When the server gets overloaded and lots of requests start queuing server
will attempt to shed load by returning 429 errors on background requests.
The client can flag a request as background by setting the header:
`Discourse-Background` to `true`
Out-of-the-box we shed load when the queue time goes above 0.5 seconds.
The only request we shed at the moment is the request to load up a new post
when someone posts to a topic.
We can extend this as we go with a more general pattern on the client.
Previous to this change, rate limiting would "break" the post stream which
would make suggested topics vanish and users would have to scroll the page
to see more posts in the topic.
Server needs this protection for cases where tons of clients are navigated
to a topic and a new post is made. This can lead to a self inflicted denial
of service if enough clients are viewing the topic.
Due to the internal security design of Discourse it is hard for a large
number of clients to share a channel where we would pass the full post body
via the message bus.
It also renames (and deprecates) triggerNewPostInStream to triggerNewPostsInStream
This allows us to load a batch of new posts cleanly, so the controller can
keep track of a backlog
Co-authored-by: Joffrey JAFFEUX <j.jaffeux@gmail.com>
We were trying to observe a non-ember object which is undefined
behavior and was leaking to odd bugs. This replaces the `filter` object
with an Ember Object and things seem to work.
I also took the opportunity with this commit to move some test specific
stuff out of `discourse-loader` which is loaded on the front end of the
application. The test module building now happens in the `test_helper`
bundle.
This misses a test because Favcount doesn't exposes a get to the counter.
Also, since this code deals with all possible notifications configs we support:
- favicon notification
- favicon new content
- title notification
- title new content
the code is a bit complicated to follow. We may look into refactoring it when a
good opportunity arises, like if https://w3c.github.io/badging/ setClientBadge() method
gives us a cleaner way to notify users.
We can't use erb in Ember CLI (since it does not have Ruby) so this has
been ported to use our `javascript:update_constants` rake test instead.
Note we don't have to run this every time a notification type as it's
only used by fixtures to fill in some specific types we test against.
This is long overdue. We had a lot of (not linted) code to initialize
our test suite as part of the Ruby `test_helper.js` bundle.
This refactor moves that out to a `setup-tests` module, which imports
all the modules properly, rather than using `require`.
It also removes the global `server` variable which some tests were using
for pretender. Those tests are fixed, and in the case of widget tests,
support for a `pretend()` was added, which mimics our acceptance tests.
One problematic test was removed, which overwrites `/posts` - this could
break tons of other tests depending on order.
The prefixing logic is moved into a `prefixProtocol` function in lib:url.
This commit also renames an incorrectly named test and uses https as default instead of http, in 2020 it's reasonable to think we most likely want https and not http. User can still specify http if required.
This commit is also moving one test to a component test.
A followup to this commit would be to ensure every dropdowns are using a regex instead of the normalize/lowercase system we have now.
These are tricky because `module.exports` is used by nodejs files as a
global, which is OK. But we don't want to allow `module` in JS tests
for qunit without importing it first.
We used many global functions to handle tests when they should be
imported like other libraries in our application. This also gets us
closer to the way Ember CLI prefers our tests to be laid out.
See https://meta.discourse.org/t/changing-a-users-email/164512 for additional context.
Previously when an admin user changed a user's email we assumed that they would need a password reset too because they likely did not have access to their account. This proved to be incorrect, as there are other reasons a user needs admin to change their email. This PR:
* Changes the admin change email for user flow so the user is sent an email to confirm the change
* We now record who the email change request was requested by
* If the requested by user is admin and not the user we note this in the email sent to the user
* We also make the confirm change email route open to anonymous users, so it can be clicked by the user even if they do not have access to their account. If there is a logged in user we make sure the confirmation matches the current user.
Allows site administrators to pick different fonts for headings in the wizard and in their site settings. Also correctly displays the header logos in wizard previews.
allowEmails used to always be set to true and did not use
can_invite_via_email, which checks for enable_local_logins.
It was a problem because on sites with local logins
disabled users were allowed to enter email addresses, but
received a generic error "error inviting that user".
Prior to this fix, weekly could be 8 days and we could have differences between period chooser text and actual results in the chart.
A good followup to this PR would be to add custom date ranges in period-chooser component.
This is where they should be as far as ember is concerned. Note this is
a huge commit and we should be really careful everything continues to
work properly.
You can now create a file in your plugin/theme in the `api-initializers`
directory which has a simpler template than previous initializers.
Example:
```
// api-initializers/my-plugin.js
import { apiInitializer } from "discourse/lib/api";
export default apiInitializer("0.8", api => {
console.log("hello world from api initializer!");
});
```
See https://meta.discourse.org/t/changing-a-users-email/164512 for context.
When admin changes an email for a user, we were incorrectly sending the password reset email to the user's old address. Also the new email does not come into effect until the reset password process is done, so this PR adds some notes to the admin to make this clearer.
Context: https://meta.discourse.org/t/stop-appending-username-to-url-when-badge-system-disabled/164904
Share links for topics and posts have the current username appended to them when the user is logged in.
If badges are disabled, those added usernames are not beneficial since they're only used to track progress towards certain badges.
This PR makes two change.
1. it moves the logic for the share link url to a centralized helper because it's shared in both topic and post models.
2. it stops username params from being added to share links when badges are disabled.
* FIX: allow editing in composer before a category is selected
If the site setting `allow_uncategorized_topics` is disabled, but no site categories have a topic template, the textarea of the composer should be enabled. This allows for a post body to be entered, however the post/topic should not be successfully created until a category is selected.
If `allow_uncategorized_topics` is disabled *and* one or more categories have a topic template, the textarea of the composer should remain disabled until a category is chosen (so that the body of their post isn’t overwritten by the template).
