Why this change?
Previously `TopicsFilter` was designed in such a way that we act on a
filter sequentially based on the order it was matched. However, this
made it hard to support filters composition where a similar filter may
be present further in the query string. Because of this limitation, I
previously introduced a private API `TopicsFilter.register_scope` which
allows us to act on a filter only after the entire query string has been
scanned. However, I felt that it made the code complicated and hard to
reason about.
In thie commit, I've changed it such that we scan through the entire
query string and group the values of each filter together. This allows
us to act on the values of a given filter in one go which I find easier
to reason about. This also opens up the possibility for us to ignore
certain filters when it has been specified multiple times.
This commit adds support for the `created-by:<username>` query filter
which will return topics created by the specified user. Multiple
usernames can be specified by comma seperating the usernames like so:
`created-by:username1,username2`. This will filter for topics created by
either of the specified users. Multiple `created-by:<username>` can also
be composed together. `created-by:username1 created-by:username2` is
equivalent to `created-by:username1,username2`.
This was inadvertently removed in 4c46c7e. In very specific scenarios,
this could be used execute arbitrary JavaScript.
Only affects instances where SVGs are allowed as uploads and CDN is not
configured.
Previously, Discourse's password hashing was hard-coded to a specific algorithm and parameters. Any changes to the algorithm or parameters would essentially invalidate all existing user passwords.
This commit introduces a new `password_algorithm` column on the `users` table. This persists the algorithm/parameters which were use to generate the hash for a given user. All existing rows in the users table are assumed to be using Discourse's current algorithm/parameters. With this data stored per-user in the database, we'll be able to keep existing passwords working while adjusting the algorithm/parameters for newly hashed passwords.
Passwords which were hashed with an old algorithm will be automatically re-hashed with the new algorithm when the user next logs in.
Values in the `password_algorithm` column are based on the PHC string format (https://github.com/P-H-C/phc-string-format/blob/master/phc-sf-spec.md). Discourse's existing algorithm is described by the string `$pbkdf2-sha256$i=64000,l=32$`
To introduce a new algorithm and start using it, make sure it's implemented in the `PasswordHasher` library, then update `User::TARGET_PASSWORD_ALGORITHM`.
This commit adds support for the `in:<topic notification level>` query
filter. As an example, `in:tracking` will filter for topics that the
user is watching. Filtering for multiple topic notification levels can
be done by comma separating the topic notification level keys. For
example, `in:muted,tracking` or `in:muted,tracking,watching`.
Alternatively, the user can also compose multiple filters with `in:muted
in:tracking` which translates to the same behaviour as
`in:muted,tracking`.
This commit adds support for the `in:pinned` filter to the topics filtering
query language. When the filter is present, it will filter for topics
where `Topic#pinned_until` is greater than `Topic#pinned_at`.
This adds a SiteSetting, which when enabled, creates a small_action post for tag/category changes to the topic. It uses `topic.add_moderator_post, and passes raw text in, to describe the change.
This was introduced to the standard library in Ruby 2.4. In my testing, it produces the same result, and is around 8x faster than our pure-ruby implementation
Before this commit, composing multiple category filters with a query such as category:category1 and category:category2 would not return any results. This is because we were filtering for topics that belonged to both category1 and category2, which is impossible since a topic can only belong to a single category.
With this commit, specifying a query like category:category1 category:category2 will now translate to filtering for topics that belong to either the category1 or category2 category.
Invite only and Discourse connect could not be enabled at the same time
because of some legacy reason. This is a follow up commit to ce04db8,
355d51a and 40f6ceb.
Followup to #20915. If we're grouping search results that don't rely on core's search, we won't have access to pg headlines. This is now configurable via the constructor, defaulting to `SiteSetting.use_pg_headlines_for_excerpt`
We use schema_migration_details to determine the age of a site in multiple migrations. This commit moves the logic into a dedicated `Migration::Helpers` module so that it doesn't need to be re-implemented every time.
This commit adds support for filtering for topics in specific
subcategories via the categories filter query language.
For example: `category:documentation:admins` will filter for topics and
subcategory topics in
the category with slug "admins" whose parent category has the slug
"documentation".
The `=` prefix can also be used such that
`=category:documentation:admins` will exclude subcategory topics of the
category with slug "admins" whose parent category has the slug
"documentation".
On the `/filter` route, the categories filtering query language is now
supported in the input per the example provided below:
```
category:bug => topics in the bug category AND all subcategories
=category:bug => topics in the bug category excluding subcategories
category:bug,feature => allow for categories either in bug or feature
=category:bug,feature => allow for exact categories match excluding sub cats
categories: => alias for category
```
Currently composing multiple category filters is not supported as we
have yet to determine what behaviour it should result in. For example,
`category:bug category:feature` would now return topics that are in both
the `bug` and `feature` category but it is not possible for a topic to
belong to two categories.
Add a modifier that will allow us to tune the results returned by suggested.
At the moment the modifier allows us to toggle including random results.
This was created for the discourse-ai module. It needs to switch off random
results when it returns related topics.
Longer term we can use it to toggle unread/new and other aspects.
This also demonstrates how to test the contract when adding modifiers.
This was added in d3f02a1270
for hashtags but later removed usage in
b2acc416e7. It was removed because
serializing the user does not include things like their
secure_categories.
It is not used by any other plugins or themes, and can cause
issues where it will error when operating on a null user. Better
to just pass in the user_id and use it to look up a user
directly in a PrettyText::Helper
Similar to the _map added for group_list SiteSettings in
e62e93f83a, this commit adds
the same extension for simple and compact `list` type SiteSettings,
so developers do not have to do the `.to_s.split("|")` dance
themselves all the time.
