
is enabled, we decided to drop the feature instead as it is hidden and
disabled by default. As such, we can consider it experiemental and drop
it without deprecation.
PG mailing list discussiong: https://www.postgresql.org/message-id/CAPpHfducQ0U8noyb2L3VChsyBMsc5V2Ej2whmEuxmAgHa2jVXg%40mail.gmail.com
Previously shortcuts added to toolbar buttons will automatically use the same action assigned to the button when clicked. This PR adds an additional optional key that can be passed when creating a new toolbar button: shortcutAction which allows for passing a custom action for the keyboard shortcut. This way a button can have a specific action when a keyboard shortcut is pressed that's different from when the button is clicked.
Followup to e25578d702
Using execCommand to replace the entire contents of the textarea is very slow for larger posts (it seems the browser does a reflow after every 'virtual keypress'.
This commit updates the `replaceText` function to be more surgical with its `insertAt` calls. Now it only selects & replaces the characters which are actually being replaced.
- fetch models inside services
- validate `user_id` in contracts
- use policy objects
- extract more logic to actions
- write specs for services and action
Previously for reviewables that could be claimed, we positioned
the "you can claim / you must claim" message and button underneath
the "Is there something wrong with this post?" message but _before_
the reviewable action buttons like Yes/No/Ignore. This was a confusing
flow.
This commit fixes the issue, and also makes it so if claiming is
required and the reviewable has not been claimed, we don't show
the "Is there something wrong with this post?" which was showing
with no buttons.
Followup 5a8e7c5f29
The admin report results need to be side by side
with the filter for the report, which sits on the
right. The previous commit made it stacked.
Followup 0323b366f3
This was happening because another spec was adding a
report using the plugin API, but there was nothing
resetting that, so later in the reports controller
when we did Report.singleton_methods, we ended up
with another report with no translation, causing another
error.
Currently, when the custom flag has the same name as the system flag (which is disabled) then it is not displayed. To fix the problem, `custom_` prefix as `name_key` is used to distinguish between the system and the custom flag.
I considered writing a migration to fix existing custom flags name key. However, at the end of migration I would need to run rails code to reset cache `Flag.reset_flag_settings!`. I decided to skip that step as it is a very edge case. If someone has the same flag name as the system flag, then all they have to do is edit the flag and click save.
In addition, I made 2 small fixes:
- edit flag title was missing translation;
- flag form UI was not showing that description is the required field.
This commit adds a new `about_page_hidden_groups` setting to exclude members of specific groups from the admin and moderator lists on the /about page.
Internal topic: t/137717.
* FIX: add "in:first" to user summary category search
The "in:first" parameter was added to the search parameters for the topic count in the user summary category search.
This ensures that the search results focus specifically on the first posts in each topic, so only topics are returned.
Meta topic: https://meta.discourse.org/t/summary-page-by-category-topic-links/215848
* add tests
Provided by @natsw
This commit converts the current chat plugin UI into the
new "show plugin" UI already followed by AI and Gamification.
In the process, I also:
* Made a dedicated /new route to create new webhooks
* Converted the webhook form to FormKit
* Made some fixes and improvements to the `AdminPluginConfigPage`, `AdminPageHeader`,
and `AdminPageSubheader` generic components, so more plugins can
adopt the UI guidelines too. This includes adding a header outlet so plugins
can add action buttons to the plugin show page header.
* Fixes the submit button loading state for FormKit (by Joffrey)
---------
Co-authored-by: Joffrey JAFFEUX <j.jaffeux@gmail.com>
Followup 14b436923c
On the standalone Site Traffic report page, we also need
to hide the 'other' and 'crawler' pageviews by default
like we do on the admin dashboard.
### UI changes
All of the UI changes described are gated behind the `use_legacy_pageviews`
site setting.
This commit changes the admin dashboard pageviews report to
use the "Consolidated Pageviews with Browser Detection" report
introduced in 2f2da72747 with
the following changes:
* The report name is changed to "Site traffic"
* The pageview count on the dashboard is counting only using the new method
* The old "Consolidated Pageviews" report is renamed as "Consolidated Legacy Pageviews"
* By default "known crawlers" and "other" sources of pageviews are hidden on the report
When `use_legacy_pageviews` is `true`, we do not show or allow running
the "Site traffic" report for admins. When `use_legacy_pageviews` is `false`,
we do not show or allow running the following legacy reports:
* consolidated_page_views
* consolidated_page_views_browser_detection
* page_view_anon_reqs
* page_view_logged_in_reqs
### Historical data changes
Also part of this change is that, since we introduced our new "Consolidated
Pageviews with Browser Detection" report, some admins are confused at either:
* The lack of data before a certain date , which didn’t exist before
we started collecting it
* Comparing this and the current "Consolidated Pageviews" report data,
which rolls up "Other Pageviews" into "Anonymous Browser" and so it
appears inaccurate
All pageview data in the new report before the date where the _first_
anon or logged in browser pageview was recorded is now hidden.
This upgrade is designed to be fully backwards-compatible. Any icon names which have changed will be automatically remapped to the new name. For now, this will happen silently. In future, once core & official themes/plugins have been updated, we will start raising deprecation errors to help theme/plugin authors update their code.
Extracted from https://github.com/discourse/discourse/pull/28715
Announcement at https://meta.discourse.org/t/were-upgrading-our-icons-to-font-awesome-6/325349
Co-authored-by: awesomerobot <kris.aubuchon@discourse.org>
* FEATURE: Log tag group changes in staff action log
This commit records every change (add, change, delete) to a tag group in
the staff action log.
It uses a modal that was originally called ThemeChangeModal to display
changes, allowing staffs to see the specific changes clearly. The modal
is renamed to StaffActionLogChangeModal in this PR.
ref: https://meta.discourse.org/t/-/325011/14
Co-authored-by: Keegan George <kgeorge13@gmail.com>
This commit adds a link to the top of the new /about page, shown to admins only, to allow them to easily navigate to `/admin/config/about` where they can edit the /about page.
Internal topic: t/137546.
This patch removes the `with_service` helper from the code base.
Instead, we can pass a block with actions directly to the `.call` method
of a service.
This simplifies how to use services:
- use `.call` without a block to run the service and get its result
object.
- use `.call` with a block of actions to run the service and execute
arbitrary code depending on the service outcome.
It also means a service is now “self-contained” and can be used anywhere
without having to include a helper or whatever.
Static pages such as /about, /faqs, /tos etc. currently overflow horizontally on some Android devices (reproducible on Samsung Galaxy A11). It seems like the `width: 100%` property on `.body-page` is what causing the problem, and removing it doesn't seem to break anything on the various devices that I've tested (desktop, iOS, Android).
* UX: Add a description about badges
* WIP: Apply admin UI guidelines
* FIX: Add routeModels to dbutton
Allows routeModels to be passed to a DButton along
with route, so we can use them as a LinkTo replacement
in more places.
