@cilefen

What is in the array passed to strlen in Unicode.php on line 395?

Possibly. I was asking what the isolation level is on the affected site. Drupal shows this configuration on the site status report.

I am not a Symfony core developer so I can't answer that. The commit does not specify, but it's an API change.

Whatever this exception is, had been triggering a deprecation in prior versions, like

trigger_deprecation('symfony/http-foundation', '5.1', 'Retrieving a non-string value from "%s()" is deprecated, and will throw a "%s" exception in Sy
mfony 6.0, use "%s::all($key)" instead.', __METHOD__, BadRequestException::class, __CLASS__);

.

Did you observe those deprecation messages? What is happening was announced if I understand this correctly.

What does is the data type of $value on line 35 of vendor/symfony/http-foundation/InputBag.php when this occurs?

What is the MySQL transaction isolation level → ?

If anyone that can reproduce this would perform a Git bisect operation on a working copy of Drupal Core you would quickly find the commit that changed this behavior, which would go a long way toward solving this.

Is Html::load intended to work on header markup?

Four alternatives are listed in the Webform Libraries → doc page.

This is a "closed (won't fix)" for me.

Are those sites on the same Drupal core versions?

It is a fact that text formats cannot be deleted in the UI. It may be possible with configuration import, but I don't know. 🐛 Disabled text formats can't be seen in the GUI Needs work has made them visible to be re-enabled if needed. Is anything else needed to close this issue?

If this is critical, are you saying that #2526150: Database cache bins allow unlimited growth: cache DB tables of gigabytes! → does not work, or that it is ineffective?

I completely understand that even with row-limiting, an attacker can essentially replace useful cached objects with useless ones, and that Drupal could do better here. But can those outcomes render a site unusable or cause database deadlocks → ? Possibly.

• https://drupal.stackexchange.com/questions/313273/how-to-display-a-block...
• https://www.drupal.org/forum/support/post-installation/2019-06-05/how-to... →

This could just be 🐛 Random sort not working for anonymous users if "Internal Page Cache" module enabled. Fixed .

What is the specific relation of this issue to 🐛 Cannot remove media on field, when there is a custom validation error Active ?

Drupal 10.1 is no longer being released. The maintainers need proposed changes to the 11.x branch. What is the specific relation of this issue to 🐛 Cannot remove media on field, when there is a custom validation error Active ?

I can't reproduce this bug. I tested with Drupal 10.2.7, 10.4.0, 10.4.1, 10.4.2, 10.4.3, and 10.4.4.

We need additional steps to reproduce the bug. Are other entity endpoints working on the affected site? Can you get a stack trace?

In addition, merging upstream into a site project is not a recommended workflow for Drupal in 2025.

The CMS edition has no publicized version number:

What is the specific Drupal core version that introduces this bug? I can't determine that from the way this issue is worded, but the wording indicates you may have this information.

There are consequences of not having concatenation enabled, like additional data transfer for your hosting provider and visitors.

Some of the system module form fields have additional documentation if you enable the help module, but not these. It’s an interesting idea to improve that documentation or to enhance the inline field help text in the way you are suggesting. Consider opening an issue for that.

On your behalf I searched this issue queue for related items and found 📌 Add save dropbutton next to 'Back to content editing' link in node preview. Active but not much else about the oddity of in fact being able to save without previewing, which is what you rightly noticed. The behavior this issue cites dates back to decisions being made thirteen years ago in #1510544: Allow to preview content in an actual live environment → , which had its origin in controlled usability tests.

It could be that this was an oversight of the original design, this is a bug introduced some time in the last twelve or so years of development, or that people have moved editorial processes to workflows → , which is part of the core system.

Drupal's Slack workspace → is the best place for interactive discussions about product decisions and development direction.

What is the actual version of the website in the screenshot? This could be 🐛 Disabled text formats can't be seen in the GUI Needs work , which is more of a feature than a bug, and if so, unrelated to CKEditor module.

At a glance this seems to be how CKEditor does things, not Drupal. That is: if you pick "Solid" for the border style and "1px" for "Width, CKEditor does not create an attribute mentioning width. CKEditor will do so if you select some other options, like 2px and other styles. But it does so differently. Sometimes CKEditor adds a style for "border", and in other cases it uses the "border-width" style.

It would be good to know:

1. What difficulties or problems does this behavior cause?
2. Is the behavior the same on the CKEditor demos? If that's the case this isn't to do with Drupal itself.

I am updating some issue metadata according to the project guidelines.

This could be 🐛 Aggregation not working css and js cache directories not being created Active .

Based on 🐛 Fatal error "getCacheTags() on null" on admin/content Views page Closed: works as designed being opened I am reopening this. It still needs steps to reproduce. 🐛 Fatal error "getCacheTags() on null" on admin/content Views page Closed: works as designed doesn't have any.

I will go and reopen the other issue. Please add your additional information there.

