The `posts` relation on `Topic` is not ordered. Using `Topic.posts.first`
is basically the same as asking for a random post, it will depend on DB
order. This breaks on Topic merge and split for example.
Additionally, a huge problem with that is that it forces active record down
a slow path. `Topic.posts.first` is extremely slow on giant topics, since
it has no default ordering it appears AR materializes the entire set prior
to doing `first`.
This commit also illustrates the importance of testing, initially I only
fixed the second instance of the problem in `post_validator.rb` but testing
revealed that the problem was repeated at the top of the file.
Longer term we should consider a larger change of default ordering the posts
relations so people do not fall down this trap anymore.
- Themes can supply translation files in a format like `/locales/{locale}.yml`. These files should be valid YAML, with a single top level key equal to the locale being defined. For now these can only be defined using the `discourse_theme` CLI, importing a `.tar.gz`, or from a GIT repository.
- Fallback is handled on a global level (if the locale is not defined in the theme), as well as on individual keys (if some keys are missing from the selected interface language).
- Administrators can override individual keys on a per-theme basis in the /admin/customize/themes user interface.
- Theme developers should access defined translations using the new theme prefix variables:
JavaScript: `I18n.t(themePrefix("my_translation_key"))`
Handlebars: `{{theme-i18n "my_translation_key"}}` or `{{i18n (theme-prefix "my_translation_key")}}`
- To design for backwards compatibility, theme developers can check for the presence of the `themePrefix` variable in JavaScript
- As part of this, the old `{{themeSetting.setting_name}}` syntax is deprecated in favour of `{{theme-setting "setting_name"}}`
This commit introduces an ultra low priority queue for post rebakes. This
way rebakes can never interfere with regular sidekiq processing for cases
where we perform a large scale rebake.
Additionally it allows Post.rebake_old to be run with rate_limiter: false
to avoid triggering the limiter when rebaking. This is handy for cases
where you want to just force the full rebake and not wait for it to trickle
* FIX: rake emails:test to bypass the "STARTTLS required" message
* FIX: too much j
* hint meta.discourse.org to people with unknown errors from the mail test
SiteSettingExtension triggers message bus which re-establishes a
DB connection in `SiteSettingExtension#process_message`. That happens
concurrently and a test that requires a connection to the db will
fail when the reconnection is happening.
Moves Highlight.js files to vendor/assets/javascripts
Adds Highlight.js in yarn package management
Removes old rake task and reliance on NPM to build Highlight.js
Highlight.js is now integrated in the "javascript:update" rake task
This makes more sense than having the guardian take an accessor.
The logic belongs in the Serializer, where the JSON is calculated.
Also removed some of the DRYness in the spec. It's fewer lines
and made it easier to test the option on the serializer.
Previously, we would initialize an ImageOptim object each time we resize.
This object init is mega expensive (170ms on a VERY fast machine):
```
[1] pry(main)> Benchmark.measure { FileHelper.image_optim }
=> #<Benchmark::Tms:0x00007f55440c1de0
@cstime=0.055742,
@cutime=0.141031,
@label="",
@real=0.17165619300794788,
@stime=0.0002750000000000252,
@total=0.19890400000000008,
@utime=0.0018560000000000798>
```
This happens cause during init it hunts for all the right binaries and sets
up internals.
We now memoize this object to avoid a huge amount of pointless work.
This ensures that unicorn master forks of sidekiq run with a lower priority
than the webs. It means that a busy sidekiq is less likely to impact web
performance
Before this patch, a high trust level user could flag something
and have an action be taken, as well as skipping the flag queue.
Now, if a TL3/TL4 cause an action, the flag will skip the minimum
visibility check and allow staff to review it.
FIX: buildTranslationTree was erroring when translations overlapped (ie. ":-)" and ":-))")
FIX: emoji translations wasn't working properly when translations overlapped
We need to handle arbitrary exceptions in this task, especially since the
task is not easily resumable.
Simply output problem uploads as you hit them for now.
Some previous migrations to S3 may have bad ACLs set on objects. This
introduces a new rake task (`rake s3:correct_acl`) that will reset ACL on
every S3 object.
Vast majority of users will never have to run it, but if you have ACL issues
this is the atomic solution.
This feature ensures optimized images run via pngquant, this results extreme amounts of savings for resized images. Effectively the only impact is that the color palette on small resized images is reduced to 256.
To ensure safety we only apply this optimisation to images smaller than 500k.
This commit also makes a bunch of image specs less fragile.
This avoids require dependency on method_profiler and anon cache.
It means that if there is any change to these files the reloader will not pick it up.
Previously the reloader was picking up the anon cache twice causing it to double load on boot.
This caused warnings.