Also fix up badges admin page header.
* UX: Reorder action buttons
* UX: Change header hierarchy to better align page's content structure
* UX: Update copy and remove unnecessary UI elements
* UX: Adjust header's icon spacing
* UX: Fix the header action buttons on mobile
* Apply prettier
---------
Co-authored-by: Martin Brennan <martin@discourse.org>
We need to start printing deprecation notices when the `show_in_ui` argument is used because it works only for the old about page which will be removed soon. For the new about page, we've introduced a new API `addAboutPageActivity` which is more flexible than a true/false argument on the server side.
Internal topic: t/136551.
- Uses `helper.renderGlimmer` with GJS to render the `<Poll` component without any widgets
- Moves some logic into component, so that only `@post`, `@poll` and `@titleHTML` need to be passed into the component (no more 'attrs')
- Updates `modifyClass` calls to modern syntax
- Replaces observer in `Post` model with a native setter & tracked property
- Replaced Poll EmberObject instances with TrackedObject
- Updated component tests with new arguments
- Updated some tests to qunit-dom
- Fixed up core `repliesBelow` and `repliesAbove` logic to create post models properly. Previously it was passing 'transformed' versions of posts into the model, which doesn't make sense.
Currently, categories support designating only 1 group as a moderation group on the category. This commit removes the one group limitation and makes it possible to designate multiple groups as mods on a category.
Internal topic: t/124648.
* DEV: Convert card-contents-base from mixin to base class
This allows us to convert it to native class syntax
Mixins are essentially deprecated, and will not be receiving native class syntax support. In this case, there is no real need for it to be a mixin, so a base class makes sense.
The old /about page has 3 plugin outlets: one after the description, one after the admins list, and one after the mods list. These outlets translate cleanly to the new /about page, so we should add them to it to make plugins/themes compatible with the new page.
This commit introduces a new hidden site setting: `group_pm_user_limit`, default to `1000` which will raise an error when attempting to create a PM target a large group.
This will bring significant improvements to install speed & storage requirements. For information on how it may affect you, see https://meta.discourse.org/t/324521
This commit:
- removes the `yarn.lock` and replaces with `pnpm-lock.yaml`
- updates workspaces to pnpm format
- adjusts package dependencies to work with pnpm's stricter resolution strategy
- updates Rails app to load modules from more specific node_modules directories
- adds a `.pnpmfile` which automatically cleans up old yarn-managed `node_modules` directories
- updates various scripts to call `pnpm` instead of `yarn`
- updates patches to use pnpm's native patch system instead of patch-package
- adds a patch for licensee to support pnpm
TagUser.rb is used to set user notification levels for a tag, we don't have a unique index on the notification level itself. This means that there might be some weird case where a user may have multiple of the same notification level on a tag.
This PR adds a migration which de-duplicates this based on defaults, where we keep the earliest record in the event there is multiple notification level per-user-per-tag.
Originally, we assumed that the href passed to the NavigationItem must
not have URLParams, and roughly appended URLParams such as
`?order=created` to the end. This will result in something like
`/latest?state=my_votes?order=created` (note the double question mark
here), for example the discourse-topic-voting plugin
This commit modifies the logic for appending URLParams to the end,
ensuring that the original URL parameters are preserved.
* add data migration to keep only unexpired or most recently expired user password
* refactor to 1:1 relationship between User and UserPassword
* add migration to remove redundant indexes on user passwords
A new setting attribute is used to define the areas (separated by `|`).
In addition, endpoint `/admin/config/site_settings.json` accepts new `filter_area` data.
A previous refactor has moved this function in the controller instead of the route making it inaccessible to the modal.
This commit is fixing this and also adding a spec.
This change introduces a new thread notification level allowing users to get notified when someone replies to the thread.
Users who watch a thread will get a green notification on the chat icon and a user notification (blue). User notifications are consolidated based on thread id to prevent cluttering the original users notification area.
---------
Co-authored-by: Régis Hanol <regis@hanol.fr>
In particular, this allows us to use gjs `<template>` for RenderGlimmer, registerWidgetShim, and decorateCooked's `helper.renderGlimmer`.
```js
// Simple component class:
registerWidgetShim(
"render-glimmer-test-component-shim",
"div.initial-wrapper-class",
SimpleComponent
);
```
```gjs
// Or an inline `<template>`
registerWidgetShim(
"render-glimmer-test-component-shim",
"div.initial-wrapper-class",
<template>Hello world</template>
);
```
We're working to remove all decorators from object-literal-properties. This commit refactors all initializers which were using these decorators into classes, where decorators are allowed. Also adds cleanup logic to the local-dates initializer, which was previously missing.
This will avoid warnings: "The code generator has deoptimised the styling of {path} as it exceeds the max of 500KB."
Should also provide a tiny improvement in build times
This commit introduces a little bit of duplication
since the old plugin UIs not using the new plugin show
page look different from ones like AI and Gamification
which have been converted. We can use the new admin
header component on the plugins list, but for the other
pages we are manually rendering a breadcrumb trail and
the list of plugin tabs.
Over time as we convert more plugins to use the new UI
guidelines and show page we can get rid of this duplication.
This commit adds a description for all the auto groups
which will be shown in the group list and show group
pages, which will help admins understand their purpose
better.
Also adds an indicator with a tooltip to explain what
the auto groups are on the group show page.
* Do not delete created external upload stubs for 2 days
instead of 1 hour if enable_upload_debug_mode is true,
this aids with server-side debugging.
* If using an API call, return the detailed error message
if enable_upload_debug_mode is true. In this case the user
is not using the UI, so a more detailed message is appropriate.
* Add a prefix to log messages in ExternalUploadHelpers, to
make it easier to find these in logster.
Prior to this fix we were calling `fetch_stats` which is never checking if we have a cache entry. This call is making a lot of SQL calls, so it's better to use the cache.
Introduce a new log line for Sidekiq jobs that consume more than
`DISCOURSE_LIVE_SLOTS_SIDEKIQ_LIMIT` live slots. This is useful to
track down jobs that may leak memory.
This is enabled only when Sidekiq's job instrumenter is enabled (set
`DISCOURSE_LOG_SIDEKIQ` to `1`).
We're seeing a lot of log noise coming from unhandled exceptions stemming from requests to TopicsController#show where id is passed in as an array.
In the implementation of the method, we assume that if id is present it will be a string. This is because one of the routes to this action uses :id as a URL fragment, and so must be a string. However, there are other routes that go to this endpoint as well. Some of them don't have this URL fragment, so you can pass an arbitrary id query parameter.
Instead of a downstream unhandled exception, we raise a Discourse::InvalidParameters upfront.
In this PR we introduced a new setting `enforce_second_factor_on_external_auth` which disables enforce 2FA when the user is authenticated with an external provider.
https://github.com/discourse/discourse/pull/27506
However, with the first registration with an external provider, we authenticate the user right after activation. In that case, we need to also keep information that the user was authenticated with an external OAuth provider.
