This was reported in another issue and I made a MR for it: https://www.drupal.org/project/samlauth/issues/3486925 🐛 Cacheability Metadata Leakage Error on SAML Login with Samlauth and r4032 Redirect Module Active
I read through all the other issues like 3232577 → and did a traceback and found the same issue as reported, it's the destination param and the URL generation that's leaking the data. Using the example in the Lullabot article I updated the call to the Url class to include the TRUE flag for collecting the metadata and returning the generated url. Available for testing in the MR. Here's a link to the patch as well: https://git.drupalcode.org/project/samlauth/-/merge_requests/26/diffs.patch
nsciacca → made their first commit to this issue’s fork.
@justcadwell thanks for your suggestion. I forked the issue, implemented your suggestion, it's working nicely for me. The only annoying part is that when you submit the page it goes back to unselected state. Meaning if you were incrementally doing changes on a role or comparing two roles with specific modules selected you have to redo those selections after save. Any thoughts?
I'm having the same issue. Installed this module to prevent the Permissions page from breaking but selecting All Roles and All Permissions will break the page and then it's stuck like that. I'd rather not have it "remember" the previous settings so revisiting /admin/people/permissions would take it back to the unselected state. In my case I don't think it's the max input vars, but rather a PHP timeout on the amount of time the page can take. And since we're on a particular platform that doesn't allow us changing this value we're stuck. Thoughts on making the "remember" state optional?
+1 Tested the MR and can confirm it is working as expected.
In addition to the issues above we noticed the that global roles that were assigned to edit one content type could no longer edit nodes that were in groups. For example:
1. Global (Outsider) role of "Person Administrator" (PA) assigned to edit all nodes of content type of "Person"
2. After upgrade to 3.3 the PA could only edit Person nodes that were NOT in groups
3. Disabling the group_support_revisions module allowed them to edit all nodes of Person again (expected/previous behavior)
FYI - in case anyone lands here - I re-opened a new issue and submitted a MR here: https://www.drupal.org/project/quiz/issues/3485142 🐛 DivisionByZero error Active
nsciacca → created an issue.
Created a MR for this. The score function within MatchingResponse needed a setEvaluated flag and there was a small change to the penalty choice logic since unchosen optino had a value of "" instead of "-1"
nsciacca → made their first commit to this issue’s fork.
I moved nguyenphan's patch into a MR based on the latest branch. This fixes the issue for me and the answers are saved. I'm still seeing an issue with the total score not being calculated but I think there's a separate issue for that I will review.
nsciacca → made their first commit to this issue’s fork.
Created a MR for the 4.x branch and removing patches from this issue.
nsciacca → created an issue.
Patch in #23 works for me, 10.2.2
I've been communicating with Pantheon's support staff and they've said:
Good day, for an update, Pantheon Platform Engineers already released an update to the Search API Pantheon module https://www.drupal.org/project/search_api_pantheon → just today. The latest version 8.1.7 contains fixes for the Sticky Solr Schema.
I'll be installing this and monitoring on site asap.
Update: one of our sites randomly reverted back to 4.2.10 on 3/1 even though there was no deploy. This is a huge problem as it prevents new items from being indexed and there's no notice or warning other than our customers reporting missing content.
nsciacca → created an issue.
FYI - I did something similar in another ticket (3337669) for files that worked nicely using their pantheon_clear_edge_paths function and passing it the file url when the file was updated. In conjunction with the Media Entity File Replace module it seems to work well.
https://www.drupal.org/project/pantheon_advanced_page_cache/issues/3337669 🐛 Not showing newly updated/changed file on Pantheon - Cache issue Needs review
https://git.drupalcode.org/project/pantheon_advanced_page_cache/-/merge_...
+1 for this - works great!
I opened another MR for the 10.2.x branch as neither existing MR/patch was applying cleanly. Includes my modification with the extra selectors that were breaking Entity Browser.
@robbdavis - If you've previously patched core for the same issue I have noted above, try the patch from the MR over on https://www.drupal.org/project/drupal/issues/2858890#comment-15337125 🐛 Drupal.views.ajaxView is not initializing pagers in nested views Needs work
I ended up making some adjustments there and it helped.
https://git.drupalcode.org/project/drupal/-/merge_requests/4242/diffs.patch
@joseph.olstad Thanks for the reply but what's interesting is that I don't have Inline Form Errors enabled. There are no javascript errors thrown and nothing on the PHP side in the watchdog logs. Here's a video of me trying to select something:
https://www.dropbox.com/scl/fi/zr0lhshhkzl4cunjuf2lm/Entity-Browser-AJAX...
I guess there could be something else in our custom code that could be conflicting so I will try to spin up a vanilla Drupal with similar config to see if I can replicate it there.
nsciacca → created an issue.
Fixed patch.
Patch attached.
Thanks for this! I added a couple other parameters for YouTube when it's used as background image to suppress the controls, set the loop and a playlist param so it continues to play.
nsciacca → made their first commit to this issue’s fork.
