Fork of #2595695: 4xx handling using subrequests: no longer able to vary by URL
The conversion of _form being in a request listener means it doesn't happen as part of routing, per se. That causes an issue for subrequests and cache contexts, as seen in #2595695: 4xx handling using subrequests: no longer able to vary by URL. The various entity magic keys are handled with a RouteEnhancer, which is part of routing.
So let's make _form do that, too. There should be no API change here, other than implicitly fixing a bug in #2595695: 4xx handling using subrequests: no longer able to vary by URL in a nicer way. (It's probably also immeasurably faster.)
Comment | File | Size | Author |
---|---|---|---|
#5 | 2595695-move-form-5.patch | 3.61 KB | Wim Leers |
Comments
Comment #2
dawehnerClosing all the
nitpick: @file, but note: #2304909: Relax requirement for @file when using OO Class or Interface per file is RTBC FUCK YEAH!!!!!!!!!!!!!!!!!!!!!!!!!
Don't we use
{@inheritdoc}
Did you typed that out, or is this coming from somewhere automatically?
Meh, out of scope
Comment #3
Wim LeersFirst rerolling without the
default.settings.php
changes.Comment #4
Wim LeersUgh, forgot to include the new file.
Comment #5
Wim Leers#2 all fixed, and fixed additional nitpicks too.
Also: #2304909: Relax requirement for @file when using OO Class or Interface per file <3 <3 <3 YES!!
Comment #7
Crell CreditAttribution: Crell at Palantir.net commentedEh, sorry for the jitter. It's been a while since I rolled a Drupal patch rather than a PR on something. :-P But yes, let's do!
Comment #8
dawehnerYeah, one less todo :)
Comment #9
tim.plunkettComment #10
catchThis blocks #2595695: 4xx handling using subrequests: no longer able to vary by URL which is an RC target.
It removes a service which theoretically someone could rely on, but as of this morning, event subscribers are treated as @internal per https://www.drupal.org/node/2562903/.
It's a real shame we didn't do #2350491: Discuss/implement our usage of private services then it would not even be a question of whether someone's using it or not except for subclassing - I've just re-opened that issue since I can't find the issue it was marked duplicate of.
Comment #11
catchCommitted/pushed to 8.0.x, thanks!
Comment #13
Crell CreditAttribution: Crell at Palantir.net commented#11 says this was committed to 8.0.x, meaning it should be in today's 8.0.0 tag. The automated message in #12, though, says it's for 8.1.x. Where'd it go? :-)
Comment #14
tim.plunketthttp://cgit.drupalcode.org/drupal/commit/?id=7e3042e
This was committed while the git server was being worked on.