We were not updating `searchTerm` when changing the input which was making us always send an empty q parameter.
This commit is also adding tests for:
- initial url with q param
- filtering the bookmarks through the input
I mistakenly replaced these with `@equal`, thinking the behavior was the same. It's not. `@propertyEqual` compares two properties, while `@equal` compares a single property with a constant.
Changes made using the ember-native-class-codemod, plus some manual tweaks
Also ensures our implicit injections applied to the prototype immediately. Without this, they will only be applied on the next `.extend()` call, which is now later than the first native-class extension.
When the tooltip items are tooltipItem = [{parsed: {y:12} }, {parsed: {y:10} } ], reducing without a initial value as a number would result in Javascript thinking it is a string. Thanks, Javascript!
There are no tests here yet since this makes use of an external library Chart.js.
Why all the manual setting of `width` and `height`? Without it all ace editors were 0x0, invisible.
Why didn't it affect the tests before? ace-editor tests only confirms that there is an element, theme-settings-editor tests were effectively unit tests before 7317b5a, fk-code test avoids interacting with ace.
Makes it easier to reach the group from the category security
tab, and moves the trash button to the right to avoid misclicks.
Also converts the category permission row to gjs
The `notifications.id` has been migrated to bigint in previous commit
799a45a291. This commit migrates one of
the related columns, `user_badges.notification_id`, to `bigint`.
This commit changes the cutoff number for the admins and mods lists on the new /about page from 12 to 6. If the admins or mods lists are bigger than 6, the about page will display the 6 most recently seen admins/mods, and tuck the rest away behind a "view more" button.
Given we are a single-page-app, we hijack the browser's default behavior
for keyboard shortcuts like End because on long topics (20+ posts) we
need to load the last post before really reaching the end of the page.
However, when the last post is already rendered, the End shortcut
currently does nothing, it takes the user to the start of the last post.
If that post is too long, the user will have to scroll down manually.
This change ensures that if the last post of a topic is already rendered
(whether it is in the viewport or not), pressing End will take the user
to the very bottom of the page.
Note for the reviewer: the test added here is for the general case, it
is too hard to test the case where the last post is already rendered,
that isn't covered here.
In FormKit you can add error on an existing field existing in the DOM, but you can also set an arbitrary error on a virtual field not existing in the DOM.
When revalidating existing data, we are only resetting real fields. This commit adds `removeError(name)` to allow you to manually manage virtual fields. `removeError` is available in the same helpers where `addError` is available.
<!-- 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. -->
Prior to this fix using:
- `replaceText(...)`
- `selectText(..., ..., { scroll: true})`
Wouldn't have the expected behaviour as the scroll from selectText will attempt to save the scroll position and restore it AFTER the replacement happened. This commit allows scroll to be a Boolean or a Number, when a number, it will be used to restore the scrollTop position.
I tried to write tests for this specific behavior but couldn't reproduce the issue in tests.
A follow-up to dd1abf91ef.
Because `ember-cli-deprecation-workflow` relied on `@ember/string` 3.1.1, that was the effective version used in the app.
Turns out there's a compatiblity issue with 4.0.0. Let's lock it to 3.x so we can easily update deprecation-workflow and then we can look into the ember/string issue.
Follow-up to e3ae57ea7a
The previous commit added an `after_create` callback that triggers a refresh for the user directory whenever a `User` record is created. Theoretically, this approach should work, however, there's a gotcha in practice, because during a real user registration, when the `User` record is created in the database, it's not marked as active until the user verifies their email address and the user directory excludes inactive users, so the initial directory refresh triggered by the `after_create` callback becomes pointless.
To make a new user appear in the user directory immediately after sign up, we need to trigger a refresh via an `after_save` callback when they verify their email address and become active.
Currently, `FlagSerializer#short_description` properly provides
`base_path` to its translations, but `FlagSerializer#description` does
not. This breaks the link to guidelines when flagging a post, for
example.
This patch provides `base_path` for `FlagSerializer#description` too.
All those addons import it but didn't have it in their package.jsons. And because the only thing that had it as a dependency (ember-cli-deprecation-workflow) removed it in the latest version, the build breaks.
File it under: a problem we wouldn't have with pnpm :P
Prior this fix a topic created in the future could generate this exception:
```
Job exception: ERROR: a negative number raised to a non-integer power yields a complex result
```
This fix and spec should ensure we don't consider topics in the future as the rest of the `update_scores` function is doing at other places.
In GroupRequestsController, request_count is incorrectly written as
user_count, which causes group member requests to be loaded infinitely
when user_count is greater than request_count.
This check was checking the wrong scope, causing problems in certain edge conditions, for example:
1. Admin adds an "on signup" field that isn't editable after signup.
2. Admin adds a "for all users" field.
3. User goes and fills up the "for all users" field from 2.
4. User is now stuck on the required fields page without any fields showing.
With this change, we only consider "for all users" fields when asking if required custom fields are filled in.
The `notifications.id` column is the most probable column to run out of
values. This is because it is an `int` column that has only 2147483647
values and many notifications are generated on a regular basis in an
active community. This commit migrates the column to `bigint`.
These migrations do not use `ALTER TABLE ... COLUMN ... TYPE` in order
to avoid the `ACCESS EXCLUSIVE` lock on the entire table. Instead, they
create a new `bigint` column, copy the values to the new column and
then sets the new column as primary key.
Related columns (see `user_badges`, `shelved_notifications`) will
be migrated in a follow-up commit.
When searching for site texts for admin using the english
version of the text, previously we would show the english
version in the results _even if_ there was another locale
translated version available when a locale was selected
from the dropdown.
This commit adds a "Only show results in selected locale"
checkbox option which will instead make it so the results
shown are in the target locale, making it easier for translators
to tell when there is actually translations vs. missing tranlsations.
Followup 76c56c8284
The change introduced above made it so the expired
bookmark reminders were cleared when using the bulk
action menu for bookmarks. However this also affected
clearing reminders for bookmarks when sending notifications.
When clearing bookmark reminders after sending notifications,
we take into account the auto delete preference:
* never - The bookmark `reminder_at` date should not be cleared,
and the bookmark is kept.
* clear_reminder - The bookmark `reminder_at` date is cleared and
the bookmark is kept
The `never` option made it so "expired" bookmark reminder show
on the user's bookmark list.
This commit fixes the change from the other commit and only
forces clearing of `reminder_at` if using the bookmark bulk
action service.
If there is a codec issue or something trying to process a video file
for thumbnail generation, uploads could hang indefinitely. This fix
ensures that we continue the upload process even if we encounter an
error trying to generate a thumbnail for it.
If you’re viewing a tag and you switch to a different tag via the sidebar or the tags dropdown, after expanding the info section of the tag page via the wrench button, the info section keeps showing the previous tag's details instead of the new one.