For example:
```
SiteSetting.markdown_linkify_tlds
=> "com|net|org|io|onion|co|tv|ru|cn|us|uk|me|de|fr|fi|gov|ddd"
SiteSetting.markdown_linkify_tlds_map
=> ["com", "net", "org", "io", "onion", "co", "tv", "ru", "cn", "us", "uk", "me", "de", "fr", "fi", "gov"]
```
There is no need to validate the user's emails when
promoting/demoting their trust level, this can cause
issues in things like Jobs::Tl3Promotions, we don't
need to fail in that case when all we are doing is changing
trust level.
Introduces a new API for plugin data modification without class-based extension overhead.
This commit introduces a new API that allows plugins to modify data in cases where they return different data rather than additional data, as is common with filtered_registers in DiscoursePluginRegistry. This API removes the need for defining class-based extension points.
When a plugin registers a modifier, it will automatically be called if the plugin is enabled. The core will then modify the parameter sent to it using the block registered by the plugin:
```ruby
DiscoursePluginRegistry.register_modifier(plugin_instance, :magic_sum_modifier) { |a, b| a + b }
sum = DiscoursePluginRegistry.apply_modifier(:magic_sum_filter, 1, 2)
expect(sum).to eq(3)
```
Key features of these modifiers:
- Operate in a stack (first registered, first called)
- Automatically disabled when the plugin is disabled
- Pass the cumulative result of all block invocations to the caller
The following are the changes being introduced in this commit:
1. Instead of mapping the query language to various query params on the
client side, we've decided that the benefits of having a more robust
query language far outweighs the benefits of having a more human readable query params in the URL.
As such, the `/filter` route will just accept a single `q` query param
and the query string will be parsed on the server side.
1. On the `/filter` route, the tags filtering query language is now
supported in the input per the example provided below:
```
tags:bug+feature tagged both bug and feature
tags:bug,feature tagged either bug or feature
-tags:bug+feature excluding topics tagged bug and feature
-tags:bug,feature excluding topics tagged bug or feature
```
The `tags` filter can also be specified multiple
times in the query string like so `tags:bug tags:feature` which will
filter topics that contain both the `bug` tag and `feature` tag. More
complex query like `tags:bug+feature -tags:experimental` will also work.
This change sets the ground work for allowing us to filter topics list
by tags in the following ways:
1. Filter for topics that matches all tags in a given set of tags
2. Filter for topics that matches any tags in a given set of tags
3. Exclude topics that matches all tags in a given set of tags
4. Exclude topics that matches any tags in a given set of tags
Before, incorrectly filled fields were marked with red border. Now, additional information under the field is displayed to notify the user what is incorrect.
/t/93696
When we renamed the `default_categories_regular` to `default_categories_normal` we missed a site setting validation method. It allowed the duplicate category ids in `default_categories_normal` site setting and caused the problem in user registration process.
5176c689e9
When setting the ACL for optimized images after setting the
ACL for the linked upload (e.g. via the SyncACLForUploads job),
we were using the optimized image path as the S3 key. This worked
for single sites, however it would fail silently for multisite
sites since the path would be incorrect, because the Discourse.store.upload_path
was not included.
For example, something like this:
somecluster1/optimized/2X/1/3478534853498753984_2_1380x300.png
Instead of:
somecluster1/uploads/somesite1/2X/1/3478534853498753984_2_1380x300.png
The silent failure is still intentional, since we don't want to
break other things because of ACL updates, but now we will update
the ACL correctly for optimized images on multisite sites.
Sites with many categories and many of them in muted by default (see
`default_categories_muted`) reported bad performance when requesting
the homepage as an anonymous user. This was the case because of the
long query that iterated over topics and categories trying to remove
those from the muted categories.
Our SafeMigrate system is designed to prevent tables/columns being dropped in pre-deploy migrations. Its regex-based detection was triggering incorrectly on `ALTER COLUMN DROP NOT NULL`.
There are many situations that may cause users to lose permission to
send messages in a chat channel. Until now we have relied on security
checks in `Chat::ChatChannelFetcher` to remove channels which the
user may have a `UserChatChannelMembership` record for but which
they do not have access to.
This commit takes a more proactive approach. Now any of these following
`DiscourseEvent` triggers may cause `UserChatChannelMembership`
records to be deleted:
* `category_updated` - Permissions of the category changed
(i.e. CategoryGroup records changed)
* `user_removed_from_group` - Means the user may not be able to access the
channel based on `GroupUser` or also `chat_allowed_groups`
* `site_setting_changed` - The `chat_allowed_groups` was updated, some
users may no longer be in groups that can access chat.
* `group_destroyed` - Means the user may not be able to access the
channel based on `GroupUser` or also `chat_allowed_groups`
All of these are handled in a distinct service run in a background
job. Users removed are logged via `StaffActionLog` and then we
publish messages on a per-channel basis to users who had their
memberships deleted.
When the user has a channel they are kicked from open, we show
a dialog saying "You no longer have access to this channel".
When they click OK we redirect them either:
* To their first other public channel, if they have any followed
* The chat browse page if they don't
This is to save on tons of requests from kicked out users getting messages
from other channels.
When the user does not have the kicked channel open, we can just
silently yoink it out of their sidebar and turn off subscriptions.
There was a lot of duplication in the svg parsing and coercion code. This reduces that duplication and causes svg sprite parsing to happen earlier so that more computation is cached.
