Fix PHPStan L2 errors "Parameter $foo of method Foo::bar() has invalid type Foo\Baz." and "Method Foo::bar() has invalid return type Foo\Baz. "

It would be good to have a coding standard change to allow using non-FQCN in comments, rather use imports on file headers and then just the class names. PHPStan fully supports that, and IMHO readibility would greatly improve.

1. +++ b/core/lib/Drupal/Core/Archiver/Tar.php
   @@ -81,7 +81,7 @@ public function listContents() {
       * Archive_Tar object for implementation-specific logic. This is for advanced
       * use only as it is not shared by other implementations of ArchiveInterface.
       *
   -   * @return Archive_Tar
   +   * @return ArchiveTar
       *   The Archive_Tar object used by this object.
       */
   

   need to adjust the docs above and below, too.

2. +++ b/core/lib/Drupal/Core/Config/Entity/ConfigEntityStorage.php
   @@ -406,7 +406,7 @@ public function createFromStorageRecord(array $values) {
   -   * @return \Drupal\Core\Config\ConfigEntityInterface
   +   * @return \Drupal\Core\Entity\EntityInterface
   

   shouldn't it be \Drupal\Core\Config\Entity\ConfigEntityInterface?

3. +++ b/core/lib/Drupal/Core/Plugin/ContextAwarePluginTrait.php
   @@ -15,7 +15,7 @@
   - * @see \Drupal\Core\Plugin\ContextAwarePluginInterface
   + * @see \Drupal\KernelTests\Core\Plugin\Context\TestContextAwarePlugin
   

   mmm sounds strange to refer a runtime trait to a test class...

4. +++ b/core/modules/jsonapi/src/Normalizer/FieldNormalizer.php
   @@ -83,7 +83,7 @@ public function denormalize($data, $class, $format = NULL, array $context = []):
   -   * @return \Drupal\jsonapi\Normalizer\Value\FieldItemNormalizerValue[]
   +   * @return array
   

   this is losing information, could it be avoided?

5. +++ b/core/modules/migrate/src/Plugin/Discovery/ProviderFilterDecorator.php
   @@ -55,7 +55,7 @@ public function __construct(DiscoveryInterface $decorated, callable $provider_ex
   -   * @return array|\mixed[]
   +   * @return array|mixed[]
   

   kinda redundant, could it be just one of the two? actually with #3309010: Support PHPDoc Types in @param @var @return annotations all this could be vastly improved

6. +++ b/core/tests/Drupal/Tests/Core/Entity/EntityUrlTest.php
   @@ -415,7 +415,7 @@ public function testUriRelationships() {
   -   *   The class name to mock. Should be \Drupal\Core\Entity\Entity or a
   +   *   The class name to mock. Should be \Drupal\Tests\Core\Entity\UrlTestEntity or a
   

   comment should be wrapped on 80 chars

Not a big brain, still I'd give some feedback :)

#4.1 IMHO we should do nothing. #3252386: Use PHP attributes instead of doctrine annotations would probably make these classes redundant at some point in time. Until then, I see no issue to keep these errors in a baseline (so that any change on that code in the meatime at least would not add errors).

#4.2 'callable' is a basic type, it should be all lowercase and no backslash in front.

+++ b/core/lib/Drupal/Core/Plugin/CategorizingPluginManagerTrait.php
@@ -74,7 +74,7 @@ public function getModuleHandler() {
-    /** @var \Drupal\Core\Plugin\CategorizingPluginManagerTrait|\Drupal\Component\Plugin\PluginManagerInterface $this */
+    /** @var \Drupal\Component\Plugin\PluginManagerInterface $this */

@@ -88,7 +88,7 @@ public function getCategories() {
-    /** @var \Drupal\Core\Plugin\CategorizingPluginManagerTrait|\Drupal\Component\Plugin\PluginManagerInterface $this */
+    /** @var \Drupal\Component\Plugin\PluginManagerInterface $this */

This seems rather strange... does it fail if we totally remove the hint? One may wonder why we need to have a trait if $this is an implementation of an interface at this point?

Adding the PHPStan-1 tag as it seems this fixes some L1 baseline errors too.

I suppose you meant PHPStan, I can't see how this could fail PHPUnit. What happens if we remove the typehint completely? What's strange to me is typehinting $this (which is the current state, not the change itself)

Looks good catch to me of all those rectifying the issues,spelling,typo etc..

I see one last thing, commented in the MR.

See inline comment.

Same here. https://www.drupal.org/project/drupal/issues/3191623#comment-14861657

I switched back to patch workflow.

Removing credit from myself as I just rebased. Will mark after that runs.

No issues LGTM

One question about the session handler parameter but otherwise this looks ready to go.

Improvements of docs. LGTM

Committed and pushed to 10.1.x, thanks!