This happens because the tag info section makes an ajax request to load the tag's details, and this request is made inside the `didInsertElement` hook which is only fired once when the component is rendered. To fix this, we need to set the result from the ajax request to null and add a `didUpdateAttrs` hook to trigger another request to load the info of the new tag.
Internal topic: t/134809.
This change is preventing the "is dirty check" from happening when clicking delete on this form. This was not good UX and was also causing bugs by leaving the form in a unexpected state.
This commit improves `TopicsController#show` to not load suggested and
related topics unless it is the last page of the topic's view.
Previously, we avoided loading suggested and related topics by the use
of conditionals in the `TopicViewSerializer` to avoid calling
`TopicView#suggested_topics` and `TopicView#related_topics`. However,
this pattern is not reliable as the methods can still be called from
other spots in the code base. Instead, we ensure that
`TopicView#include_suggested` and `TopicView#include_related` is set
correctly on the instance of `TopicView` which ensures that for the
given instance, `TopicView#suggested_topics` and
`TopicView#related_topics` will be a noop.
Previous to this change there is no clean way to apply keyboard shortcuts
to things such as "add poll" and other hidden options in the toolbar
This allows shortcuts to be specified similar to how they are on the toolbar
Co-authored-by: Joffrey JAFFEUX <j.jaffeux@gmail.com>
If we don't get a `videoWidth` back for a video don't try and generate a
thumbnail for it.
Also as part of this change I switched getImageData, the function
throwing the error, to use video.videoWidth instead of canvas.width
because it's very likely we were setting canvas.width too early before
the width could be read. Now that we are reading the value inside of the
setTimeout hopefully we will actually have a width. Just incase we don't
detect a width we will now exit early instead of throwing an error.
We only need to check for `0` and not null because the value is an
integer and will always return a 0 if it can't be read. https://developer.mozilla.org/en-US/docs/Web/API/HTMLVideoElement/videoWidth
See https://meta.discourse.org/t/322363
This commit moves the business logic in the `Admin::UsersController#suspend` and `Admin::UsersController#silence` actions to dedicated service classes. There's no functional changes in this commit.
Internal topic: t/130014.
This lays the groundwork for converting SelectKit subclasses to native class syntax. This commit is designed to be entirely backwards-compatible, so it should not affect any existing subclasses.
Of interest:
- Any properties which are designed to be overridden by subclasses are implemented using a local `@protoProp` decorator. That means they are applied to the prototype, so that they can be overridden in subclasses by both legacy `.extend()` prototype extensions, and by modern native-class fields.
- New class decorators are introduced: `@selectKitOptions` and `@pluginApiIdentifiers`. These are native class versions of the legacy `concatenatedProperties` system. This follows the pattern Ember has introduced for `@className`, `@classNameBindings`, etc.
This would cause the infamous error:
```
index.js:118 Uncaught (in promise)
Error: Assertion Failed: You attempted to update `href` on `<ExtraNavItem:ember384>`, but it had already been used previously in the same computation.
```
Moreover, set didnt seem necessary here.
In some situations, the filesystem cache will be read and persisted to the database. If the file being read is still being written, then that can lead to empty/partial caches being stored in the database.
This commit ensures that cannot happen by switching to our `atomic_write_file` helper (which writes to a temp file, and then does an atomic `mv` operation to move it to the destination)
Followup 1446596089
The link to inform admins that restore is disabled
was not correct. This fixes it and also changes it
to go to /admin/backups/settings
This commit changes the custom flags admin config area
to use the AdminPageHeader component and conforms
to the new admin UI guidelines. Also makes some slight
modifications to the reusable components to handle using
a @route for an action button.
This commit implements 2 new metrics/stats in the /about page for the _estimated_ numbers of unique visitors from the EU and the rest of the world. This new feature is currently off by default, but it can be enabled by turning on the hidden `display_eu_visitor_stats` site settings via the rails console.
There are a number of assumptions that we're making here in order to estimate the number of unique visitors, specifically:
1. we're assuming that the average of page views per anonymous visitor is similar to the average number of page views that a logged-in visitor makes, and
2. we're assuming that the ratio of logged in visitors from the EU is similar to the ratio of anonymous visitors from the EU
Discourse keeps track of the number of both logged-in and anonymous page views, and also the number of unique logged-in visitors and where they're from. So with those numbers and the assumptions above, we can estimate the number of unique anonymous visitors from the EU and the rest of the world.
Internal topic: t/128480.
We need a way to disable certain checks programatically, e.g. on Discourse hosting. This PR adds a configuration option for this, and makes it so that disabled checks aren't run as part of #run_all.
This patch allows using an AR relation as a model in services without
fetching associated records. It will just check if the relation is empty
or not. In the former case, the execution will stop at that point, as
expected.
This commit introduces a new frontend API to add custom items to the "Site activity" section in the new /about page. The new API is called `addAboutPageActivity` and it works along side the `register_stat` serve-side API which serializes the data that the frontend API consumes. More details of how the two APIs work together is in the JSDoc comment above the API function definition.
Internal topic: t/128545/9.
This commit improves the hilight-ing of mentions in posts and chat messages.
- `@here` and `@all` will generate a `<a class="mention --wide">`
- bots will generate a `<a class="mention --bot">`
- current user will generate a `<a class="mention --current">`
To achieve this change the following value transformer has been added: "mentions-class". It will be run in posts and chat messages after the mention is rendered.
A bug were bots were not considered in mentioned users has also been fixed as part of this PR.
This commit fixes a bug where the silence button is incorrectly displayed on the admin page of a staff user. It's not actually possible to silence a staff user because the backend correctly prevents it, but the frontend isn't checking if the button should be displayed.
Another small bug that this commit fixes is the similar users list not showing up inside the silence/suspend modals due to also a bug in the frontend.
I've also changed the way similar users are loaded so that they're not returned by the `admin/users#show` endpoint anymore and moved them into a new endpoint that the penalize modals (suspend and silence) can call directly to retrieve the list of users. This is done because the similar users list is never shown on the admin user page (`/admin/users/:user_id/:username`); they're only needed when the suspend or silence modals are opened.
Internal topic: t/130014.
This is a variation on bc3e8a9963cf9a64d114ec751c875025af169690, which was reverted due to issues on iOS. Safari's "in response to user action" check cannot follow the `runAfterFramePaint` chain of interaction -> requestAnimationFrame -> messageChannel, and so some sensitive browser APIs (e.g. clipboard, upload, etc.) were blocked.
This commit is similar, but uses `next()` instead of `runAfterFramePaint()`. The result seems the same, but doesn't have the same issue on iOS.
The chat-emoji-picker change was required to resolve a test failure. The emoji picker has never closed-on-scroll on desktop, so there is no user-facing change in behavior.
This commit converts the Backups page in the admin interface
to follow our new admin interface guidelines.
As part of this work, I've also made `AdminPageHeader` and `AdminPageSubheader`
components that can be reused on any admin page for consistency, that handle
the title and action buttons and also breadcrumbs.
