The cause of this is likely a misbehaving gateway plugin. The plugin should add a report for each recipient.
Otherwise some other custom code is likely to blame.
Without more info, its not possible to guess offhand.
If you believe its a problem with the core framework, please post steps to repro / patch.
Sounds like we're on the right path.
To make it clear, if your Kernel tests are testing openid_connect, then it you will need to also manually enable externalauth (add to $modules) since it has been a hard dependency for a while. It is only by luck (the service provider) that you havn't needed it.
Also,
Which core version.
And are you running any patches with each/either of the OIDC / Keycloack modules.
I don't think there is any tone implied. It is merely communicating clearly, similar or according to RFC2119.
It would make sense for a coding standard to adopt clear language.
The project doesnt utilise `openid_connect.openid_connect` in 2.2.1?
Are you sure its this project.
Can we get a stack trace from the error?
The service shows up if I do \Drupal::service('openid_connect.openid_connect');.
Not much has happened with this project recently, however the other user is reporting quite a leap. v1.8 to 2.2.1 ??
Sounds like you're on it. Yeah lets add tags.
Is form caching on by default now, easily testable nowadays?
As found in 3.9.3 and later.
Merged. CI: PHPUnit still red so no release quite yet.
@claudiu.cristea if you want to rebuild you can try looking at the fork, and looking at the tags with `previous-` prefix. @ https://git.drupalcode.org/issue/drupal-2879087/-/network/9.2.x?ref_type... / https://git.drupalcode.org/issue/drupal-2879087/-/tags
Thanks!
Can we get a new release?
Can we change default branch in gitlab?
+1 to allowing optional comments.
+1 to allowing those comments to be formatted as a formal docblock
+ 1 to allowing optional new lines where appropriate for legibility.
-1 to requiring comments.
0 to requiring comments to be formal "docblock" style.
-1 to requiring newlines.
-1000% for requiring comments.
Noting that Symfony, has chosen this also. But it doesnt seem to be enforced by rulesets.
Expanding on new lines. New line before comment, so long as there is a comment. Otherwise enums bumping into eachother is allowed.
And new line after and before the enum brace, same as classes.
More importantly, enums should be self explanatory. Or otherwise easily explainable by an optional class-level/top-level comment.
/**
* Optional comment.
*/
enum Thing {
case Foo1;
case Foo2;
case Foo3;
/**
* Description.
*/
case Foo4;
case Foo5;
}
v0.2.2
Activity again in the original repo, @shish got access.
Ill commit this since its a change to dev, Ill switch back to coding machine later.
I figured out the problem with CI, looks like theres a difference where code lies when ran as a fork. In your MR, the composer step is failing despite it being green.
My green MR @ https://git.drupalcode.org/project/pinto/-/merge_requests/16 with the same code as yours. Ill close this one and commit yours since I know it works.
I'd accept a PR with each of the proposed bug fixes.
As found in 1.8.0
FYI interface change is in 11.1 and higher. The change is not in 11.0.
Might be easier to make this a change on only the 11.1 and higher branch.
Havn't tried it but I trust you ;)
Yep, posted there also → .
Im now a -1 on this issue, copying my comment from the core issue:
I've been making use of phpstan strict ruleset for a couple years now, and I have to mention that I agree with one of their rules, that would be at odds with this issue. That is, to never use
?:. Because it has weak comparison semantics.Justification, from docs:
Disallow short ternary operator (?:) - implies weak comparison, it's recommended to use null coalesce operator (??) or ternary operator with strict condition.
So, since this issue is about core itself, not coding standards. I'd support enabling this rule from strict, and/or in the interim finding a codesniffer rule that disallows it.
So I'd be advocating for SlevomatCodingStandard.ControlStructures.DisallowShortTernaryOperator
I've been making use of phpstan strict ruleset for a couple years now, and I have to mention that I agree with one of their rules, that would be at odds with this issue. That is, to never use ?:. Because it has weak comparison semantics.
Justification, from docs:
Disallow short ternary operator (?:) - implies weak comparison, it's recommended to use null coalesce operator (??) or ternary operator with strict condition.
So, since this issue is about core itself, not coding standards. I'd support enabling this rule from strict, and/or in the interim finding a codesniffer rule that disallows it.
The project is presently maintained by the community.
By that I mean, as sole maintainer I'm not currently being proactive about maintenance.
To be clear, I dont intend to add new maintainers. That won't help anything. I continue to be a steward of the project.
If you create a D11 compat issue, and we have community consensus that it works, I'll review and merge if ready.
