[jirkace]

Psalm gives a static method that is declared : static the parent's class instead of the called class when it is called as static::, self:: or Child::, so a property of the child is "not defined".

Before

abstract class ParentException extends Exception
{
    public static function withMessage(string $message): static
    {
        return new static($message);
    }
}

class MyException extends ParentException
{
    private int $user = 0;

    public static function create(string $message, int $user): static
    {
        $exception = static::withMessage($message);
        $exception->user = $user; // UndefinedPropertyAssignment: Instance property ParentException::$user is not defined

        return $exception;
    }
}

After

        $exception->user = $user; // no issue ($exception is MyException&static)

Refs #7485

[jirkace]

Behaviour change

• A native : static return type is scanned as Parent&static (class of the declaring class, is_static). TypeExpander::expandNamedObject() only replaced the static of a docblock @return static (value static) with the class the method is called on, so for a native return type called through static::, self::, parent:: or a class name from a child, the result stayed Parent&static. A new branch resolves such a type to the called class (Child&static) when it is given as a class name that is, or extends, the declaring class; instance calls already did.
• Visible effects: no UndefinedPropertyAssignment, LessSpecificReturnStatement or MoreSpecificReturnType for code like the above (the report of Invalid InvalidPropertyAssignmentValue when assigning value to a private property #7485, also with self::withMessage()); $x = static::make() traces C&static instead of P&static, as it did for a docblock @return static before.
• Verified: the new valid case fails on master (UndefinedPropertyAssignment) and passes here; the neighbouring single cases with static/fluent/self in their names (up to 12 per file in MethodCallTest, ReturnTypeTest, ClassTest, ClassTemplateTest) all pass (the loop printed no failure; I did not record how many it ran); phpcs and full self-analysis are clean; the psalm/endtoend-test-phpunit run with its .psalm/config.xml reports no error. Not run: the rest of the test suite and the perf report.

[github-actions]

✅ Performance: unchanged · Issues: decreased

Project	Memory (MB)	Wall (s)	Issues	Weighted
✅ laravel	1 821	63.9 $\color{gray}{\textsf{+0.2\%}}$	25 916 $\color{red}{\textsf{+1}}$ $\color{green}{\textsf{-2}}$	51 156 $\color{green}{\textsf{-3}}$
✅ symfony	4 002	131.7 $\color{gray}{\textsf{+2.1\%}}$	51 065 $\color{gray}{\textsf{+0}}$ $\color{green}{\textsf{-16}}$	101 056 $\color{green}{\textsf{-48}}$
✅ doctrine-orm	428	13.2 $\color{gray}{\textsf{+1.7\%}}$	3 505 $\color{gray}{\textsf{+0}}$ $\color{green}{\textsf{-2}}$	5 574 $\color{green}{\textsf{-6}}$
✅ psl	651	17.3 $\color{gray}{\textsf{+0.3\%}}$	1 424	5 514
✅ phpunit	506	10.4 $\color{gray}{\textsf{-0.7\%}}$	1 627	1 265
✅ php-parser	245	9.5 $\color{gray}{\textsf{-0.3\%}}$	17 169	20 624
✅ tempest	922	20.1 $\color{gray}{\textsf{-6.1\%}}$	5 016	9 616
✅ amphp-amp	60	1.4 $\color{gray}{\textsf{-0.7\%}}$	32	79
✅ webmozarts-assert	272	3.0	577	2 714
✅ guzzle-promises	55	1.4 $\color{gray}{\textsf{-0.7\%}}$	163	407
✅ monolog	161	3.5 $\color{gray}{\textsf{+1.2\%}}$	1 004	2 722
✅ drupal	5 196	146.1 $\color{gray}{\textsf{+0.6\%}}$	98 829 $\color{gray}{\textsf{+0}}$ $\color{green}{\textsf{-1}}$	161 671 $\color{green}{\textsf{-4}}$
Total	14 319	421.3 $\color{gray}{\textsf{+0.6\%}}$	206 327 $\color{red}{\textsf{+1}}$ $\color{green}{\textsf{-21}}$	362 398 $\color{green}{\textsf{-61}}$

How to read this table

• Memory: Psalm's own peak figure, threads=1. More than +5 % fails, unless the PR changes composer.json.
• Wall: base and head run on the same runner; ±10 % is noise. Above +20 % warns, never fails. re-run: a reversed second run was taken and the faster kept.
• Issues: head count, then $\color{red}{\textsf{+added}}$ $\color{green}{\textsf{-removed}}$ against base.
• Weighted: issues weighted by error level (levels 1–8 as is, always-errors 10, feature-specific and purity suggestions 0). The projects' own code is pinned and their dependencies resolve identically for base and head, so a change means Psalm reports more, fewer or differently serious issues. A rise is not necessarily a regression.
• Total: sum over the projects compared on both sides. Thresholds apply per project, not to the total.
• Colors: $\color{red}{\textsf{red}}$ past a fail or warn threshold (or more issues), $\color{green}{\textsf{green}}$ an improvement past noise (or fewer issues), $\color{gray}{\textsf{gray}}$ noise.
• Performance: regressed if a head run crashed or timed out or a memory or wall delta is red, else improved if one is green, else unchanged. Its icon is the check result: ❌ fail · ⚠️ warn · ✅ pass; the icon next to each project is that project's result, 💥 = not comparable: base crashed or timed out (counts as ⚠️).
• Issues (heading): the total weighted change, increased · decreased · unchanged. Informational, never fails the check.