Long term my plan is to give up on require dependency and instead use:
https://github.com/Shopify/autoload_reloader
As per the documentation for KEYS
```
Warning: consider KEYS as a command that should only be used in production environments with extreme care. It may ruin performance when it is executed against large databases. This command is intended for debugging and special operations, such as changing your keyspace layout.
```
Instead SCAN
```
Since these commands allow for incremental iteration, returning only a small number of elements per call, they can be used in production without the downside of commands like KEYS or SMEMBERS that may block the server for a long time (even several seconds) when called against big collections of keys or elements.
```
This generates a 10x10 PNG thumbnail for each lightboxed image.
If Image Lazy Loading is enabled (IntersectionObserver API) then
we'll load the low res version when offscreen. As the image scrolls
in we'll swap it for the high res version.
We use a WeakMap to track the old image attributes. It's much less
memory than storing them as `data-*` attributes and swapping them
back and forth all the time.
This validation makes sure that the s3_upload_bucket and the
s3_backup_bucket have different values. The backup bucket is
allowed to be a subfolder of the upload bucket. The other way
around is forbidden because the backup system searches by
prefix and would return all files stored within the backup
bucket and its subfolders.
* Dashboard doesn't timeout anymore when Amazon S3 is used for backups
* Storage stats are now a proper report with the same caching rules
* Changing the backup_location, s3_backup_bucket or creating and deleting backups removes the report from the cache
* It shows the number of backups and the backup location
* It shows the used space for the correct backup location instead of always showing used space on local storage
* It shows the date of the last backup as relative date
`SiteSerializer#is_readonly` is cached for an anonymous user so we have
to clear the cache when disabling readonly mode. Otherwise, the site may
appear to be in readonly mode for an extended period of time.
A per process cache is hard to reason about. During PostgreSQL
failovers. The site may bounce in and out of readonly mode depending on
which server and process that a request hits.
Historically due to https://meta.discourse.org/t/why-is-discourse-so-slow-on-android/8823
we decreased page sizes of both home page and topic page on android by half.
This was done on the server side and as a side effect and caused page sizes on android
to mismatch between Android and non Android.
Unfortunately about a year ago googlebot started pretending it is Android,
this cause Google to start indexing pages as what android would see. So
it saw double the amount of pages in the index as what exists on desktop.
This in turn caused double the amount of indexing work and a large amount
of broken links on long topics.
This fix removes all special behavior which is no longer needed due to
other performance work in Discourse including raw handlebars on home page
and virtual dom on topic pages.
I tested we do not need this on Blu Advance 5.0 it has 1.3 GHZ mediatec mt6580
This phone retails for around $50 USD.
If we decide long term that we want any hacks like this we will shift them
to the client side. It can just hold data in memory without rendering.
This feature is used for defer loading of images and in future for post cloaking
This gives us a polyfill so we can safely use the feature in problem browsers
The polyfill supports "polling" but it does not appear we need it yet.
If we discover anything odd here, consider setting poll interval per:
https://github.com/w3c/IntersectionObserver/tree/master/polyfill
```
var io = new IntersectionObserver(callback);
io.POLL_INTERVAL = 100; // Time in milliseconds.
```
Keeping the mutation observer cause we often mutate the DOM
Previously the 'reconnect' process was a bit magic - IF you were already logged into discourse, and followed the auth flow, your account would be reconnected and you would be 'logged in again'.
Now, we explicitly check for a reconnect=true parameter when the flow is started, store it in the session, and then only follow the reconnect logic if that variable is present. Setting this parameter also skips the 'logged in again' step, which means reconnect now works with 2fa enabled.
Some URLs in browsers are non compliant and contain twos `#` this commit adds
special handling for this edge case by auto encoding any fragments containing `#`
The wizard searches for:
* a topic that with the "is_welcome_topic" custom field
* a topic with the correct slug for the current default locale
* a topic with the correct slug for the English locale
* the oldest globally pinned topic
It gives up if it didn't find any of the above.
Also added some helpful functionality for plugin developers:
- Raises RuntimeException if the auth provider has been registered too late
- Logs use of deprecated parameters
* FEATURE: allow plugins and themes to extend the default CSP
For plugins:
```
extend_content_security_policy(
script_src: ['https://domain.com/script.js', 'https://your-cdn.com/'],
style_src: ['https://domain.com/style.css']
)
```
For themes and components:
```
extend_content_security_policy:
type: list
default: "script_src:https://domain.com/|style_src:https://domain.com"
```
* clear CSP base url before each test
we have a test that stubs `Rails.env.development?` to true
* Only allow extending directives that core includes, for now
Changes to functionality
- Removed syncing of user metadata including gender, location etc.
These are no longer available to standard Facebook applications.
- Removed the remote 'revoke' functionality. No other providers have
it, and it does not appear to be standard practice in other apps.