New issue for any changes, please.
Notifier is ready at https://www.drupal.org/project/notifier →
The essential parts of Notifier required for this are now ready at https://www.drupal.org/project/notifier →
Development of SMS Framework continues locally for now.
Great thanks <3
dpi → created an issue. See original summary → .
@avpaderno is the project name issue "notifier-0" in IS, something you can address?
Thanks all.
Removing existing maintainers per above
dpi → created an issue. See original summary → .
Moved feedback FU to ✨ Split report to page per entity type Active
Autowired services are available by their classname as ID. So for plugins, or anything using containerinjection interfaces, or even \Drupal::service(...), you should do $container->get(DatetimeRangeViewsHooks::class),
As it is, I dont think 3.8 is compatible with 11.1 (11.0?) anymore. Due to various incompatibilities introduced. So ignoring tests on D11, and hoping for the best.
Requested in Slack @ https://drupal.slack.com/archives/C1BMUQ9U6/p1737427230495489
Thanks @avpaderno, does that mean you've contacted them directly?
Transferring as plenty of time and notice has passed.
Intent to fully take over and remove existing maintainers as there is no lineage from the old project at all.
Hoping to revisit this with D11 upgrades, but that could end up being any time this year.
Could we get a new @beta release including this one <3
Supporting @smustgrave request as he's a active, proven, competent, member of the community.
Checked the linting issues, you can ignore.
Feedback already there.
Does this work for 11.0 also? If not, might need to make the to-be-changed-to-autowired service arg optional (make nullable) and call function if service not injected. e.g $this->hookservice?->method() ?? \call_old_function();
Thanks, I can see weekly test runs are failing since core 11.1.
If you can get the tests passing again, that would satisfy this issue.
Might be a dupe of ✨ Phone number verification event dispatcher RTBC
The typical 2 week period has passed, though I'll grant extra time due to this special time period.
Thanks @gbyte, @walkingdexter, I'd accept high-level/architectural review over macro/technical review.
However, pinging with a query anytime the client collection is fetched means a lot of db roundtrips, and as such kill performance - if I am not mistaken, getClientConnection() is called anytime a transaction is opened. BTW, that would also mean that non-transactional statement executions will not be covered.
As the code is written right now, this is not correct. ->ping() is called in ->getClientConnection(), but since ->connection is always a PDO, it exits. It would only continue if the PDO object in ->connection was set to null by an explicit call to ->close().
The behavior is not wrapping every call in a safety net. If the connection was lost, and ping was not called by userland code, then indeed the app would crash. Encapulating all database calls would be quite a scope increase, and fwiw is not the behavior of Symfony+Doctrine+Messenger.
Another aspect to take care of is understanding the implications of a reconnect occurring while a db transaction is open - is the db transaction rolled back, how will the db failure reflect in the transaction manager, etc.
I'm not sure about this one.
Seems to me this is likely outdated, given version and time since last comment.
This isnt just a Gin problem, we can de-escalate, but it still seems necessary to adhere to standards.
I sent this to dbehrman today.
bdlangton has a disabled contact form.
Both users are not active in the community, so I expect this request to elapse and move into the site mainainers queue.
Hi, Barrett Langton (bdlangton) and Daniel Behrman (dbehrman),
I'm looking to take over the Notifier (notifier) namespace in order to implement Symfony Notifier functionality for Drupal.
https://www.drupal.org/project/notifier/issues/3494214 🌱 Taking over the Notifier namespace (attn: bdlangton, dbehrman) Active
I understand you may not be interested in the Drupal community any longer, for which you may choose to reply or disregard
Brought up at #3301682-21: Define bundle classes via attributes →
The discovery should probably make sure that it's not possible to register multiple classes for the same bundle. Can also happen with the alter hook, where it's much harder to detect, but at least here we could?
What about subclasses? Lets say a contrib module defines a node type and adds a bundle class for it. In a custom project you want to subclass and change that. I guess you still have the option to use the alter hook then and set it there.
Maybe you'd be interested in reviewing ✨ Add logic to determine which bundle class should take effect when multiple classes extend the same bundle Active and associated PR.
dpi → created an issue. See original summary → .
Inspired by ✨ Use attributes for DateRecurInterpreter plugin discovery Active
'source time zone' should already be UTC, how is this any different.
Your patch hard codes fields names, doesnt make much sense to me.
@socialnicheguru current plan is for sec-only. No other support. #3325334-6: Stan linting: return types, parameters, modern services, etc. →