Here's an updated patch for Drupal 10:
- Updated to simplify theme functions into .module
- Replaced core/jquery.once with core/once
- Replace file_munge_filename
- Replace usage of $_POST with request_stack service
- Drupal 10 code standards
Also there seems to be activity/work on the other issue here: Automated Drupal 10 compatibility fixes
-
https://www.drupal.org/project/dropzonejs_webform/issues/3296994 →
The patch in #91 no longer applies to 3.x-dev and it's not an easy re-roll... the split of modes with Basic, Intermediate, and Advanced complicates this.
nsciacca → created an issue.
This should be fixed in the simplesaml 2.1 release: https://github.com/simplesamlphp/simplesamlphp/commit/8de020f7b2815b5d67...
+1 for the patch. I didn't have any issue with the browser back as reported by arturs.v
nsciacca → created an issue. See original summary → .
Thanks @Wilfred Waltman - patch helped me when the View failed to display to logged in users.
New patch attached - the first one didn't account for multiple misspelled terms.
nsciacca → created an issue.
Thanks for the re-roll, working well for Drupal 9.5.10 + Facets 2.0.6 + Views Ajax History 1.7.0
I got a similar error with Views Bulk Operations:
Error: Call to undefined method Drupal\views_bulk_operations\Form\ConfigureAction::getEntity() in Drupal\rabbit_hole\Plugin\Field\FieldWidget\RabbitHoleDefaultWidget->formElement() (line 168 of modules/contrib/rabbit_hole/src/Plugin/Field/FieldWidget/RabbitHoleDefaultWidget.php).
Adding the below check to the top of the formElement function works for my instance.
if (!method_exists($form_state->getFormObject(), 'getEntity')) {
return;
}I can confirm this bug - I was just about to log it as well. It happens in both D9 & D10 with CKEditor5. I'm not sure where the fix needs to be... but it's the source that doesn't get updated. If you update the image and switch to view the source you'll see the old image url in the "src" attribute, so when you switch back it shows the original image.
+1 this patch worked for me
Updated patch for D10 as #66 had a duped line.
@GuStAvE33 Here's a patch for D10 using core/once instead of jquery once which is likely what was failing.
I was able to get this to work as desired with the patch but I had to enable the facets_summary block, which I wasn't using previously. The JS library was added to that submodule, but I don't think it needs to be there in the long run. I was able to enable it and just put a button with class "facets-apply-button" on the page with my existing checkbox facets that were updated to not auto-submit and it worked.
+1 for this patch. Just ran across this in the newer version of Fullcalendar View when upgrading to Drupal 10 and this saved the day. Thanks!
Patch in #32 worked for me. I had Layout Builder blocks referencing paragraphs and found that cloned nodes that were unpublished were setting the source node's paragraphs to be unpublished - so anonymous visitors would see a blank page visiting the source node. This is working, thank you.
Is it possible to get these changes into a dev branch? Currently you cannot install 2.0.0 and then patch it as composer will throw requirement conflicts.
@mherchel - I just applied this to my 10.1 and other than a warning "] rmdir(/files/css): Directory not empty FileSystem.php:267" it did appear to fix the aggregate CSS/JS on my Pantheon deploy.
I think this should be re-opened, I tried with the latest version and found the following issues:
1. The actions don't exist for the Bulk Form for media. I manually copied the "system.action.publish_latest" and "system.action.unpublish_latest" to media specific ones to get this to work from within my config sync. However, there should be examples in the module's config/optional folder.
2. Both action plugins for "PublishLatestRevisionAction" and "UnpublishCurrentRevisionAction" had checks in the access function for whether the entity was a node, therefore it would return an Access Denied when applying the action. I updated this to check for either nodes or media.
3. There should be similar settings for the module for "Enable dialog on media edit form" and "Enable dialog on admin media listing". This affects the dialog pop ups and the code is all hard coded to look for node forms.
I have attached a patch for 1 & 2 which at least makes the module functional for media entities. #3 is a larger lift I don't have time for at the moment.
The patch fixes the error from breaking the page, however the functionality is broken. Not sure if they're related but this issue describes the context that's not working: https://www.drupal.org/project/term_condition/issues/3363916 🐛 2.0.2 breaks backward compatibility Fixed
Re-rolled #23 for 9.5.9
nsciacca → made their first commit to this issue’s fork.
This didn't work for me, the tokens can't be enclosed in quotes.. not sure if it's a newer database incompatibility. Anyhow, this is a super old issue but in case anyone else runs across it, here's an updated patch
nsciacca → created an issue.
I couldn't apply the patch to 2.1 when using composer to manage patches, so I re-created it. Attaching in case anyone else needs.
I can no longer reproduce the error. I updated back to 2.1.1 and it's connecting fine, even without the patch added. Strange, but I think this is good to go.
nsciacca → made their first commit to this issue’s fork.
nsciacca → created an issue.
+1 for the patch in #16, thanks!
@smustgrave I think my issue is actually caused by the change in https://www.drupal.org/project/drupal/issues/3294720 🐛 The attachBehaviors() for document is only called after Big Pipe chunks are processed Fixed . As that change was the one that introduced another attachBehaviors call whereas this patch just modified an existing call. Okay to move forward, and thanks for checking in.
Retracting my comment above. Using version 3.1.3 and this no longer is an issue.