Also renamed `AdminPluginFilteredSiteSettings` to `AdminFilteredSiteSettings` since
it can be used generally to show a subset of filtered site settings, not only
settings for a plugin. Not sure if it's ideal to have to define a new route for this
for every config area, but not sure how else to do it right now.
This commit fixes two codepaths which where incorrectly working with capitalized usernames as we were doing a mix of username_lower and non lower username.
Also adds two specs for these cases.
This commit adds a blue dot next to the "What's New"
link in the admin sidebar if the user has not seen the
new features yet, as a followup to 3e5976f843
which removed the tab on the dashboard that had this same
functionality.
When the admin visits the "What's New" page they count
as having seen all the features straight away. This could
be something we want to change, but for now this keeps the
same functionality.
`routeAction` is testing at runtime that a route exists when in debug mode. However in the case of components tested in isolation there's no existing route which was causing an exception, this commit prevents this check in this case as it's irrelevant.
* System user attachment size WIP
* spec check
* controller update
* add max to system_user_max_attachment_size_kb
* DEV: update to use static method for `max_attachment_size_for_user`
add test to use large image.
add check for failure.
* DEV: update `system_user_max_attachment_size_kb` default value to 0
remove unecessary test.
update tests to reflect the new default value of `system_user_max_attachment_size_kb`
* DEV: update maximum_file_size to check when is an attachment made by a system user
Add tests for when `system_user_max_attachment_size_kb` is over and under the limit
Add test for checking interaction with `max_attachment_size_kb`
* DEV: move `max_attachment_size_for_user` to private methods
* DEV: turn `max_attachment_size_for_user` into a static method
* DEV: typo in test case
* DEV: move max_attachment_size_for_user to private class method
* Revert "DEV: move max_attachment_size_for_user to private class method"
This reverts commit 5d5ae0b715.
---------
Co-authored-by: Gabriel Grubba <gabriel@discourse.org>
Long code is considered 1000 characters atm. 10000+ characters would take seconds due to how highlightjs will consider every installed lang before choosing the best candidate.
Move it into the preceding `dl` element and remove the `div` wrapper.
The two plugins that use this outlet:
* d-follow - actually jumps through hacky hoops to get its elements into that `dl` anyway
* d-gamification - would look better if its element was in that `dl`
We used to show New Features in a tab on the dashboard,
but this could get pushed down the page especially on
our hosting. In 043117ca13
we made a separate What's New page, so this commit removes
the dashboard tab and changes the admin notification to
send the admin to /admin/whats-new instead of the dashboard
tab.
This commit fixes a number of bugs in `file_size_restriction` settings and does a little of refactoring to reduce duplicated code in site setting types (the refactoring is necessary to fix one of the bugs).
The bugs in `file_size_restriction` settings that are fixed in this commit:
1. Save/cancel buttons next to a `file_size_restriction` setting are shown upon navigating to the settings page without changes being made to the setting
2. Cancel button that discards changes made to the setting doesn't work
3. Reset button that resets the setting to its default doesn't work
4. Validation error message isn't cleared when resetting/cancelling changes
To repro those bugs, navigate to `/admin/site_settings/category/files` and observe the top 2 settings in the page (`max image size kb` and `max attachment size kb`).
Internal topic: t/134726.
Desktop chat notification sounds have stopped working on most desktop browsers.
This is due to Notifications API being disabled when Push Notifications are supported in the browser, which means that we never iterate on the desktopNotificationHandlers and trigger the callback since we return early.
When browsing through a sidebar with this feature enabled (i.e. admin, or docs), it's weird to have the scroll jump around when you click an item. This commit adds a check, so that we only `scrollIntoView` for items which are not already in the viewport.
Followup to b7cce1a0dc
We changed the design of the member access wizard step to use toggle groups instead of switches. To support existing designs for notices, we need another plugin outlet.
0 is falsy in JavaScript, so the original code would treat 0 as if it
were not input. This unique exception was added to prevent 0 from being
treated as empty input.
This commit fixes a bug in the redesigned about page where if there's no banner image configured for the page, the top of the page where the banner goes is occupied with large white space. Additionally, this commit also fixes a related bug in the admin config area for the /about page where it's not possible to remove the uploaded banner image.
### Why?
Before, all flags were static. Therefore, they were stored in class variables and serialized by SiteSerializer. Recently, we added an option for admins to add their own flags or disable existing flags. Therefore, the class variable had to be dropped because it was unsafe for a multisite environment. However, it started causing performance problems.
### Solution
When a new Flag system is used, instead of using PostActionType, we can serialize Flags and use fragment cache for performance reasons.
At the same time, we are still supporting deprecated `replace_flags` API call. When it is used, we fall back to the old solution and the admin cannot add custom flags. In a couple of months, we will be able to drop that API function and clean that code properly. However, because it may still be used, redis cache was introduced to improve performance.
To test backward compatibility you can add this code to any plugin
```ruby
replace_flags do |flag_settings|
flag_settings.add(
4,
:inappropriate,
topic_type: true,
notify_type: true,
auto_action_type: true,
)
flag_settings.add(1001, :trolling, topic_type: true, notify_type: true, auto_action_type: true)
end
```
`defaultCategoryLinkRenderer` is using a fake category object which doesn’t have access to the functions and getters of category model.
This had been incorrectly set in c197daa04c
As we don't get a real category object, we have to call the transformers manually and also pass the fake category object as context, this is not ideal as people might try to access properties in the transformer which are not available on the category object given they will be different based on the context. Hopefully one day this helper and all the chain can be refactored to use a real category model.
This commit also adds tests for these two properties in the category-link helper.
<!-- 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 forces the textarea to check if the list is inside a codefence and won't continue the list if it's the case.
Note this commit also uses the message param of qunit assertions to make them more explicit. It has no impact on behavior.
This change ensures native push notifications respect the site setting for push_notification_time_window_mins. Previously only web push notifications would account for the delay, now we can bring more consistency between Discourse in browser vs Hub, by applying the same delay strategy to both forms of push notifications.
### Why?
Before, all flags were static. Therefore, they were stored in class variables and serialized by SiteSerializer. Recently, we added an option for admins to add their own flags or disable existing flags. Therefore, the class variable had to be dropped because it was unsafe for a multisite environment. However, it started causing performance problems.
### Solution
When a new Flag system is used, instead of using PostActionType, we can serialize Flags and use fragment cache for performance reasons.
At the same time, we are still supporting deprecated `replace_flags` API call. When it is used, we fall back to the old solution and the admin cannot add custom flags. In a couple of months, we will be able to drop that API function and clean that code properly. However, because it may still be used, redis cache was introduced to improve performance.