Our schema allows `category.topic_id` to be NULL. Null values shouldn't actually happen in production, but it is very common in tests because `Fabricate(:category)` skips creating the definition topic to improve performance. Before this commit, a NULL category.topic_id would cause all subcategory topics to be excluded from a TopicQuery result. This is because, in postgres, `NULL <> anything` is falsy. Instead, we can use `IS DISTINCT FROM`, which will return true when NULL is compared to a non-NULL value.
Why is this change required?
Prior to this change, we would list all group messages that a user
has access to in the user menu messages notifications panel dropdown.
However, this did not respect the topic's notification level setting and
group messages which the user has set to 'normal' notification level were
being displayed
What does this commit do?
With this commit, we no longer display all group messages that a user
has access to. Instead, we only display group messages that a user is
watching in the user menu messages notifications panel dropdown.
Internal Ref: /t/94392
Follow up to 4d2a95ffe6. Sometimes
due to the original UploadReference migration or other issues,
multiple UploadReference records can have the exact same
created_at date and time. To tiebreak and correct the SQL order
when this happens, we can add a secondary `id ASC` ordering
when we check for the first upload reference.
Way back when this was introduced way back in b96c10a903
I didn't have any frame of reference for what these max rate
limit numbers should be, so 10 seemed like a reasonable limit
until a real world case where this did not make sense came
along.
The time has come.
Moving these into site settings, which are hidden since in most
cases there is no need to change these.
Similar spirit to e195e6f614,
this moves the Bookmarkable registration to DiscoursePluginRegistry
so plugins which are not enabled do not register additional
bookmarkable classes.
When a theme setting of type `upload` has a default upload, it should return the URL of the specified default upload until a custom upload is used for the setting. However, currently this isn't the case and we get null instead of the default upload URL.
The reason for this is because the `super` method of `#value` already returns the default upload URL (if there's one), so we can't pass that to `cdn_url` which expects an upload ID:
c961dcc757/lib/theme_settings_manager.rb (L212)
This commit fixes the bug by skipping the call to `cdn_url` when we fallback to the default upload for the setting value.
* UX: add type tag and design update
* UX: clarify status copy in reviewQ
* DEV: switch to selectKit
* UX: color approve/reject buttons in RQ
* DEV: regroup actions
* UX: add type tag and design update
* UX: clarify status copy in reviewQ
* Join questions for flagged post with "or" with new I18n function
* Move ReviewableScores component out of context
* Add CSS classes to reviewable-item based on human type
* UX: add table header for scoring
* UX: don't display % score
* UX: prefix modifier class with dash
* UX: reviewQ flag table styling
* UX: consistent use of ignore icon
* DEV: only show context question on pending status
* UX: only show table headers on pending status
* DEV: reviewQ regroup actions for hidden posts
* UX: reviewQ > approve/reject buttons
* UX: reviewQ add fadeout
* UX: reviewQ styling
* DEV: move scores back into component
* UX: reviewQ mobile styling
* UX: score table on mobile
* UX: reviewQ > move meta info outside table
* UX: reviewQ > score layout fixes
* DEV: readd `agree_and_keep` and fix the spec tests.
* Fix the spec tests
* fix the quint test
* DEV: readd deleting replies
* UX: reviewQ copy tweaks
* DEV: readd test for ignore + delete replies
* Remove old
* FIX: Add perform_ignore back in for backwards compat
* DEV: add an action alias `ignore` for `ignore_and_do_nothing`.
---------
Co-authored-by: Martin Brennan <martin@discourse.org>
Co-authored-by: Vinoth Kannan <svkn.87@gmail.com>
Follow up to 098ab29d41. Since
we just used a `cattr_reader` on `About` this was not safe
for multisite, since some sites could have the chat plugin
enabled and some may not. Using `DiscoursePluginRegistry` gets
around this issue, and makes it so the chat stats only show
for a site if `chat_enabled` is true.
Sometimes we get Maps URL containing a zoom level as a float (17.5z and
not 17z) but this doesn’t work with our current onebox implementation.
While Google accepts those float zoom levels, it removes automatically
the floating part in the URL (thus when visiting a Maps URL containing
17.5z, the URL will be rewritten shortly after as 17z). When putting a
float zoom level in an embedded URL, this actually breaks (Maps API
returns a 400 error).
This patch addresses the issue by allowing the onebox engine to match on
a zoom level expressed as a float but we only keep the integer part thus
rendering properly maps.
The ensure_consistency rake task was not marking posted as true for post authors in the TopicUser table, post migration. Create another step to set posted='t'.
Forcing distributed muted to raise when a notify reviewable job is running
leads to excessive errors in the logs under many conditions.
The new pattern
1. Optimises the counting of reviewables so it is a lot faster
2. Holds the distributed lock for 2 minutes (max)
The downside is the job queue can get blocked up when tons of notify
reviewables are running at the same time. However this should be very
rare in the real world, as we only notify when stuff is flagged which
is fairly infrequent.
This also give a fair bit more time for the notifications which may be
a little slow on large sites with tons of mods.
This commit implements many changes to topic and comments embedding. It
deprecates the class_name field from EmbeddableHost and suggests using
the className parameter. discourse_username parameter has been
deprecated and it will fetch it from embedded site from the author or
discourse-username meta.
See the updated code sample from Admin > Customize > Embedding page.
