PostAnalyzer and CookedPostProcessor both replace URLs with oneboxes.
PostAnalyzer did not use the max_oneboxes_per_post site and setting and
CookedPostProcessor replaced at most max_oneboxes_per_post URLs ignoring
the oneboxes that were replaced already by PostAnalyzer.
Follow-up to eb237e634a.
Some `{{topic-list}}` instances, like the one for suggested topics, opt out of focusing the row of the last visited topic in the list, but we currently still add listeners for focus/blur events even if when the topic-list instance opts out. This commit adds a check so that we only register focus/blur listeners if the topic-list opts in for last visited topic focus.
Another attempt at fixing https://meta.discourse.org/t/discourse-with-a-screen-reader/178105/88?u=osama. Previous PR (reverted): #16240.
The problems with the previous PR were:
1. As you scrolled down a topics list, the first topic of every new batch of topics would receive focus and the indicator would show up.
2. Similar to 1, clicking the `See X new or updated topics` notice would also focus a random topic from the new topics that were just loaded.
3. Topics in the suggested topics list received focus too
4. Our custom focus indicator appeared on mobile, but it shouldn't.
This commit should have none of these problems.
This commit is a redo of2f1ddadff7dd47f824070c8a3f633f00a27aacde
which we reverted because it blew up an internal CI check. I looked
into it, and it happened because the old migration to add the bookmark
columns still existed, and those columns were dropped in a post migrate,
so the two migrations to add the columns were conflicting before
the post migrate was run.
------
This commit only includes the creation of the new columns and index,
and does not add any triggers, backfilling, or new data.
A backfill will be done in the final PR when we switch this over.
Intermediate PRs will look something like this:
Add an experimental site setting for using polymorphic bookmarks,
and make sure in the places where bookmarks are created or updated
we fill in the columns. This setting will be used in subsequent
PRs as well.
Listing and searching bookmarks based on polymorphic associations
Creating post and topic bookmarks using polymorphic associations,
and changing special for_topic logic to just rely on the Topic
bookmarkable_type
Querying bookmark reminders based on polymorphic associations
Make sure various other areas like importers, bookmark guardian,
and others all rely on the associations
Prepare plugins that rely on the Bookmark model to use polymorphic
associations
The final core PR will remove all the setting gates and switch over
to using the polymorphic associations, backfill the bookmarks
table columns, and ignore the old post_id and for_topic colummns.
Then it will just be a matter of dropping the old columns down the
line.
This reverts commit 5d77f485cb.
There are some edge cases that we need to handle better. Reverting this
commit because we're going to do a beta release later today.
This is done by defining a `/all` route for use when a category's default filter is 'none'. This was defined for regular category routes in 3e7f7fdd, but not for tag routes.
This commit also corrects the route name TagsShowNoneCategory*Route -> TagsShowCategoryNone*Route, which fixes an error when setting subcategories=none while filtering by tags.
Previously we would issue a 403 for all invalid routes under `/tags/c/...`, which is not semantically correct. In some cases, these 403'd routes would then be handled successfully in the Ember app, leading to some very confusing behavior.
These methods have been natively supported in all our target browsers for many years. We're now feature-detecting `String.prototype.replaceAll`, which is a much more recent addition. If a browser has `replaceAll`, it'll have `padStart` and `padEnd`
This commit is a redo of e21c640a3c
which we reverted to not include half-done work in a release.
This commit is slightly different though, in that it only includes
the creation of the new columns and index, and does not add any
triggers, backfilling, or new data.
A backfill will be done in the final PR when we switch this over.
Intermediate PRs will look something like this:
1. Add an experimental site setting for using polymorphic bookmarks,
and make sure in the places where bookmarks are created or updated
we fill in the columns. This setting will be used in subsequent
PRs as well.
2. Listing and searching bookmarks based on polymorphic associations
3. Creating post and topic bookmarks using polymorphic associations,
and changing special for_topic logic to just rely on the Topic
bookmarkable_type
4. Querying bookmark reminders based on polymorphic associations
5. Make sure various other areas like importers, bookmark guardian,
and others all rely on the associations
6. Prepare plugins that rely on the Bookmark model to use polymorphic
associations
The final core PR will remove all the setting gates and switch over
to using the polymorphic associations, backfill the bookmarks
table columns, and ignore the old post_id and for_topic colummns.
Then it will just be a matter of dropping the old columns down the
line.
Plugins like chat add custom score type to override the title in the UI, but that should be reserved for situations when you need to manage the flag priority separately, which is configurable in the queue settings page.
Currently, if a plugin creates a custom score type, it won't be able to associate a priority, so there's no real gain from doing so. Priorities are tightly related to post-action types, which is something we might want to revise. For now, this change lets plugins move away from custom score types without compromises.
Previously we were loading almost all the data in an afterModel hook, storing it temporarily in route properties, and then passing it to the controller in `setupController`.