To test backward compatibility you can add this code to any plugin
```ruby
replace_flags do |flag_settings|
flag_settings.add(
4,
:inappropriate,
topic_type: true,
notify_type: true,
auto_action_type: true,
)
flag_settings.add(1001, :trolling, topic_type: true, notify_type: true, auto_action_type: true)
end
```
Adds a new statistics (hidden from the UI, but available via the API) that tracks daily participating users.
A user is considered as "participating" if they have
- Reacted to a post
- Replied to a topic
- Created a new topic
- Created a new PM
- Sent a chat message
- Reacted to a chat message
Internal ref - t/131013
This commit continues on work laid out by 6039b513fe to redesign the /about page. In this commit, we add sections for showing the site admins and moderators.
The lists of admins and moderators display the 10 most recently seen admins/moderators, with a button to display the rest of admins or moderators. Admins or moderators that have not logged in to the site in the last year will not be shown. Clicking on an admin's or moderator's name/avatar will show their user card.
e.g.
```
WARNING: Binding style attributes may introduce cross-site scripting vulnerabilities; please ensure that values being bound are properly escaped. For more information, including how to disable this warning, see https://deprecations.emberjs.com/v1.x/#toc_binding-style-attributes. Style affected: \"height: 60px\"
```
This message indicates broken behavior, so it should be an error rather than a warning.
An early-return is added, so that we don't even attempt to make the modification. This will make the behavior consistent, and easier to understand.
Also updates the normalization logic to use the resolver's own logic. This will handle all sorts of normalization in addition to our deprecations.
In development, Ember raises an error when previously-used values are updated during a render. This is to avoid 'backtracking', where parts of templates have to be re-rendered multiple times. In general, this kind of pattern should be avoided, and Ember's warning helps us do that.
However, for the deprecation warning banner, it is quite reasonable for some rendering to trigger a deprecation, and thereby require the global-notice to be re-rendered. We can use our `DeferredTrackedSet` to achieve that. Its `.add` method will delay adding an item to the Set until after the current render has completed.
e.g. we map `controller:composer` to `service:composer` in resolver lookups. So, when doing the cache check in modifyClass, we need to check against the normalized name, not the deprecated name.
Very similar to move up/down flag problem fixed here - https://github.com/discourse/discourse/pull/28272
Those are the steps to toggle the flag:
1. click toggle - `saving` CSS class is added;
2. request to backend;
3. `saving` CSS class is removed.
And check if the flag was toggle was:
```ruby
def has_saved_flag?(key)
has_css?(".admin-flag-item.#{key}.saving")
has_no_css?(".admin-flag-item.#{key}.saving")
end
```
If the save action is very fast, then the saving class is removed before the first check.
Therefore I decided to invert it, and once action is finished add `saved` CSS class.
Then we can have a quick positive check:
```ruby
def has_saved_flag?(key)
has_css?(".admin-flag-item.#{key}.saved")
end
```
This commit adds two new getters to the category model:
- `displayName`
- `descriptionText`
These getters are used instead of `name` and `description_text` where appropriate.
On top of this two transformers have been added to allow plugins to alter these getters:
```javascript
api.registerValueTransformer(
"category-display-name",
({ value, context }) =>
value + "-" + context.category.id + "-transformed"
);
```
```javascript
api.registerValueTransformer(
"category-description-text",
({ value, context }) =>
value + "-" + context.category.id + "-transformed"
);
```
Currently, in the list controller, when encountering an unsafe redirect
error, a 404 is rendered. The problem is that it’s done in a way that it
also logs a fatal error (because a `Discourse::NotFound` exception was
raised inside a `rescue_from` block).
This patch addresses that issue by simply rendering a 404 without
raising any error.
Those are the steps to move the flag:
1. open menu;
2. click move up - `saving` CSS class is added;
3. request to backend;
4. `saving` CSS class is removed.
To check if the action was finished we are using this method:
```
def move_up(key)
open_flag_menu(key)
find(".admin-flag-item__move-up").click
has_saved_flag?(key)
self
end
def has_saved_flag?(key)
has_css?(".admin-flag-item.#{key}.saving")
has_no_css?(".admin-flag-item.#{key}.saving")
end
```
However, sometimes specs were failing with `expected to find CSS ".admin-flag-item.spam.saving" but there were no matches`
I think that the problem is with those 2 lines:
```
find(".admin-flag-item__move-up").click
has_closed_flag_menu?
```
If the save action is very fast, then the `saving` class is removed before the first check.
Therefore, to determine that the move action is finished, I am checking if the menu is closed.
Currently, when a badly named category slug is provided, it can lead to
an infinite redirect.
This patch addresses the issue by properly unescaping `request.fullpath`
so the path is successfully rewritten and the redirect happens as
expected.
Currently to handle stub topics after merging, there are only options to (1) never delete a stub topic and (2) delete a stub topic after X amount of days. This adds the option to immediately delete a stub topic upon merge.
---------
Co-authored-by: Mark VanLandingham <markvanlan@gmail.com>
Co-authored-by: Renato Atilio <renato@discourse.org>
This resolves issues when a mix of callback-based modifications and Ember-reopen-based modifications are used on the same target. In summary:
- Fixes `pluginId` exception logic for callback-based modifications
- Moves `pluginId` storage to a WeakMap so it doesn't pollute the target's descriptors
- When applying a legacy modifyClass, we will temporarily rollback any modern callback-based modifications. This means all of Ember's reopen calls apply to un-prepended classes, and then we add our modern prepends on top.
- Calls `.proto()` on CoreObject descendants before prepending, to ensure that pending Ember mixins have been applied
This commit continues on work laid out by 6039b513fe to redesign the /about page. In this commit, we add the site age and a section on the right hand side to show site activities/statistics such as topics, posts, sign-ups, likes etc.
- Added `addLogSearchLinkClickedCallbacks` which allows plugins/TCs to register a callback when a search link is clicked and before a search log is created
Following a recent refactor, some methods from `FlagSettings` have been
renamed (`custom_types` -> `additional_message_types`). The
`PostActionType` model was using `custom_types` but when the renaming
was done, it was renamed to `with_additional_message` instead of
`additional_message_types`, which under the right circumstances will
raise an error.
During our refactoring of admin badges we decided to link to:
`adminSiteText.edit locale=locale`
Instead of:
`adminSiteText q=key`
After feedback from the community we are reverting this change.
A recent change in FormKit has changed the syntax of this specific component. It's also better to use `<CheckboxGroup />` for this use case too.
Im mixed on writing tests for labels, it's a lot of tests to write for a rather low value.
This commit also slightly tweaks the width of the icon picker, from medium to small.
Admin can create up to 50 custom flags. It is limited for performance reasons.
When the limit is reached "Add button" is disabled and backend is protected by guardian.
This commit patches `Net::HTTP` to reduce the default timeouts of 60
seconds when we are processing a request. There are certain routes in
Discourse which makes external requests and if the proper timeouts are
not set, we risk having the Unicorn master process force restarting the
Unicorn workers once the `30` seconds timeout is reached. This can
potentially become a vector for DoS attacks and this commit is aimed at
reducing the risk here.