* FEATURE: Add className parameter for Discourse embed
* DEV: Hide class_name from EmbeddableHost
* DEV: Deprecate class_name field of EmbeddableHost
* FEATURE: Use either author or discourse-username meta tag
* DEV: Deprecate discourse_username parameter
* DEV: Improve embed code sample
This commit introduces a few experimental changes to the New topics list and "Everything" link in the sidebar:
1. Make the New topics list include unread topics
2. Make the Everything section in the sidebar link to the New topics list (`/new`)
3. Remove "unread" or "new" text next to the count and keep the count
4. The count is a sum of new and unread topics counts
All of these of changes are behind an off-by-default feature flag. I've not written extensive tests for these changes because they're highly experimental.
Internal topic: t/77234.
When invoking e.g. `can_see?(Foo.new)`, the guardian checks if there's a method `#can_see_foo?` defined and if so uses that to determine whether the user can see it or not.
When such a method is not defined, the guardian currently returns `true`, but it is probably a better call (pun intended) to make it "safe by default" and return `false` instead. I.e. if you can't explicitly see it, you can't see it at all.
This change makes the change to `Guardian#can_see?` to fall back to `false` if no visibility check method is defined.
For `#can_see_user?` and `#can_see_tag?` we don't have any particular logic that prevents viewing. We previously relied on the implicit `true` value, but since that's now change to `false`, I have explicitly implemented these two methods in `UserGuardian` and `TagGuardian` modules. If in the future we want to add some logic for it, this would be the place.
To be clear, **the behaviour remains the same**, but the `true` value is now explicit rather than implicit.
We were only supporting the main name of each HighlightJS language. So, by default, you could not use `js` or `jsx` to highlight Javascript, given they are aliases for `javascript`.
This PR adds a list of aliases as a constant to core (built via a rake task), and then checks against the `highlighted_languages` site settings plus the list of aliases when processing a code block.
This commit 57caf08e13 broke
`bin/turbo_rspec` timing recording via `TurboTests::Runner`,
because we changed to using all `spec/*` folders except
`spec/system` as default for the runner, rather than
the old `['spec']` array, which is what `TurboTests::Runner`
was relying on to determine whether to record test run
time with `ParallelTests::RSpec::RuntimeLogger`.
Instead, we can just pass a new `use_runtime_info` boolean to the
runner class and use it when running against the default set of
spec files using `bin/turbo_rspec` and the turbo rspec rake task.
The current default timeout is hardcoded to 2 seconds which is proving
too low for certain cases, and resulting in sporadic timeouts due to slow DNS queries.
As of ba3f62f576, handlebars templates are colocated with js files so the path to hbs templates referenced by this rake task is no longer valid. This commit fixes the path to hbs templates and updates a couple of files that are generated by the rake task.
We call `post.update_uploads_secure_status` in both
`PostCreator` and `PostRevisor`. Only the former was checking
if `SiteSetting.secure_uploads?` was enabled, but the latter
was not. There is no need to enqueue the job
`UpdatePostUploadsSecureStatus` if secure_uploads is not
enabled for the site.
- Reduce duplication of terms in post index from unlimited to 6. This will
result in reduced index size and reduced weighting for posts containing
a huge amount of duplicate terms. (Eg: a post containing "sam sam sam sam
sam sam sam sam", will index as "sam sam sam sam sam sam", only including
the word up to 6 times.) This corrects a flaw where title weighting could
be ignored.
- Prioritize exact matches of words in titles. Our search always performs
a prefix match. However we want to give special weight to exact title matches
meaning that a search for "sum" will find topics such as "the sum of us" vs
"summer in spring".
- Pick up fixes to our search algorithm which are missing from old indexes.
Specifically pick up the fix that indexes URLs properly. (`https://happy.com`
was stemmed to `happi` in keywords and then was not searchable)
see also:
https://meta.discourse.org/t/refinements-to-search-being-tested-on-meta/254158
Indexing will take a while and work in batches, in the background.
* Add username and name_or_username variables to SystemMessage defaults
* Allow username and name variables on welcome_user email template overrides
* Satisfy linting
* Add test
This commit adds backend support for a new topics list that combines both the current unread and new topics lists. We're going to experiment with this new list (name TBD) internally and decide if this feature is something that we want to fully build.
Internal topic: t/77234.
Currently, clicking on the unsubscribe link with a `key` associated with a
deleted topic results in an HTTP 500 response.
This change fixes that by skipping any attempt to run topic related flow if topic
isn't present.
* FIX: do not notify admins on suppressed categories
Avoid notifying admins on categories where they are not explicitly members
in cases where SiteSetting.suppress_secured_categories_from_admin is
enabled.
This helps keep notification stream clean and avoids admins mistakenly
being invited to discussions that should be suppressed
This is a combined work of Martin Brennan, Loïc Guitaut, and Joffrey Jaffeux.
---
This commit implements a base service object when working in chat. The documentation is available at https://discourse.github.io/discourse/chat/backend/Chat/Service.html
Generating documentation has been made as part of this commit with a bigger goal in mind of generally making it easier to dive into the chat project.