This does not follow Ember best-practices, and causes a number of unexpected behaviours. For example, Ember only calls `setupController` **when the model value changes**. Since `model()` was only returning the tag, that meant that category changes and `additionalTag` changes wouldn't always trigger a `setupController` call, and things would get into a very weird state. This is visible when using the 'loading-slider' component because the category navigation dropdown gets 'stuck' when switching categories.
This commit moves all the data-fetching into `model()`. To make things cleaner, it also:
- removes most uses of route-level variables
- introduces async/await in the model() function
- removes some unneeded `get()` usage
- re-uses DiscoverySortableController for queryParam default handling
- Removes override of `renderTemplate()` so that queryParams are correctly passed through to the controller
- Removes some `transitionToRoute` hacks which were working around the queryParams issue
- Switches to `@action`
In certain cases (like chat quoting) we need to be able
to call the API with an async AJAX call before copying
the results to the clipboard. The only way to reliably
do this is by handing off the AJAX promise to a ClipboardItem.
This commit introduces a new clipboardCopyAsync function
to handle this, which will stand alongside the existing
clipboardCopy function which can be used when no AJAX
request is necessary.
The meaning of reminder_at and reminder_last_sent_at changed after
commit 6d422a8033. A bookmark reminder
will fire only if reminder_last_sent_at is null, but before that it
fired everytime reminder_at was set. This is no longer true because
sometimes reminder_at continues to exist even after a reminder fired.
When using the loading-slider, the component instance is re-used across different pages and so the didInsertElement/willDestroyElement hooks are not fired during page transitions. Instead, we can lean on `didReceiveAttrs`.
Similar fix to 87b98e2862
Note that the `scrollTop` feature is still problematic under the loading slider. That will need to be addressed in a future commit.
* DEV: Allow params to be passed on topic redirects
There are several places where we redirect a url to a standard topic url
like `/t/:slug/:topic_id` but we weren't always passing query parameters
to the new url.
This change allows a few more query params to be included on the
redirect. The new params that are permitted are page, print, and
filter_top_level_replies. Any new params will need to be specified.
This also prevents the odd trailing empty page param that would
sometimes appear on a redirect. `/t/:slug/:id.json?page=`
* rubocop: fix missing space after comma
* fix another page= reference
If a user copies a gif from a website into their clipboard and then
tries to paste it into the Discourse composer, we would only paste a
static single frame of the original gif. This happens because the
browser doesn't store the original image in the clipboard, but two
entries:
1. image/png with the frame of the copy moment
2. text/html with the markup of the gif img element
This commit adds an heuristic that detects this and makes us pick the
clipboard content of text/html instead of the image/png when this
happens.
From there our existing HTML paste logic handles and converts the HTML
img tag into markdown, preserving even the alt text.
See https://meta.discourse.org/t/-/218720 for context.
share-topic modal is used everywhere expect when clicking on the top
right corner of the post. This changes standardize on share-topic modal
and add the missing features from share-popup.
The `testem.scss` include triggers a live reload locally. We need these
styles when running `ember test --server`, so this loads that stylesheet
only in that scenario.
* FEATURE: use canonical links in posts.rss feed
Previously we used non canonical links in posts.rss
These links get crawled frequently by crawlers when discovering new
content forcing crawlers to hop to non canonical pages just to end up
visiting canonical pages
This uses up expensive crawl time and adds load on Discourse sites
Old links were of the form:
`https://DOMAIN/t/SLUG/43/21`
New links are of the form
`https://DOMAIN/t/SLUG/43?page=2#post_21`
This also adds a post_id identified element to crawler view that was
missing.
Note, to avoid very expensive N+1 queries required to figure out the
page a post is on during rss generation, we cache that information.
There is a smart "cache breaker" which ensures worst case scenario is
a "page drift" - meaning we would publicize a post is on page 11 when
it is actually on page 10 due to post deletions. Cache holds for up to
12 hours.
Change only impacts public post RSS feeds (`/posts.rss`)
* DEV: Deprecate /posts/:id/reply-ids/all
It was added in ed4c0c4a63 and its only use was removed in b58867b6e9
Nothing in all-the* seems to be using this endpoint.
* Update app/controllers/posts_controller.rb
Co-authored-by: Alan Guo Xiang Tan <gxtan1990@gmail.com>
Eventually we want to remove JQuery, but that's a long way off. Installing this package will stop ember-cli printing the deprecation warning on every boot
```
DEPRECATION: The integration of jQuery into Ember has been deprecated and will be removed with Ember 4.0. You can either opt-out of using jQuery, or install the `@ember/jquery` addon to provide the jQuery integration. Please consult the deprecation guide for further details: https://emberjs.com/deprecations/v3.x#toc_jquery-apis
```
The `blocked onebox domains` setting lets site owners change what sites
are allowed to be oneboxed. When a link is entered into a post,
Discourse checks the domain of the link against that setting and blocks
the onebox if the domain is blocked. But if there's a chain of
redirects, then only the final destination website is checked against
the site setting.