The Safari 15 bugfix has been rolled into @babel/preset-env in the most recent version, so we no longer need to carry our vendored copy.
This commit updates @babel/preset-env, runs npx yarn-deduplicate yarn.lock, and removes the vendored transform.
This commit also refactors our theme transpiler to use @babel/preset-env, with the same list of target browsers as our ember-cli build uses. This means we no longer need to maintain a separate list of babel transforms for themes.
We support a low-level construct called "inline checks", which you can use to register a problem ad-hoc from within application code.
Problems registered by inline checks never show up in the admin dashboard, this is because when loading the dashboard, we run all realtime checks and look for problems. Because of an oversight, we considered inline checks to be "realtime", causing them to be run and clear their problem status.
To fix this, we don't consider inline checks to be realtime, to prevent them from running when loading the admin dashboard.
Followup to f70a65ea02
1. Update a second regex in `routeTo` to avoid stripping domain/protocol from middle of string
2. Update `URL.handleURL` to strip double-slashes in paths, before calling the ember router. This mimics what Ember does on initial page-load
Additional tests are added for both
This change is mainly a refactor of the desktop notifications service to improve readability and have standardised values for tracking state for current user in regards to the Notification API and Push API.
Also improves readability when handling push notification jobs, especially in scenarios where the push_notification_time_window_mins site setting is set to 0, which will allow sending push notifications instantly.
To achieve this, a new notifications service is set up with an `isInDoNotDisturb` tracked property. While a user is in do-not-disturb mode, it runs a regular timer until do-not-disturb is over.
We were writing theme-transpiler JS files to the filesystem on a per-process basis, and then immediately reading them back in. Plus, there was no cleanup mechanism, so the tmp directory would grow indefinitely.
This commit refactors things so that the `build.js` script outputs the theme-transpiler source to stdout. That way, we can read it directly into the process, and then into mini-racer, without needing to go via the filesystem. No cleanup required!
In production, the theme-transpiler is still cached in a file during `assets:precompile`
In the formkit conversion in 2ca06ba236
we missed setting a type for the UppyImageUploader for badges. Also,
we were not passing down the `image_url` as form data, so when we used
`data.image` for that field the badge was not updating in the UI after
page loads and the image URL was not loading for preview.
Co-authored-by: Joffrey JAFFEUX <j.jaffeux@gmail.com>
This commit introduces the `behaviorTransformer` API to safely override behaviors defined in Discourse.
Two new plugin APIs are introduced:
- `addBehaviorTransformerName` which allows plugins and theme-components to add a new valid transformer name if they want to provide overridable behaviors;
- `registerBehaviorTransformer` to register a transformer to override behaviors.
It also introduces the function `applyBehaviorTransformer` which can be imported from `discourse/lib/transformer`. This is used to mark a callback containing the desired behavior as overridable and applies the transformer logic.
How does it work?
## Marking a behavior as overridable:
To mark a behavior as overridable, in Discourse core, first the transformer name must be added to `app/assets/javascripts/discourse/app/lib/transformer/registry.js`. For plugins and theme-components, use the plugin API `addBehaviorTransformerName` instead.
Then, in your component or class, use the function `applyBehaviorTransformer` to mark the Behavior as overridable and handle the logic:
- example:
```js
...
@action
loadMore() {
applyBehaviorTransformer(
"discovery-topic-list-load-more",
() => {
this.documentTitle.updateContextCount(0);
return this.model
.loadMore()
.then(({ moreTopicsUrl, newTopics } = {}) => {
if (
newTopics &&
newTopics.length &&
this.bulkSelectHelper?.bulkSelectEnabled
) {
this.bulkSelectHelper.addTopics(newTopics);
}
if (moreTopicsUrl && $(window).height() >= $(document).height()) {
this.send("loadMore");
}
});
},
{ model: this.model }
);
},
...
```
## Overriding a behavior in plugins or themes
To override a behavior in plugins, themes, or TCs use the plugin API `registerBehaviorTransformer`:
- Example:
```js
withPluginApi("1.35.0", (api) => {
api.registerBehaviorTransformer("example-transformer", ({ context, next }) => {
console.log('we can introduce new behavior here instead', context);
next(); // call next to execute the expected behavior
});
});
```
Ember's legacy mixin system does not support native-class syntax, so we have to use the non-decorator syntaxes for `action()` and `computed()`.
Eventually, we will need to refactor things to remove these mixins... but today is not that day.
When creating a shared draft, we're recording topic view stats on the draft and then pass those on when the draft is published, conflating the actual view count.
This fixes that by not registering topic views if the topic is a shared draft.
When `SiteSetting.review_every_post` is true and the category `require_topic_approval` system creates two reviewable items.
1. Firstly, because the category needs approval, the `ReviewableQueuePost` record` is created - at this stage, no topic is created.
2. Admin is approving the review. The topic and first post are created.
3. Because `review_every_post` is true `queue_for_review_if_possible` callback is evaluated and `ReviewablePost` is created.
4. Then `ReviewableQueuePost` is linked to the newly generated topic and post.
At the beginning, we were thinking about hooking to those guards:
```
def self.queue_for_review_if_possible(post, created_or_edited_by)
return unless SiteSetting.review_every_post
return if post.post_type != Post.types[:regular] || post.topic.private_message?
return if Reviewable.pending.where(target: post).exists?
...
```
And add something like
```
return if Reviewable.approved.where(target: post).exists?
```
However, because the callback happens in point 3. before the `ReviewableQueuePost` is linked to the `Topic`, it was not possible.
Therefore, when `ReviewableQueuePost` is creating a `Topic`, a new option called `:reviewed_queued_post` is passed to `PostCreator` to avoid creating a second `Reviewable`.
Currently, descriptions for flag types aren’t interpolated, returning
`%{base_path}` in their string, for example. This breaks the navigation
on the sites.
The behavior changed probably because of an upgrade of Ruby, as two
hashes were passed to `I18n.t` (`vars` and `default`) without using the
splat operator.
Similar to https://github.com/discourse/discourse/pull/28061, merging topics with many posts can exceed the 30 seconds timeout that Unicorn workers are limited to, so we should move the operation into a background thread to get around this limit.
Internal topic: t/133710.
* SECURITY: Update default allowed iframes list
Change the default iframe url list to all include 3 slashes.
* SECURITY: limit group tag's name length
Limit the size of a group tag's name to 100 characters.
Internal ref - t/130059
* SECURITY: Improve sanitization of SVGs in Onebox
---------
Co-authored-by: Blake Erickson <o.blakeerickson@gmail.com>
Co-authored-by: Régis Hanol <regis@hanol.fr>
Co-authored-by: David Taylor <david@taylorhq.com>
Followup 4aea12fdcb
In certain config areas (like About) we want to be able
to fetch specific site settings by name. In this case,
sometimes we need to be able to fetch hidden settings,
in cases where a config area is still experimental.