Working with services generally involves 3 parts:
- The service object itself, which is a series of steps where few of them are specialized (model, transaction, policy)
```ruby
class UpdateAge
include Chat::Service::Base
model :user, :fetch_user
policy :can_see_user
contract
step :update_age
class Contract
attribute :age, :integer
end
def fetch_user(user_id:, **)
User.find_by(id: user_id)
end
def can_see_user(guardian:, **)
guardian.can_see_user(user)
end
def update_age(age:, **)
user.update!(age: age)
end
end
```
- The `with_service` controller helper, handling success and failure of the service within a service and making easy to return proper response to it from the controller
```ruby
def update
with_service(UpdateAge) do
on_success { render_serialized(result.user, BasicUserSerializer, root: "user") }
end
end
```
- Rspec matchers and steps inspector, improving the dev experience while creating specs for a service
```ruby
RSpec.describe(UpdateAge) do
subject(:result) do
described_class.call(guardian: guardian, user_id: user.id, age: age)
end
fab!(:user) { Fabricate(:user) }
fab!(:current_user) { Fabricate(:admin) }
let(:guardian) { Guardian.new(current_user) }
let(:age) { 1 }
it { expect(user.reload.age).to eq(age) }
end
```
Note in case of unexpected failure in your spec, the output will give all the relevant information:
```
1) UpdateAge when no channel_id is given is expected to fail to find a model named 'user'
Failure/Error: it { is_expected.to fail_to_find_a_model(:user) }
Expected model 'foo' (key: 'result.model.user') was not found in the result object.
[1/4] [model] 'user' ❌
[2/4] [policy] 'can_see_user'
[3/4] [contract] 'default'
[4/4] [step] 'update_age'
/Users/joffreyjaffeux/Code/pr-discourse/plugins/chat/app/services/update_age.rb:32:in `fetch_user': missing keyword: :user_id (ArgumentError)
from /Users/joffreyjaffeux/Code/pr-discourse/plugins/chat/app/services/base.rb:202:in `instance_exec'
from /Users/joffreyjaffeux/Code/pr-discourse/plugins/chat/app/services/base.rb:202:in `call'
from /Users/joffreyjaffeux/Code/pr-discourse/plugins/chat/app/services/base.rb:219:in `call'
from /Users/joffreyjaffeux/Code/pr-discourse/plugins/chat/app/services/base.rb:417:in `block in run!'
from /Users/joffreyjaffeux/Code/pr-discourse/plugins/chat/app/services/base.rb:417:in `each'
from /Users/joffreyjaffeux/Code/pr-discourse/plugins/chat/app/services/base.rb:417:in `run!'
from /Users/joffreyjaffeux/Code/pr-discourse/plugins/chat/app/services/base.rb:411:in `run'
from <internal:kernel>:90:in `tap'
from /Users/joffreyjaffeux/Code/pr-discourse/plugins/chat/app/services/base.rb:302:in `call'
from /Users/joffreyjaffeux/Code/pr-discourse/plugins/chat/spec/services/update_age_spec.rb:15:in `block (3 levels) in <main>'
```
The #pluck_first freedom patch, first introduced by @danielwaterworth has served us well, and is used widely throughout both core and plugins. It seems to have been a common enough use case that Rails 6 introduced it's own method #pick with the exact same implementation. This allows us to retire the freedom patch and switch over to the built-in ActiveRecord method.
There is no replacement for #pluck_first!, but a quick search shows we are using this in a very limited capacity, and in some cases incorrectly (by assuming a nil return rather than an exception), which can quite easily be replaced with #pick plus some extra handling.
Ember CLI will automatically run babel transformations in parallel when the config is 'serializable', and can therefore be applied in multiple processes automatically. If any plugin is defined in an unserializable way, parallelisation will be disabled.
Our discourse-widget-hbs transformer was causing parallelisation to be disabled. This commit fixes that, and also enables the throwUnlessParallelizable flag so that we catch this kind of issue more easily in future.
This commit also refactors our deprecation silencing system into its own file, and uses a fake babel plugin to ensure deprecations are silenced in babel worker processes.
In our GitHub CI jobs, this doubles the speed of ember builds (1m30s -> 45s). It should also improve production deploy times, and cold-start dev builds.
This PR is a major change to Sass compilation in Discourse.
The new version of sass-ruby moves to dart-sass putting we back on the supported version of Sass. It does so while keeping compatibility with the existing method signatures, so minimal change is needed in Discourse for this change.
This moves us
From:
- sassc 2.0.1 (Feb 2019)
- libsass 3.5.2 (May 2018)
To:
- dart-sass 1.58
This update applies the following breaking changes:
>
> These breaking changes are coming soon or have recently been released:
>
> [Functions are stricter about which units they allow](https://sass-lang.com/documentation/breaking-changes/function-units) beginning in Dart Sass 1.32.0.
>
> [Selectors with invalid combinators are invalid](https://sass-lang.com/documentation/breaking-changes/bogus-combinators) beginning in Dart Sass 1.54.0.
>
> [/ is changing from a division operation to a list separator](https://sass-lang.com/documentation/breaking-changes/slash-div) beginning in Dart Sass 1.33.0.
>
> [Parsing the special syntax of @-moz-document will be invalid](https://sass-lang.com/documentation/breaking-changes/moz-document) beginning in Dart Sass 1.7.2.
>
> [Compound selectors could not be extended](https://sass-lang.com/documentation/breaking-changes/extend-compound) in Dart Sass 1.0.0 and Ruby Sass 4.0.0.
SCSS files have been migrated automatically using `sass-migrator division app/assets/stylesheets/**/*.scss`
Allows users to configure their own custom sidebar sections with links withing Discourse instance. Links can be passed as relative path, for example "/tags" or full URL.
Only path is saved in DB, so when Discourse domain is changed, links will be still valid.
Feature is hidden behind SiteSetting.enable_custom_sidebar_sections. This hidden setting determines the group which members have access to this new feature.
Previously due to an error archived topics were more prominent in search
than closed topics.
This amends our internal logic to ensure archived topics are bumped down
the list.
