🇪🇸Spain @lpeidro

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

I have checked the code and I looks great. I think is ready for new version more stable.

Thank you very much.

lpeidro → created an issue.

lpeidro → created an issue.

I’ve moved all the logic for calculating the path and route name of tasks and groups from the Plugin Manager to the Repository service, to decouple the logic and improve modularity.

I also created a trait to handle the logic for building task items in the controller, since the same logic applies to both cases.

The rest of the implementation looks great and works as expected. Thanks, Antonio, for your work!

I merge the branch.

I have merged the branch. Thank you Antonio and Jonathan.

There are many cases to manage. To avoid reducing the cases, a method has been implemented as the first one executed to detect whether the route corresponds to an entity route in the most flexible way possible, regardless of the entity type.

To achieve this, we use two criteria:

• The route name starts with "entity".
• There is a parameter of type entity converter.

Additionally, this parameter provides the key to retrieving the entity from the $route_match variable.

Ready for validation.

Thank you Albefer, I merge the branch.

lpeidro → changed the visibility of the branch 3512879-false-access-denied-link-with to hidden.

lpeidro → changed the visibility of the branch 3512879-false-access-denied-link-with to hidden.

lpeidro → created an issue.

I’m using this issue to introduce two minor changes:

• Restrict access to the ability to launch the Batch.
• Add the subcategory field and filter to the Entity Mesh view of nodes.

It works as expected.
I merge the issue.

I have to refactor how the database query is built. Using Aggregate Query and conditions is not working properly.

I tried adding a condition to the query, but I suddenly realized that it wasn't necessary. It was enough to simply add a new groupBy to the query.

This way, there is no need to run two queries—one for published and another for unpublished content.

    $query = $this->entityTypeManager
      ->getStorage('node')
      ->getAggregateQuery()
      ->accessCheck(FALSE)
      ->aggregate('nid', 'count', NULL, $alias_count)
      ->groupBy('type')
      ->groupBy('status');

    $query = $this->entityTypeManager
      ->getStorage("node")
      ->getAggregateQuery()
      ->accessCheck(FALSE)
      ->aggregate('nid', 'COUNT', NULL, $alias_count)
      ->groupBy('langcode')
      ->groupBy('type')
      ->groupBy('status');

Jonathan please, could review the code.

Thanks a lot.

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

It is ready to review and to merge. But I am not create unit test for this case. Maybe we can create a new issue to generate the unit tests.

lpeidro → created an issue.

The problem was in this part of the code in the class EntityRender:

  protected function setDataTargetFromRoute($target) {
    if (empty($target->getPath())) {
      return;
    }

    try {
      // @phpstan-ignore-next-line
      $route_match = \Drupal::service('router.no_access_checks')->match($target->getPath());
    }
    catch (\Exception $e) {
      $target->setSubcategory('broken-link');
      return;
    }

    // It is a view route.
    if (isset($route_match['view_id'])) {
      $target->setEntityType('view');
      $target->setEntityId($route_match['view_id'] . '.' . $route_match['display_id']);
      return;
    }

    // This case apply with entity canonical routes.
    $route_parts = explode('.', $route_match['_route']);
    if (count($route_parts) === 3 && $route_parts[0] === 'entity' && $route_parts[2] === 'canonical') {
      $entity_id = '';
      $entity = $route_parts[1];
      if (isset($route_match[$entity]) && $route_match[$entity] instanceof EntityInterface) {
        $entity_id = $route_match[$entity]->id();
      }
      $target->setEntityType($entity);
      $target->setEntityId((string) $entity_id);
    }
  }

The taxonomy term page is built using a view. For this reason, when the code detected that the view_id parameter was set in the $route_match variable, it handled the link as a view, without considering other parameters such as route_name.

When a taxonomy term had an alias, the entity was processed earlier, and this method was not called.

I have modified the code to be more precise in this aspect:

protected function setDataTargetFromRoute($target) {
    if (empty($target->getPath())) {
      return;
    }

    try {
      // @phpstan-ignore-next-line
      $route_match = \Drupal::service('router.no_access_checks')->match($target->getPath());
    }
    catch (\Exception $e) {
      $target->setSubcategory('broken-link');
      return;
    }

    if (empty($route_match['_route'])) {
      $target->setSubcategory('broken-link');
      return;
    }

    // @todo Maybe here is possible to get as well the entity object
    if (str_starts_with($route_match['_route'], 'view.')) {
      $target->setEntityType('view');
      $target->setEntityId($route_match['view_id'] . '.' . $route_match['display_id']);
      return;
    }

    // This case apply with entity canonical routes.
    $route_parts = explode('.', $route_match['_route']);
    if (count($route_parts) > 1) {
      $entity_id = '';
      $entity = $route_parts[1];
      if (isset($route_match[$entity]) && $route_match[$entity] instanceof EntityInterface) {
        $entity_id = $route_match[$entity]->id();
        $entity = $route_match[$entity]->getEntityTypeId();
      }
      $target->setEntityType($entity);
      $target->setEntityId((string) $entity_id);
    }
  }

lpeidro → created an issue.