Welcome to drupal.org, @yasarteceze:

The area where you opened this issue is for Drupal Core, the main software, but your question is about a "contributed" module, so the issue must be moved to that specific module's queue. I think this is about "reCAPTCHA" module so I am moving it there.

Well I see more issues with nginx and Drupal's asset system so we can ignore nginx as a cause. I am astonished there are not 404 errors, and if there really are none, then I don't really know where to look next.

Have you taken the step of comparing (diff-ing) the HTML output of a page with a broken layout vs the same page with a working layout? If not, that is something you could consider doing. It would provide more information.

Are there no errors logged by Drupal?

The web server type (Apache? nginx?) and its configuration would be helpful.

It would be possible to do something with that if you disable JavaScript aggregation, rebuild caches, then get that error with its stack trace for us.

Additionally, any information about what makes the affected Drupal site different from a newly-installed Drupal site will tend to help lead toward a resolution. Otherwise, we can only guess.

If this is the same report as 🐛 CKEditor5 table properties not visible after Drupal 10.4.0 update Active , what is the purpose of this issue?

Can you confirm that there are no browser console errors or Drupal log errors whatsoever?

Thanks for that update. I repeated the steps in I outlined in comment #4, in Drupal 10.4.4, which as I write this is the latest release, and I could not reproduce the bug.

I can't tell from this report but I assume these are 404 errors? The browser console will show them, if you would please check there.

If so, this is probably caused by the way CSS and JavaScript assets are concatenated has changed in recent Drupal releases. What is the web server platform configuration?

I can't reproduce this.

I did the following:

1. I installed a new Drupal site from the 11.x branch, which is at commit 501bbdbe21c257fc
2. I browsed to /admin/structure/types/manage/article/form-display.
3. I moved field Image to the Disabled section and saved the form.
4. Image remains in the disabled section.
5. I checked at /node/add/article that Image is in fact disabled.

Because of that, we need additional reproduction steps from you and perhaps error logging (including browser console errors, if any), to proceed here.

Is this with the experimental "Field Layout" module? In that case be aware of 📌 Inform users that Field Layout is deprecated Active .

Please add clear steps to reproduce in any case.

To reproduce this bug we need a commercial theme Martis and a custom theme. Do I have that right? I am tagging this bug "needs steps to reproduce" so that another community member can set up a reproduction site.

You could execute a git blame operation on a Git working copy of Drupal core to quickly find the commit that broke your code.

It seems 11.1.4 and 11.1.3 have the same number of untranslated strings.

Could you be more specific about exactly what you have seen?

greggles → credited cilefen → .

All support for Drupal 9 ended on 01 Nov 2023, so bug requests for it are not accepted. Look into setting the MySQL transaction isolation level → as that may help.

Please add steps to reproduce on a new Drupal site.

This will surely be blocked on 🌱 what's missing for a stable release ? Active . We would expect a stable release to exist before considering incorporating into Core.

Thanks for the patch. Can you create a merge request → of the same so the bots can test your changes?

Thanks for the report. If the resolution is that very clear I am tagging this "Novice" for the next steps.

Is any of the modules affected by 🐛 This form can only be submitted in the default workspace Closed: duplicate installed?

I can't reproduce the bug. Could you please add more steps to reproduce the bug, including any modules installed and in particular, CKEditor plugins?

To set up a reproduction someone using S3 will have to try. On the website with this issue:

1. Is the private file system working normally outside of Webform? That is, do ordinary private file fields function as expected on the website?
2. Are the private files accessible in the submissions by clicking on the attachments?

What is the S3 integration with the site? Is there a contributed module like https://www.drupal.org/project/s3fs → involved?

@jons I am setting these issue values

I watched the screen recording and based on what I saw, I am unsure what this bug is reporting. Is the issue that as you enable or disable filters, there may be errors flashing at the top of the page, which push page content down? For example CKEditor 5 needs at least the <p> and <br> tags to be allowed to be able to function may show if you disable then enable certain filters, and if it does show, content on the page moves downward in response.

Unfortunately nobody reading this will understand what "emp login" means because I don't know what that is. Is it part of the problem?

Also, as I mentioned above, Drupal 10.2 is not supported here. Upgrade to a supported release. It's possible this has been fixed.

No. I suggest you search the Internet for "Drupal\Core\DependencyInjection\ContainerNotInitializedException: \Drupal::$container is not initialized yet. \Drupal::setContainer() must be called with a real container".

Drupal 10.2 is no longer supported. The best advice I can give is to try with a supported release. Additionally, there could be a coding problem with a custom module in the site.

Can we get the stack trace and direct and simple steps to reproduce added to the issue summary please?

I think the best-known configuration for nginx is in ✨ Provide documentation/default server block for Nginx server. Needs work . If you are using a different configuration than that you are experimenting.

Got a stack trace?