Previously to_tsquery would split terms and join with &
In PG 14 terms are split and use <-> which means followed directly by.
In PG 13:
discourse_test=# SELECT to_tsquery('english', '''hello world''');
to_tsquery
---------------------
'hello' & 'world'
(1 row)
In PG 14:
discourse_test=# SELECT to_tsquery('english', '''hello world''');
to_tsquery
---------------------
'hello' <-> 'world'
(1 row)
Change is very unobtrosive, we simply amend our to_tsquery to behave like
it used to behave and make no use of the `<->` operator
More detail at: https://akorotkov.github.io/blog/2021/05/22/pg-14-query-parsing/
Note that plainto_tsquery used elsewhere in Discourse keeps the exact
same function.
This also corrects a faulty test that was passing by a fluke on older
version of PG
We've had a couple of problems with the R2 gem where it generated a broken RTL CSS bundle that caused a badly broken layout when Discourse is used in an RTL language, see a3ce93b and 5926386. For this reason, we're replacing R2 with `rtlcss` that can handle modern CSS features better than R2 does.
`rltcss` is written in JS and available as an npm package. Calling the `rltcss` from rubyland is done via the `rtlcss_wrapper` gem which contains a distributable copy of the `rtlcss` package and loads/calls it with Mini Racer. See https://github.com/discourse/rtlcss_wrapper for more details.
Internal topic: t/76263.
`--d-hover` is calculated to be equivalent to primary-100 in light mode, or primary-low in dark mode
`--d-selected` is calculated to be equivalent to primary-low in light mode, or primary-100 in dark mode
`lib/color_math` is introduced to provide some utilities for making these calculations.
The new `prioritize_exact_search_match` can be used to force the search
algorithm to prioritize exact term matches in title when ranking results.
This is scoped narrowly to titles for cases such as a topic titled:
"organisation chart" and a search of "org chart".
If we scoped this wider, all discussion about "org chart" would float to
the top and leave a very common title de-prioritized.
This is a hidden site setting and it has some performance impact due
to double ranking.
That said, performance impact is somewhat mitigated cause ranking on
title alone is a very cheap operation.
Many users seems surprised by prefix matching in search leading to
unexpected results.
Over the years we always would return results starting with a search term
and not expect exact matches.
Meaning a search for `abra` would find `abracadabra`
This introduces the Site Setting `enable_search_prefix_matching` which
defaults to true. (behavior unchanged)
We plan to experiment on select sites with exact matches to see if the
results are less surprising
* FIX: Ensure soft-deleted topics can be deleted
The topic was not found during the deletion process because it was
deleted and `@post.topic` was nil.
* DEV: Use @topic instead of finding the topic every time
Under some situations, we would inadvertently return a public (unauthenticated) result to an authenticated API request. This commit adds the `Api-Key` header to our anonymous cache bypass logic.
Under scenarios of extremely high load where large numbers of `Reviewable*` items are being created, it has been observed that multiple instances of the `NotifyReviewable` job may run simultaneously.
These jobs will work satisfactorily if the concurrency is limited to 1, and the different types of jobs (items reviewable by admins, vs moderators, vs particular groups, etc.) are run eventually.
This change introduces a new option to `DistributedMutex` which allows the `max_get_lock_attempts` to be specified. If the number is exceeded an error will be raised, which will cause Sidekiq to requeue the job. Sidekiq has existing logic to back-off on retry times for jobs that have failed multiple times.
There was an issue where if hashtag-cooked HTML was sent
to the ExcerptParser without the keep_svg option, we would
end up with empty </use> and </svg> tags on the parts of the
excerpt where the hashtag was, in this case when a post
push notification was sent.
Fixed this, and also added a way to only display a plaintext
version of the hashtag for cases like this via PrettyText#excerpt.
When checking whether an existing upload should be secure
based on upload references, do not count deleted posts, since
there is still a reference attached to them. This can lead to
issues where e.g. an upload is used for a post then later on
a custom emoji.
Currently, `Tag#topic_count` is a count of all regular topics regardless of whether the topic is in a read restricted category or not. As a result, any users can technically poll a sensitive tag to determine if a new topic is created in a category which the user has not excess to. We classify this as a minor leak in sensitive information.
The following changes are introduced in this commit:
1. Introduce `Tag#public_topic_count` which only count topics which have been tagged with a given tag in public categories.
2. Rename `Tag#topic_count` to `Tag#staff_topic_count` which counts the same way as `Tag#topic_count`. In other words, it counts all topics tagged with a given tag regardless of the category the topic is in. The rename is also done so that we indicate that this column contains sensitive information.
3. Change all previous spots which relied on `Topic#topic_count` to rely on `Tag.topic_column_count(guardian)` which will return the right "topic count" column to use based on the current scope.
4. Introduce `SiteSetting.include_secure_categories_in_tag_counts` site setting to allow site administrators to always display the tag topics count using `Tag#staff_topic_count` instead.
This fixes a longstanding issue for sites with the
secure_uploads setting enabled. What would happen is a scenario
like this, since we did not check all places an upload could be
linked to whenever we used UploadSecurity to check whether an
upload should be secure:
* Upload is created and used for site setting, set to secure: false
since site setting uploads should not be secure. Let's say favicon
* Favicon for the site is used inside a post in a private category,
e.g. via a Onebox
* We changed the secure status for the upload to true, since it's been
used in a private category and we don't check if it's originator
was a public place
* The site favicon breaks :'(
This was a source of constant consternation. Now, when an upload is _not_
being created, and we are checking if an existing upload should be
secure, we now check to see what the first record in the UploadReference
table is for that upload. If it's something public like a site setting,
then we will never change the upload to `secure`.