* Avoid recalculating hasTopicTemplates repeatedly
This replaces an invalid `window.scrollTop` occurrence, left over from a
jQuery refactoring. It should fix an undesirable "shake effect" I have
run into in Safari when navigating to a reply in the post stream.
eslint --fix is capable of fix it automatically for you, ensure prettier is run after eslint as eslint --fix could leave the code in an invalid prettier state.
It didn't provide much value to users as it often didn't match the number visible in the topic timeline. That's because `postNumber` is an implementation detail, while the timeline counts only full visible posts (no whispers, topic action notices, etc.)
typographer can change " to ” leading to breakages in parser
At least codify this. Longer term we want to re-prioritize typographer so
it always runs after bbcode parsing.
Previously attributes such as `[test a='a"a' b="a'a"]` were not correctly
handled.
This amends the regex parser to ensure it correctly parses attributes
without breaking incorrectly on the first nested quote
We don't delegate `mouseMove` so this code has not been called in a long
time. It also seems that `screenX`/`screenY`/`lastX`/`lastY` was never
used so they were removed too.
Before this patch, discourse-markdown depended on the modules in its
bundle being defined in a specific order or it wouldn't load properly.
Now, any file in the bundle can export a `priority` const (default 0)
and files will be loaded in order of ascending priority instead. This
allows us to use a bundle packaged in any order we want.
To check if a post contains any embedded media, we look if the "image_sizes" attribute is present in the new post manager arguments.
We want to see one boxed links, but we only store the raw content of the post. To work around this, I extracted the onebox logic from the composer editor into a module.
This commit also:
- deprecate an old way of using actions for future removal
- removes all the hidden/block/... logic, as we just rely on parent being displayed or not
* FEATURE: Diffrentiate between group + individual mentions
This commit adds the necessary code for Discorse core to differentiate between group + individual mentions in the notification user panel and notification page.
It changes the group mention icon from `at` to `users` as well as adds context as to which group was mentioned in the topic.
The current situation could cause a transition on the button to end after/during modal has shown and causing the button to get focus again. Browsers would then refuse to switch focus.
This is a kinda convulted solution, but it's a general purpose solution which doesn't involve changing anything in plugins/themes or core templates.
* REFACTOR: reworks all the search-advanced-options panel
This commit includes the following changes:
- prevents any mutation of external (to the component) values
- get rid of observers
- uses @action
- minor UI tweaks
- dropped the unecessary debounce
- drops all the legacy code for badges/groups which is not being used
- replaces user-selector by user-chooser and improves multi-select to not show `search` if maximum has been reached
Most importantly this refactor should fix multiple bugs due to _update() being called multiple times if searchTerm was empty and other various bugs where some changes in searchTerm was not applied to the sidebar.
This PR removes the user reminder topic timers, because that system has been supplanted and improved by bookmark reminders. The option is removed from the UI and all existing user reminder topic timers are migrated to bookmark reminders.
Migration does this:
* Get all topic_timers with status_type 5 (reminders)
* Gets all bookmarks where the user ID and topic ID match
* Loops through the found topic timers
* If there is no bookmark for the OP of the topic, then we just create a bookmark with a reminder
* If there is a bookmark for the OP of the topic and it does **not** have a reminder set, then just
update it with the topic timer reminder
* If there is a bookmark for the OP of the topic with a reminder then just discard the topic timer
* Cancels all outstanding user reminder topic timers
* **Trashes (not deletes) all user reminder topic timers**
Notes:
* For now I have left the user reminder topic timer job class in place; this is so the jobs can be cancelled in the migration. It and the specs will be deleted in the next PR.
* At a later date I will write a migration to delete all trashed user topic timers. They are not deleted here in case there are data issues and they need to be recovered.
* A future PR will change the UI of the topic timer modal to make it look more like the bookmark modal.
* DEV - versions of JS files written to a JS file to be included by load-script and appended as params to URLs
* Formatting
* Incorporate feedback from PR
* Update filename of public-js-versions
Before this commit, onShow code could be impacted by code happening after the onShow call. This should ensure this code works for example:
```
onShow() {
afterRender(() => {
someInput.focus();
})
}
```
This moves the logic for horizontally placing the topic progress wrapper from the JS component to SCSS. Doing so means it is more easily overridable by themes and plugins.
This also changes the left/right spacing from 1em to 2em for non-mobile screens (it fits better on iPad portrait especially).
Admins can currently add the bookmarks discovery route link
to the homepage interface, but users can't presently select
that as their default home view. This change facilitates that,
adding the option to the existing Default Home Page dropdown on
the User Preferences Interface page.
Incorrect search filters like `in:personalasd` will end up checking the
checkbox for `in:personal` because the regexp used was only doing prefix
matching.
When a category is removed from `auto_watch_category` we are removing
CategoryUser. However, there are still TopicUser with notification level
set to `watching` which was inherited from Category.
We should move them back to `regular` unless they were modified by a user.
This helps us out in a few ways:
1. It lessens our reliance on jQuery
2. It's slightly less code because it omits options we don't use
3. It is one less library to import and put into ES6 modules
didRender will be called each time the widget is rendered
willRerender will be called the second time a widget is rendered to give an opportunity to clean some state before the tree is replaced
* Category Page / Tag Class Additions
These additions add:
-the name of the tag as a css class to discourse tags.
-'pinned' as a css class to pinned topics in the category view
eg repro before:
- visit http://pr-discourse.test/top/weekly?f=foo
- select another period in the period chooser
- f=foo was gone
After this commit it should still be present
This indication covers all cases of network errors, not just "cancelled by user".
The post upload component already has its own handling for user-cancelled uploads, but the generic upload component does not.
Tested by stopping my localhost server right before attempting to upload a file.