It is ready for testing. Unit test added and ready to review.

Thank you, Jonhattan, I merge the branch.

The issue with false broken links occurred when removing the langcode from the href to search for the alias. The removal of the langcode deleted its occurrence anywhere in the string, which in some cases modified the URL.

The code has been updated so that it is only removed at the beginning of the string, if it exists.

lpeidro → created an issue.

lpeidro → created an issue.

It works as expected. I merge the branch. Thank you Antonio.

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

Merge the Issue, thank you Sady and Anotnio.

I added some minor points.

It works as is expected.

Thank you very much Antonio.

I merge the issue.

lpeidro → created an issue.

lpeidro → created an issue.

Taking advantage of this specific issue for the Paragraph case, we have implemented functionality to manage the dependency between insights and their plugin tasks.

The dependency on the modules is set in the Task Plugin definition. If the dependency is broken, the Task Plugin will be disabled. Similarly, in the insight, if the Task Plugin is inactive due to dependency issues, the corresponding insight will not be executed under any circumstances.

It is ready for testing.

Perfect, it works as expected. I merge the branch. Thanks Edu.

lpeidro → created an issue.

Hello,

We are experiencing this issue on a client project, but the solution proposed by maximpodorov is not working properly in our case and does not resolve the issue.

We have created a simple module that might be useful to help solve this problem Url Purge Aggregator → .

It is a very simple queuer that adds the URLs to the Purge queue when a node is edited or removed. This functionality can be enabled through configuration.

In this case, it would be necessary to implement the hook_entity_insert() hook to use the service provided by the module (url_purge_aggregator.queuer) to add the node's URL when it is created.

I am going to create an issue to add this option as well.

I hope this module can be useful.

lpeidro → created an issue.

lpeidro → created an issue.

Hello Omar:

Thank you very much for your contribution. However, the most appropriate way to fix this issue is to prevent the insight logic from running when the Paragraphs module is not enabled.

I believe it would be better to ask the plugins if they have any dependencies that are not available.

There is already some logic for this in the task plugins, but the result is not currently propagated to the insights.

lpeidro → created an issue.

Branch merged, and a new issue was created to apply the definitive fix: https://www.drupal.org/project/entity_mesh/issues/3509575 🐛 Manage dependencies in sites that are not multilingual Active

lpeidro → created an issue.

It works as expected. Thank you very much, Edu.
I merge the branch.

It works as expected. Thank you Edu, I merged the branch.

• I have added the GitLab CI configuration to run both static code analysis and unit tests in GitLab.
• I have resolved some issues and added compatibility with Drupal 10.
• I just merged it into the 2.x development branch.

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

I created a new branch (3506358-manage-dependency) to manage the service dependency without requiring the Language module to be enabled.

Great, thank you very much!!!

Hello, Joel:

We need this for a client to invalidate the CDN cache, as they are facing issues when editors update content or remove files.

At first, we planned to create a custom module, but then we found yours and didn’t want to reinvent the wheel.

However, due to the client's security policies, we can only implement custom modules or modules that are in stable versions.

I work for Metadrop, and we are very active in the community: https://www.drupal.org/metadrop →

Thank you.

lpeidro → created an issue.

Review the merge request and QA done.
It has been tested both with and without filters, and it works correctly.

It is mandatory to inform the user that is needed to manage the dependency with the module "views_data_export" before the update.

I merge the branch

Thanks, it works as expected. I merge the branch.

Thank you, guys. Great work. I merge the issue.

Hello Mortona:

Thank you very much for your collaboration. Yes, it is a good idea to show a statistic about the field usage.

We will include that information.

It works as expected. I merge the fix.

I’ve only made one modification, which is extracting the logic that calculates the language prefix from the target object to the Repository service, as it’s also needed during the processing of redirects.

Could you please take a look and do a quick QA? Thanks!

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

It works as expected. I merge the branch.

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

Thanks you very much Edu, fix merged.

We have added the possibility to download the report of block layout to be more useful for the user.

I have refactored the code to show the summary in a table:

This is different from the original proposal, but I realized that this format is clearer for the user.

I merge the branch.

Hello Sirclickalot:

Thank you very much for your feedback. We will fix the issue as soon as possible to ensure compatibility with sites that are not multilingual.

Best regards.

Merged funcionality.

lpeidro → created an issue.

Thank you Antonio and Sady for your colaboration.

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

Merge the branch. Thank you very much Antonio y Sady.

lpeidro → created an issue.

The types errors has been fixed and I used the service "router.route_provider" to get information the configuration of the views using their machine name. This information includes the path and if it is an admin route.

Maybe is a better way to this info.

I marge with the main branch. Thank you Sady for your feedback.

lpeidro → created an issue.

Added information about the access configuration of views to the Views report, along with a column indicating whether anonymous users can access each view. If it is an administrative view, the value will be displayed in red.

An insight has been included to alert about such cases.

A configuration object has been created, which can only be modified in the corresponding YAML configuration file, to include those views that, despite being on the administrative path, are configured to be potentially accessible to anonymous users.

Ready for testing.