Prior to this change, we were parsing `Post#cooked` every time we
serialize a post to extract the usernames of mentioned users in the
post. However, the only reason we have to do this is to support
displaying a user's status beside each mention in a post on the client side when
the `enable_user_status` site setting is enabled. When
`enable_user_status` is disabled, we should avoid having to parse
`Post#cooked` since there is no point in doing so.
Since the new hashtag format has been added, we want site
admins to be able to rebake old posts with the old hashtag
format. This can now be done with `rake hashtags:mark_old_format_for_rebake`
which goes and marks posts with the old cooked version of hashtags
in this format for rebake:
```
<a class=\"hashtag\" href=\"/c/ux/14\">#<span>ux</span></a>
```
c.f. https://meta.discourse.org/t/what-rebake-is-required-for-the-new-autocomplete-styling/249642/12
This commit fixes the following issue:
* User creates a post
* Akismet or some other thing like requiring posts to be approved puts
the post in the review queue, deleting it
* Admin approves the post
* Email is never sent to mailing list mode subscribers
We intentionally do not enqueue this for every single post when
recovering a topic (i.e. recovering the first post) since the topics
could have a lot of posts with emails already sent, and we don't want
to clog sidekiq with thousands of notify jobs.
TL4 users can already list and unlist topics, but they can't see
the unlisted topics. This change brings this to par by allowing
TL4 users to also see unlisted topics.
So it can easily be overwritten in a plugin for example.
### Added more tests to provide better coverage
We previously only had `u.silenced_till IS NULL` but I made it consistent with pretty much every other places where we check for "active" users.
These two new lines do change the query a tiny bit though.
**Before**
- You could not get the badge if you were currently silenced (no matter what period is being checked)
- You could get the badge if you were suspended 😬
**After**
- You can't get the badge if you were silenced during the past year
- You can't get the badge if you were suspended during the past year
### Improved the performance of the query by using `NOT EXISTS` instead of `LEFT JOIN / COUNT() = 0`
There is no difference in behaviour between
```sql
LEFT JOIN user_badges AS ub ON ub.user_id = u.id AND ...
[...]
HAVING COUNT(ub.*) = 0
```
and
```sql
NOT EXISTS (SELECT 1 FROM user_badges AS ub WHERE ub.user_id = u.id AND ...)
```
The only difference is performance-wise. The `NOT EXISTS` is 10-30% faster on very large databases (aka. posts and users in X millions). I checked on 3 of the largest datasets I could find.
When the thread is aborted, an exception is raised before the `start` of a job is set, and therefore raises an exception in the `ensure` block. This commit checks that `start` exists, and also adds `abort_on_exception=true` so that this issue would have caused test failures.
The `enable_new_notifications_menu` site setting allows sites that have
`navigation_menu` set to `legacy` to use the redesigned notifications
menu before switching to the new sidebar navigation menu.
This is a very subtle one. Setting the redirect URL is done by passing
a hash through a Discourse event. This is broken on Ruby 2 since the
support for keyword arguments in events was added.
In Ruby 2 the last argument is cast to keyword arguments if it is a
hash. The key point here is that creates a new copy of the hash, so
what the plugin is modifying is not the hash that was passed.
This patch introduces a cookies rotator as indicated in the Rails
upgrade guide. This allows to migrate from the old SHA1 digest to the
new SHA256 digest.
This will give us some aggregate stats on the defer queue performance.
It is limited to 100 entries (for safety) which is stored in an LRU cache.
Scheduler::Defer.stats can then be used to get an array that denotes:
- number of runs and completions (queued, finished)
- error count (errors)
- total duration (duration)
We can look later at exposing these metrics to gain visibility on the reason
the defer queue is clogged.
There was an issue with channel archiving, where at times the topic
creation could fail which left the archive in a bad state, as read-only
instead of archived. This commit does several things:
* Changes the ChatChannelArchiveService to validate the topic being
created first and if it is not valid report the topic creation errors
in the PM we send to the user
* Changes the UI message in the channel with the archive status to reflect
that topic creation failed
* Validate the new topic when starting the archive process from the UI,
and show the validation errors to the user straight away instead of
creating the archive record and starting the process
This also fixes another issue in the discourse_dev config which was
failing because YAML parsing does not enable all classes by default now,
which was making the seeding rake task for chat fail.
* Firefox now finally returns PerformanceMeasure from performance.measure
* Some TODOs were really more NOTE or FIXME material or no longer relevant
* retain_hours is not needed in ExternalUploadsManager, it doesn't seem like anywhere in the UI sends this as a param for uploads
* https://github.com/discourse/discourse/pull/18413 was merged so we can remove JS test workaround for settings
Currently when generating a onebox for Discourse topics, some important
context is missing such as categories and tags.
This patch addresses this issue by introducing a new onebox engine
dedicated to display this information when available. Indeed to get this
new information, categories and tags are exposed in the topic metadata
as opengraph tags.
If unaccent is called with quote-like Unicode characters then it can
generate invalid queries because some of the transformed quotes by
unaccent are not escaped and to_tsquery fails because of bad input.
This commits replaces more quote-like Unicode characters before
unaccent is called.
Fixes the support for kwargs in `DiscourseEvent.trigger()` on Ruby 3, e.g.