The steps to reproduce are normal activities so this is probably more of a web server setup issue on your end. For example, update.ph/selection isn't a PHP script, unless that is just a typo. No one is going to be able to assist without more information that only you can provide.

This sounds like a bug.

git log origin/11.1.x..origin/11.x --oneline will show the unique commits in 11.x. A git bisect operation will determine the commit that changed a behavior.

@catch: You wrote the comment I had halfway written before I was distracted onto something else. Yes, we want these algos to be very fast. Moving to cryptographically-secure algos here makes no sense insomuch as I understand the use case.

On the 11.x branch, currently at 93ec7a93b738f0fd, I followed the steps to reproduce. I got the following user warning, which, if it's new information, also needs looking into, but no exception. Is there an additional action or step to be taken to get the LogicException?

User warning: Trying to overwrite a cache redirect for "response:[cookies:big_pipe_nojs]=:[languages:language_interface]=en:[request_format]=html:[route]=view.frontpage.page_19c08cb355eb3b07aca692af8c68cc11a934f5e596bf92a0449a69ff3afcd6d3f:[session.exists]=1:[theme]=olivero:[url.path.is_front]=is_front.1:[url.path.parent]=:[url.query_args]=:[url.site]=https://drupal.ddev.site:[user.node_grants:view]=all:[user.permissions]=is-admin:[user.roles:authenticated]=true" with one that has nothing in common, old one at address "user.permissions, session.exists, cookies:big_pipe_nojs, user.roles:authenticated, languages:language_interface, theme, url.query_args, user.node_grants:view, url.site, route, url.path.parent, url.path.is_front, route.menu_active_trails:account, route.menu_active_trails:main, url.query_args:_wrapper_format, request_format" was pointing to "user.roles:anonymous", new one points to "workspace". in Drupal\Core\Cache\VariationCache->set() (line 185 of core/lib/Drupal/Core/Cache/VariationCache.php).
Drupal\Core\Cache\VariationCache->set() (Line: 215)
Drupal\dynamic_page_cache\EventSubscriber\DynamicPageCacheSubscriber->onResponse() (Line: 246)
Symfony\Component\EventDispatcher\EventDispatcher::Symfony\Component\EventDispatcher\{closure}() (Line: 206)
Symfony\Component\EventDispatcher\EventDispatcher->callListeners() (Line: 56)
Symfony\Component\EventDispatcher\EventDispatcher->dispatch() (Line: 216)
Symfony\Component\HttpKernel\HttpKernel->filterResponse() (Line: 204)
Symfony\Component\HttpKernel\HttpKernel->handleRaw() (Line: 76)
Symfony\Component\HttpKernel\HttpKernel->handle() (Line: 53)
Drupal\Core\StackMiddleware\Session->handle() (Line: 48)
Drupal\Core\StackMiddleware\KernelPreHandle->handle() (Line: 28)
Drupal\Core\StackMiddleware\ContentLength->handle() (Line: 32)
Drupal\big_pipe\StackMiddleware\ContentLength->handle() (Line: 116)
Drupal\page_cache\StackMiddleware\PageCache->pass() (Line: 90)
Drupal\page_cache\StackMiddleware\PageCache->handle() (Line: 48)
Drupal\Core\StackMiddleware\ReverseProxyMiddleware->handle() (Line: 51)
Drupal\Core\StackMiddleware\NegotiationMiddleware->handle() (Line: 53)
Drupal\Core\StackMiddleware\AjaxPageState->handle() (Line: 51)
Drupal\Core\StackMiddleware\StackedHttpKernel->handle() (Line: 715)
Drupal\Core\DrupalKernel->handle() (Line: 19)

Do we really need two issues for this?

To move this issue forward, provide the YAML of a simple form that exhibits the bug.

Better?

Yes, it's just broken. I posted a naive merge here. Test it out.

The form element I fixed has @see \Drupal\webform\Plugin\Field\FieldWidget\WebformEntityReferenceAutocompleteWidget::formElement but that is gone or has moved, and I don't know the relevance.

cilefen → made their first commit to this issue’s fork.

Actually this title is more apt.

Most of that work took place at 💬 The hidden webform text format leads to text format selects with only one option Active . I don't think it needs to be enabled. It is disabled by default on install. What actual problem is this causing to the site you are working on?

I have taken those steps and including selecting British English as suggested in the issue summary. Unfortunately, I too cannot reproduce the bug. I suppose this is something to do with the translations download. Are there any browser console errors?

Because the steps to reproduce need some refinement as or additional information as #17 suggests I am tagging this issue that way.

Is that after clicking on User Interface -> Navigation Blocks? The description doesn't say. This may be essentially intentional because that administration screen is for editing the navigation blocks. What should happen instead?

@jatingupta40 Is that 10.4.1 or 11.1.1? Your steps to reproduce say both.

Additionally there is a specific release announcement for 10.4.3 → .

I am returning the issue metadata.