Apply fully_qualified_strict_types in annotations

[ostrolucky]

I would like to apply this fixer across annotations as well. Here's what I expect to fix

<?php

class SomeClass
{
    /** @Foo\Bar */
    private $foo;
}

into

<?php

use Foo\Bar;

class SomeClass
{
    /** @Bar */
    private $foo;
}

How do you think solution should look like? As far as I see, so far no fixer outside DoctrineAnnotation fixes annotations

[julienfalque]

👍 though I think I would be better to have a dedicated fixer in the DoctrineAnnotation namespace.

[Slamdunk]

Already discussed in #3644

[ostrolucky]

Ah thanks for a link. Indeed, it would make sense to fix both cases (phpdoc param/return types and doctrine style classes prefixed with @).

[ostrolucky]

After looking at it some more, I have realized FullyQualifiedStrictTypesFixer does not do what I want anyway. It changes FCQNs to short versions only if class has been imported. My goal is to import the class if it hasn't been imported.

[ostrolucky]

For all of those who are looking for importing fcqn classes which were not imported, there are

SlevomatCodingStandard\Sniffs\Namespaces\ReferenceUsedNamesOnlySniff which does this
For PHP-CS-Fixer, there is Symplify\CodingStandard\Fixer\Import\ImportNamespacedNameFixer

[mvorisek]

I belive this was fixed by #5620

[Wirone]

@mvorisek no, #5620 was about phpDoc tags only (@tag \F\Q\C\N), not annotations (@F\Q\C\N).

[keradus]

btw, what is this /** @MyClass */ annotation usage?

[mvorisek]

"attributes defined in phpdoc"

/* @ORM\Column */ for example

https://packagist.org/packages/doctrine/annotations

[keradus]

I think Doctrine moved to attributes by now, if it's the only use case, maybe there is no big value of creating support for it.

[Wirone]

From doctrine/annotations:

⚠️ PHP 8 introduced attributes, which are a native replacement for annotations. As such, this library is considered feature complete, and should receive exclusively bugfixes and security fixes.

So it has limited support, but is not strictly deprecated. From our perspective it shouldn't be a priority IMHO, but if anyone is willing to contribute, then we can consider merging 😉.

[mvorisek]

with #7679 this can be easily implemented