```rb
DiscourseEvent.trigger(:before_system_message_sent, message_type: type, recipient: @recipient, post_creator_args: post_creator_args, params: method_params)
```
Fixes https://github.com/discourse/discourse-local-site-contacts
If a secure upload's access_control_post was trashed, and an anon user
tried to look at that upload, they would get a 500 error rather than
the correct 403 because of an error inside the PostGuardian logic.
This commit does a couple of things:
1. Changes the limit of tags to include a subject for a
notification email to the `max_tags_per_topic` setting
instead of the arbitrary 3 limit
2. Adds both an X-Discourse-Tags and X-Discourse-Category
custom header to outbound emails containing the tags
and category from the subject, so people on mail clients
that allow advanced filtering (i.e. not Gmail) can filter
mail by tags and category, which is useful for mailing
list mode users
c.f. https://meta.discourse.org/t/headers-for-email-notifications-so-that-gmail-users-can-filter-on-tags/249982/17
This commit fixes an issue where the chat message bookmarks
did not respect the user's `bookmark_auto_delete_preference`
which they select in their user preference page.
Also, it changes the default for that value to "keep bookmark and clear reminder"
rather than "never", which ends up leaving a lot of expired bookmark
reminders around which are a pain to clean up.
When sending emails out via group SMTP, if we
are sending them to non-staged users we want
to mask those emails with BCC, just so we don't
expose them to anyone we shouldn't. Staged users
are ones that have likely only interacted with
support via email, and will likely include other
people who were CC'd on the original email to the
group.
Co-authored-by: Martin Brennan <martin@discourse.org>
This commit changes the default return value of `Auth::ManagedAuthenticator#primary_email_verified?` to false. We're changing the default to force developers to think about email verification when building a new authentication method. All existing authenticators (in core and official plugins) have been updated to explicitly define the `primary_email_verified?` method in their subclass of `Auth::ManagedAuthenticator` (example commit 65f57a4d05).
Internal topic: t/82084.
The `TopicView#bookmarks` method is called by `TopicViewSerializer` and `PostSerializer`
so we want to avoid running a meaningless query when user is not
present.
When loading posts in a topic, the topic level guardian
checks are run multiple times even though all the posts belong to the
same topic. Profiling in production revealed that this accounted for a
significant amount of request time for a user that is not staff or anon.
Therefore, we're optimizing this by adding memoizing the topic level
calls in `PostGuardian`. Speficifally, the result of
`TopicGuardian#can_see_topic?` and `PostGuardian#can_create_post?`
method calls are memoized per topic.
Locally profiling shows a significant improvement for normal users
loading a topic with 100 posts.
Benchmark script command: `ruby script/bench.rb --unicorn --skip-bundle-assets --iterations 100`
Before:
```
topic user:
50: 114
75: 117
90: 122
99: 209
topic.json user:
50: 67
75: 69
90: 72
99: 162
```
After:
```
topic user:
50: 101
75: 104
90: 107
99: 184
topic.json user:
50: 53
75: 53
90: 56
99: 138
```
This commit removes 3 redundant DB queries when loading posts.
1. `@posts` will eventually have to be loaded so we can avoid two
additional queries.
2. No need to preload topic association of posts as we're already
dealing with a fixed topic in `TopicView`.
In some situations (e.g. disaster recovery), it may make sense to spin up a temporary readonly version of a cluster. In that situation, the s3 `expire_missing_assets` job would delete assets which are still in use by the canonical read-write version of the cluster.
To avoid that, this commit will skip deletion if the site is currently in readonly mode.
1. Fix bug where we were not waiting for all unicorn workers to start up
before running benchmarks.
2. Fix a bug where headers were not used when benchmarking. Admin
benchmarks were basically running as anon user.
3. Disable rate limits when in profile env. We're pretty much going to
hit the rate limit every time as a normal user.
4. Benchmark against topic with a fixed posts count of 100. Previously profiling script was just randomly creating posts
and we would benchmark against a topic with a fixed posts count of 30.
Sometimes, the script fails because no topics with a posts count of 30
exists.
5. Benchmarks are not run against a normal user on top of anon and
admin.
6. Add script option to select tests that should be run.
# Context
When a topic is reviewable by a group we give those group moderators some admin abilities including the ability to delete a topic.
# Problem
There are two main problems:
1. Currently when a group moderator deletes a topic they are redirected to root (not the same for staff)
2. Viewing the categories deleted topics (`c/foo/1/?status=deleted`) does not display the deleted topic to the group moderator (not the same for staff).
# Fix
If the `deleted_by` user is part a group that matches the `reviewable_by_group` on a topic then don't redirect. This is the default interaction for staff to give them the ability to do things like restore the topic in case it was accidentally deleted.
To render the deleted topics as expected for the group moderator I am utilizing [the guardian scope of `guardian.can_see_deleted_topics?` for said category](https://github.com/discourse/discourse/pull/19618/files#diff-288e61b8bacdb29d9c2e05b42da6837b0036dcf1867332d977ca7c5e74a44297R802-R803)
Previously, browser logs would be printed to STDOUT halfway through the test run. This commit changes the behaviour so that the logs are included in the failure summary along with other rspec failure information.
There is an issue where chat message processing breaks due to
unhandles `SocketError` exceptions originating in the SSRF check,
specifically in `FinalDestination::Resolver`.
This change gives `FinalDestination::SSRFDetector` a new error class
to wrap the `SocketError` in, and haves the `RetrieveTitle` class
handle that error gracefully.
A few specs in `dashboard_controller_spec.rb` set some state in redis but don't clean it up afterwards which causes other specs to fail when they're ran after `dashboard_controller_spec.rb`.
Related commit: 18467d4.