This was originally introduced in d87583d0ef, but has not seen usage, since usually when the session needs to be modified, the request is available.
It causes issues with certain queue drivers, as it can't be serialized.
It's also not entirely accurate, as a user can have multiple sessions at once. Therefore, a given session is a property of the request, not of the user.
The reason this causes issues in the Queue is that when a Job has payload that consists User(s), the Queue will try to serialize that. Serializing the User object will require serializing the session too; this causes a Serialization of Closure is not allowed error, see image.
One can circumvent that in many ways, the most obvious one is adding a __sleep and __wakeup implementation in the User class (or the session handler). But as we aren't really using the session on the User model anywhere in core, bundled or most community extensions it is best to simply detach this from the user.
* Add an ActorReference class to store the actor `$request->getAttribute('actorReference')->getActor()`
* Add a middleware to inject the actor reference
* Deprecate `$request->getAttribute('actor')`
* Rename `app` helper to `resolve`, deprecate old version
* Rename $this->app to $this->container in service providers
We no longer couple Flarum\Foundation\Application to the Laravel container; instead, we use the container separately. Changing our naming to reflect that will make things clearer.
- Make session token-based instead of user-based
- Clear current session access tokens on logout
- Introduce increment ID so we can show tokens to moderators in the future without exposing secrets
- Switch to type classes to manage the different token types. New implementation fixes#2075
- Drop ability to customize lifetime per-token
- Add developer access keys that don't expire. These must be created from the database for now
- Add title in preparation for the developer token UI
- Add IP and user agent logging
- Delete all non-remember tokens in migration
- Support slug drivers for core's sluggable models, easily extends to other models
- Add automated testing for affected single-model API routes
- Fix nickname selection UI
- Serialize slugs as `slug` attribute
- Make min search length a constant
- Split DispatchRoute. This allows us to run middleware after we figure out which route we're on, but before we actually execute the controller for that route.
- By making the route name explicitly available to middlewares, applications like CSRF and floodgate can set patterns based on route names instead of the path, which is an implementation detail.
- Support using route name match for CSRF extender, deprecate path match
* Fixes#2157, Explicitly set SameSite value for cookies by making samesite a config option in config.php. Also contains an update for the cookie library dependency
In Laravel 5.8, the `Container::tagged()` method was changed to return
an iterator [1].
We only use the result for iteration, or, in this case, to pass a bunch
of "reporters" to the error handler middleware, therefore we need to
accept an iterable here.
[1]: https://laravel.com/docs/5.8/upgrade#container-generators
- Stop trying to implement Laravel's Application contract, which
has no value for us.
- Stop inheriting from the Container, injecting one works equally
well and does not clutter up the interfaces.
- Inject the Paths collection instead of unwrapping it again, for
better encapsulation.
This brings us one step closer toward upgrading our Laravel
components (#2055), because we no longer need to adopt the changes
to the Application contract.
This extends our boot exception handling block to also catch and format
all exceptions that could be thrown while building our request handler,
i.e. the middleware stack handling requests.
The only exceptions that would now not be handled in this way could be
raised by Zend's `RequestHandlerRunner` and its delegates, which we
should be able to rely on.
Exceptions on request execution will be handled by the error handler in
the middleware stack.
Fixes#1607.
...not based on status code.
To simplify this logic, we now use the same error "type" both when
routes are not found and specific models are not found. One exception is
ours, one is from Laravel, but for the purposes of error handling they
should be treated the same.
Fixesflarum/core#1641.
The error handling middleware now expects an array of reporters.
Extensions can register new reporters in the container like this:
use Flarum\Foundation\ErrorHandling\Reporter;
$container->tag(NewReporter::class, Reporter::class);
Note that this is just an implementation detail and will be hidden
behind an extender.
This separates the error registry (mapping exception types to status
codes) from actual handling (the middleware) as well as error formatting
(Whoops, pretty error pages or JSON-API?) and reporting (log? Sentry?).
The components can be reused in different places (e.g. the API client
and the error handler middleware both need the registry to understand
all the exceptions Flarum knows how to handle), while still allowing to
change only the parts that need to change (the API stack always uses the
JSON-API formatter, and the forum stack switches between Whoops and
pretty error pages based on debug mode).
Finally, this paves the way for some planned features and extensibility:
- A console error handler can build on top of the registry.
- Extensions can register new exceptions and how to handle them.
- Extensions can change how we report exceptions (e.g. Sentry).
- We can build more pretty error pages, even different ones for
exceptions having the same status code.
* Integration tests: Memoize request handler as well
This is useful to send HTTP requests (or their PSR-7 equivalents)
through the entire application's middleware stack (instead of
talking to specific controllers, which should be considered
implementation detail).
* Add tests for CSRF token check
* Integration tests: Configure vendor path
Now that this is possible, make the easy change...
* Implement middleware for CSRF token verification
This fixes a rather large oversight in Flarum's codebase, which was that
we had no explicit CSRF protection using the traditional token approach.
The JS frontend was actually sending these tokens, but the backend did
not require them.
* Accept CSRF token in request body as well
* Refactor tests to shorten HTTP requests
Multiple tests now provide JSON request bodies, and others copy cookies
from previous responses, so let's provide convenient helpers for these.
* Fixed issue with tmp/storage/views not existing, this caused tmpname to notice.
Fixed csrf test that assumed an access token allows application access, which is actually api token.
Improved return type hinting in the StartSession middleware
* Using a different setting key now, so that it won't break tests whenever you re-run them once smtp is set.
Fixed, badly, the test to create users etc caused by the prepareDatabase flushing all settings by default.
* added custom view, now needs translation
- Simpler class naming:
Frontend\CompilerFactory → Frontend\Assets
Frontend\HtmlDocumentFactory → Frontend\Frontend
Frontend\HtmlDocument → Frontend\Document
- Remove AssetInterface and simply collect callbacks in Frontend\Assets
instead
- Remove ContentInterface because it serves no purpose (never type-
hinted or type-checked)
- Commit and add asset URLs to the Document via a content callback
instead of in the Document factory class itself
- Add translations and locale assets to Assets separate to the assets
factory, as non-forum/admin asset bundles probably won't want them
- Update Frontend Extender to allow the creation of new asset bundles
- Make custom LESS validation listener a standalone class instead of
extending RecompileFrontendAssets