- The 'facebook_no_email' event is no longer logged. The system can
cope fine with a missing email address.
Data is migrated to the new user_associated_accounts table.
facebook_user_infos can be dropped once we are confident the data has
been migrated successfully.
A generic implementation of Auth::Authenticator which stores data in the
new UserAssociatedAccount model. This should help significantly reduce the duplicated
logic across different auth providers.
Previously, the site setting was only effective on the client side of
things. Once the site setting was been reached, all oneboxes are not
rendered. This commit changes it such that the site setting is respected
both on the client and server side. The first N oneboxes are rendered and
once the limit has been reached, subsequent oneboxes will not be
rendered.
* Add missing icons to set
* Revert FA5 revert
This reverts commit 42572ff
* use new SVG syntax in locales
* Noscript page changes (remove login button, center "powered by" footer text)
* Cast wider net for SVG icons in settings
- include any _icon setting for SVG registry (offers better support for plugin settings)
- let themes store multiple pipe-delimited icons in a setting
- also replaces broken onebox image icon with SVG reference in cooked post processor
* interpolate icons in locales
* Fix composer whisper icon alignment
* Add support for stacked icons
* SECURITY: enforce hostname to match discourse hostname
This ensures that the hostname rails uses for various helpers always matches
the Discourse hostname
* load SVG sprite with pre-initializers
* FIX: enable caching on SVG sprites
* PERF: use JSONP for SVG sprites so they are served from CDN
This avoids needing to deal with CORS for loading of the SVG
Note, added the svg- prefix to the filename so we can quickly tell in
dev tools what the file is
* Add missing SVG sprite JSONP script to CSP
* Upgrade to FA 5.5.0
* Add support for all FA4.7 icons
- adds complete frontend and backend for renamed FA4.7 icons
- improves performance of SvgSprite.bundle and SvgSprite.all_icons
* Fix group avatar flair preview
- adds an endpoint at /svg-sprites/search/:keyword
- adds frontend ajax call that pulls icon in avatar flair preview even when it is not in subset
* Remove FA 4.7 font files
We were looking up each mention one by one without any form of caching and that results
in a problem somewhat similar to an N+1. When we have to do alot of DB
lookups, it also increased the time spent in the V8 context which may
eventually lead to a timeout. The change here makes it such that mention lookups only does a single
DB query per post that happens outside of the V8 context.
Previously in some cases we would queue logging of invalid post numbers
The impact would be we would miss logging an incoming link and would leak
an error.
This also adjusts the algorithm to expect
- 30% saving for JPEG conversion
AND
- Minimum of 75K bytes saved
The reasoning for increase of saving requirements is cause PNG may have been
uploaded unoptimized, 30% saving on PNG is very possible
This removes a monkey patch we no longer need since our containers require
2.5.2 or up for all Discourse installs.
If you are looking to deploy on 2.5.1 which is highly not recommended you
will need to figure out how to apply this diff.
* Prioritizes non-image uploads
* Does one remap per upload instead of 3 remaps previously
* Every 100 uploads migrated, do 2 remaps which fixes broken
URLs
* Exclude email_logs table from remap
* Cuts number of queries from 273 to 89
* Add some specs
* For a table with 500 posts, benchmarks locally shows a runtime
reduction from 0.046929135 to 0.032694705.
* First take on subsetting svg icons
* FontAwesome 5 svg subset WIP
* Include icons from plugins/badges into svg sprite subset
* add svg icon support to themes
* Add spec for SvgSprite
* Misc. SVG icon fixes
* Use FA5 svgs in local-dates plugin
* CSS adjustments, fix SVG icons in group flair
* Use SVG icons in poll plugin
* Add SVG icons to /wizard
We regressed and optimized images no longer worked with svg
The following adds the correct logic to simply copy file for svgs
and bypasses resizing for svg avatars
On sites with many flages, it could take quite a long time for
ActiveRecord to return all the joined data.
It's now 3 queries instead of one, but significantly faster, especially
if you have a minimum threshold set.
Previously the related PMs were last meaning you would have to work through
all unread to see them.
Also amends it so it either asks for related by group OR user not both.
4481836 introduced accent stipping in search_indexer,
but we need to strip it from the query itself as well
TODO in search with diacritics:
- Still need to fix excerpts on search page
- need to support accent stripping in in_topic search
- need to make sure that in:title works correctly
- need to fix "word boldening" in titles
* Enable user email PM when posting to group or replying to topic via email
* remove extra line
* Add test and fix snake_case
* Only reenable email_private_messages for PM replies
This splits off the logic between SSO keys used incoming vs outgoing, it allows to far better restrict who is allowed to log in using a site.
This allows for better auditing of the SSO provider feature