Splitting out a different endpoint for this purpose
allows us to be stricter with what we return for config
areas without affecting the main site settings UI, revealing
hidden settings before they are ready.
`addCommunitySectionLink` API function accepts secondary argument to determine if the link should be added to the primary or secondary (more) section. There was a bug and all links were mounted in the secondary section.
We have a dedicated admin page (`/admin/customize/email_templates`) that lets admins customize all emails that Discourse sends to users. The way this page works is that it lists all translations strings that are used for emails, and the list of translation strings is currently hardcoded and hasn't been updated in years. We've had a number of new emails that Discourse sends, so we should add those templates to the list to let admins easily customize those templates.
Meta topic: https://meta.discourse.org/t/3-2-x-still-ignores-some-custom-email-templates/308203.
In this case, there is no 'nearPost' param in the URL. Instead, the server preloads a post-stream with whichever page of posts is requested. We can check for that situation using `postStream.firstPostPresent`.
Also updates the widget-header version to fetch a value from the service on initial render, instead of relying on the observer triggering.
Followup to bdec564d14
Currently, if MF definitions are missing (typically because there’s a
compilation error), `I18n.messageFormat` will try to access
`I18n._mfMessages.hasMessage` resulting in a crash that will in turn
crash Ember.
This patch addresses the issue by using the optional chaining operator
making the `I18n.messageFormat` method return a "Missing Key" message.
MF strings won’t be rendered properly, but the site will stay usable.
* FIX: Ensure JsLocaleHelper to obly outputs up-to-date translations
The old implementation forgot to filter out deprecated
translations, causing these translations to incorrectly override the new
locale in the frontend.
This commit fills in the forgotten where clause, filtering only the
up-to-date part.
Related meta topic: https://meta.discourse.org/t/outdated-translation-replacement-causing-missing-translation/314352
By default, the swc minifier seems to unwrap 'unneeded' IIFE. That means it was undoing the 'bugfix' transformation we have for class fields in Safari 15. Disabling the 'inline' and 'reduce_funcs' options seems to stop this behavior.
Currently, when adding translation overrides, values aren’t validated
for MF strings. This results in being able to add invalid plural keys or
even strings containing invalid syntax.
This patch addresses this issue by compiling the string when saving an
override if the key is detected as an MF one.
If there’s an error from the compiler, it’s added to the model errors,
which in turn is displayed to the user in the admin UI, helping them to
understand what went wrong.
When we show user tips, we immediately send an AJAX request to mark the
tiup as seen. This is done in the background. However, when system tests
are run, sometimes that request is not completed before the test ends.
This causes the test to be flakey.
One way to fix this is to force the system test run to wait for the AJAX
request to complete. However, this is not ideal because it makes the
test suite slower on each run.
Instead, this commit removes the flakey assertion and adds an alternative
assertion in the frontend tests that ensures the background request is
sent when the user tip is shown.
Form Kit is our new form library/framework for unifying the way forms look across Discourse. The admin config area for the /about page is a new form that isn't currently used, so it makes sense for it to be one of the first forms to be migrated to Form Kit to test the library.
Co-authored-by: Joffrey JAFFEUX <j.jaffeux@gmail.com>
* `@ember/owner` instead of `@ember/application`
* `discourse-i18n` instead of `I18n`
* `{ service } from "@ember/service"` instead of `inject as service`
In an attempt to improve build performance, 9db5eafb mistakenly removed minimization for some of our JS assets, leading to a significant increase in the size of some files.
This commit restores minimization to those files. To avoid regressing on the build time improvements, this commit switches to using the `webpack-terser-plugin`'s "swcMinify" option. On an entry-level 1CPU/1GB-ram/2GB-swap DO droplet, this commit increases build time from ~16 minutes to ~18 minutes.
Co-authored-by: Alan Guo Xiang Tan <gxtan1990@gmail.com>
Performing a bulk action on many topics can exceed the 30 seconds timeout that Unicorn workers have which results in the request failing and the operation getting aborted. To get around this 30 seconds timeout, we can push the operation into a background thread using the rack `hijack` API.
Internal topic: t/133779.
In development, classes are lazy loaded so `Jobs::Onceoff.onceoff_job_klasses`
may not have been set. This is not a problem in production cause stuff
is eager loaded.
Follow-up to f4d06f195d
* FIX: Add post id to the anchor to prevent two identical anchors
We generate anchors for headings in posts. This works fine if there is
only one post in a topic with anchors. The problem comes when you have
two or more posts with the same heading. PrettyText generates anchors
based on the heading text using the raw context of each post, so it is
entirely possible to generate the same anchor for two posts in the same
topic, especially for topics with template replies
Post1:
# heading
context
Post2:
# heading
context
When both posts are on the page at the same time, the anchor will only
work for the first post, according to the [HTML specification](https://html.spec.whatwg.org/multipage/browsing-the-web.html#scroll-to-the-fragment-identifier).
> If there is an a element in the document tree whose root is document
> that has a name attribute whose value is equal to fragment, then
> return the *first* such element in tree order.
This bug is particularly serious in forums with non-Latin languages,
such as Chinese. We do not generate slugs for Chinese, which results in
the heading anchors being completely dependent on their order.
```ruby
[2] pry(main)> PrettyText.cook("# 中文")
=> "<h1><a name=\"h-1\" class=\"anchor\" href=\"#h-1\"></a>中文</h1>"
```
Therefore, the anchors in the two posts must be in exactly the same by
order, causing almost all of the anchors in the second post to be
invalid.
This commit solves this problem by adding the `post_id` to the anchor.
The new anchor generation method will add `p-{post_id}` as a prefix when
post_id is available:
```ruby
[3] pry(main)> PrettyText.cook("# 中文", post_id: 1234)
=> "<h1><a name=\"p-1234-h-1\" class=\"anchor\" href=\"#p-1234-h-1\"></a>中文</h1>"
```
This way we can ensure that each anchor name only appears once on the
same topic. Using post id also prevents the potential possibility of the
same anchor name when splitting/merging topics.
We are investigating a memory leak in Sidekiq and saw the following line
when comparing heap dumps over time.
`Allocated IMEMO 14775 objects of size 591000/7389528 (in bytes) at:
/var/www/discourse/app/jobs/onceoff/onceoff.rb:36`
That line in question was doing a `.select { |klass| klass < self }` on
`ObjectSpace.each_object(Class)`. This for some reason is allocating a
whole bunch of `IMEMO` objects which are instruction sequence objects.
Instead of diving deeper into why this might be leaking, we can just
save our time by switching to an implementation that is more efficient
and does not require looping through a ton of objects.
Disabling webpack minimize is a bug we are working to resolve but we
have to consider self-hosters that deploy on low cost hardware
and reenabling this for them drastically increases the build time.
For now, add a `DISCOURSE_WEBPACK_MINIMIZE` env to allow sites to opt
back in.