This commit amends that behavior so that every website in the redirect
chain is checked against the site setting, and if anything is blocked
the original link doesn't onebox at all in the post. The
`Discourse-No-Onebox` header is also checked in every response and the
onebox is blocked if the header is set to "1".
Additionally, Discourse will now include the `Discourse-No-Onebox`
header with every response if the site requires login to access content.
This is done to signal to a Discourse instance that it shouldn't attempt
to onebox other Discourse instances if they're login-only. Non-Discourse
websites can also use include that header if they don't wish to have
Discourse onebox their content.
Internal ticket: t59305.
In the API keys page where admins can create API keys with restricted scopes, each scope shows a list of URLs that it allows. But currently, this list of allowed URLs shows incomplete URLs for scopes that are added by plugins. For example, the allowed URL for the "run queries" scope of the data-explorer plugin is shown as `/queries/:id/run` when the correct URL for this scope is `/admin/plugins/explorer/queries/:id/run`. The first 3 segments of the path are the mount path of the plugin's engine and it's missing because the routes set of the engine doesn't include the mount path. To fix this, this commit gets the mount path and prepends it to the URL so the complete URL is shown to the user.
It's not possible to write tests for this change because plugins are not loaded in the test environment by default when core's tests suite is running.
Previously, if an admin user tried to add/remove
users to another user's ignored list, it would
be added to their own ignore list because the
controller used current_user. Now for admins only
a source_user_id parameter can be passed through,
which will be used to ignore the target user for
that source user.
The user can select what happens with a bookamrk after it expires. New
option allow bookmark's reminder to be kept even after it has expired.
After a bookmark's reminder notification is created, the reminder date
will be highlighted in red until the user resets the reminder date.
User can do that using the new Clear Reminder button from the dropdown.
The search_ignore_accents site setting can be used to make the search
indexer remove the accents before indexing the content. The unaccent
function from PostgreSQL is better than Ruby's unicode_normalize(:nfkd).
When creating files with create-multipart, if the file
size was somehow zero we were showing a very unhelpful
error message to the user. Now we show a nicer message,
and proactively don't call the API if we know the file
size is 0 bytes in JS, along with extra console logging
to help with debugging.
This change adds support for the categories endpoint to have an api
scope. Only adds GET scope for listing categories and for fetching a
single category.
See: https://meta.discourse.org/t/218080/4
This PR adds an extra description to the 2FA page when granting a user admin access. It also introduces a general system for adding customized descriptions that can be used by future actions.
(Follow-up to dd6ec65061)
Stop Discourse from prompting for push notification on latest iOS beta
where the navigators exposes a broken Push object.
We had some feature detection functions that where outside our
pre-initializer that it dedicated for this stuff. All feature detection
now lives on sniff-capabilities file.
Also removed some old browser detection from the push notifications
code, and simplified the function signature because of it.
Co-authored-by: Jarek Radosz <jradosz@gmail.com>
Discourse users and associated accounts are created or updated when a
user logins or connects the account using their account preferences.
This new API can be used to create associated accounts and users too,
if necessary.
Our @mention user search prioritized users based on prefix matches.
So if searching for `sa` we will display `sam`, `asam` in that order
Previously, we did not prioritize group matches based on prefix. This change ensures better parity.
Implementation notes:
1. User search only prioritizes based on username prefix, not name prefix. TBD if we want to change that.
2. @mention on client side will show 0 group matches if we fill up all the spots with user matches. TBD if we want to unconditionally show the first / second group match.
Co-authored-by: Alan Guo Xiang Tan <gxtan1990@gmail.com>
Include a calculation of offset in main - this allows docking
to be calculated (again) using global offsets, in case there is a
custom header outside of Discourse that pushes forum content down.
The current implementation ties the filter query params tightly to the
`summary` attribute on the post stream model making it hard to support
other filters.
This feature was rarely used, could be used for spamming users and was
impossible to add a context to why the user was notified of a topic. A
simple private messages that includes the link and personalized message
can be used instead.
* DEV: Show only top level replies
Adds a new query param to the topic view so that we can filter out posts
that aren't top level replies. If a post is a reply to another post
instead of the original topic post we should not include it in the
response if the `filter_top_level_replies` query param is present.
* add rspec test
Bianca Nenciu
Penar Musaraj
Jarek Radosz
Osama Sayegh
Martin Brennan
Mark VanLandingham
David Taylor
Alan Guo Xiang Tan
Robin Ward
Roman Rizzi
Blake Erickson
Rafael dos Santos Silva
Sam
Ella E
Ayke Halder
Joffrey JAFFEUX
Kris
Jordan Vidrine
Dan Ungureanu
Jeff Wong
Krzysztof Kotlarek