✅ laravel: issues +1 −2

Memory 1 821 → 1 821 MB (<0.1 %) · Wall 63.8 → 63.9 s (+0.2 %)

Weighted 51 159 → 51 156 (−3, −0.01 %): L3 −1

Issue type	added	removed
LessSpecificReturnStatement	0	1
MoreSpecificReturnType	1	1

Added (1)

• src/Illuminate/Database/Eloquent/HasCollection.php:24 MoreSpecificReturnType (error): The declared return type 'Illuminate\Database\Eloquent\Collection<array-key, Illuminate\Notifications\DatabaseNotification&static>' for Illuminate\Database\Eloquent\HasCollection::newCollection is more specific th…

Removed (2)

• src/Illuminate/Http/Request.php:565 MoreSpecificReturnType (error): The declared return type 'Illuminate\Http\Request&static' for Illuminate\Http\Request::duplicate is more specific than the inferred return type 'Symfony\Component\HttpFoundation\Request&static'
• src/Illuminate/Http/Request.php:570 LessSpecificReturnStatement (error): The type 'Symfony\Component\HttpFoundation\Request&static' is more general than the declared return type 'Illuminate\Http\Request&static' for Illuminate\Http\Request::duplicate

✅ symfony: issues +0 −16

Memory 4 002 → 4 002 MB (<0.1 %) · Wall 129.0 → 131.7 s (+2.1 %)

Weighted 101 104 → 101 056 (−48, −0.05 %): L3 −16

Issue type	added	removed
LessSpecificReturnStatement	0	8
MoreSpecificReturnType	0	8

Removed (16)

• src/Symfony/Component/String/AbstractUnicodeString.php:485 MoreSpecificReturnType (error): The declared return type 'Symfony\Component\String\AbstractUnicodeString&static' for Symfony\Component\String\AbstractUnicodeString::trimPrefix is more specific than the inferred return type 'Symfony\Componen…
• src/Symfony/Component/String/AbstractUnicodeString.php:488 LessSpecificReturnStatement (error): The type 'Symfony\Component\String\AbstractString&static' is more general than the declared return type 'Symfony\Component\String\AbstractUnicodeString&static' for Symfony\Component\String\AbstractUnicod…
• src/Symfony/Component/String/AbstractUnicodeString.php:524 MoreSpecificReturnType (error): The declared return type 'Symfony\Component\String\AbstractUnicodeString&static' for Symfony\Component\String\AbstractUnicodeString::trimSuffix is more specific than the inferred return type 'Symfony\Componen…
• src/Symfony/Component/String/AbstractUnicodeString.php:527 LessSpecificReturnStatement (error): The type 'Symfony\Component\String\AbstractString&static' is more general than the declared return type 'Symfony\Component\String\AbstractUnicodeString&static' for Symfony\Component\String\AbstractUnicod…
• src/Symfony/Component/String/UnicodeString.php:204 MoreSpecificReturnType (error): The declared return type 'Symfony\Component\String\UnicodeString&static' for Symfony\Component\String\UnicodeString::join is more specific than the inferred return type 'Symfony\Component\String\AbstractUnicodeString…
• src/Symfony/Component/String/UnicodeString.php:209 LessSpecificReturnStatement (error): The type 'Symfony\Component\String\AbstractUnicodeString&static' is more general than the declared return type 'Symfony\Component\String\UnicodeString&static' for Symfony\Component\String\UnicodeString::join
• src/Symfony/Component/String/UnicodeString.php:288 MoreSpecificReturnType (error): The declared return type 'Symfony\Component\String\UnicodeString&static' for Symfony\Component\String\UnicodeString::replaceMatches is more specific than the inferred return type 'Symfony\Component\String\AbstractUni…
• src/Symfony/Component/String/UnicodeString.php:293 LessSpecificReturnStatement (error): The type 'Symfony\Component\String\AbstractUnicodeString&static' is more general than the declared return type 'Symfony\Component\String\UnicodeString&static' for Symfony\Component\String\UnicodeString::replaceM…
• src/Symfony/Component/String/UnicodeString.php:402 MoreSpecificReturnType (error): The declared return type 'Symfony\Component\String\UnicodeString&static' for Symfony\Component\String\UnicodeString::trimPrefix is more specific than the inferred return type 'Symfony\Component\String\AbstractUnicode…
• src/Symfony/Component/String/UnicodeString.php:428 LessSpecificReturnStatement (error): The type 'Symfony\Component\String\AbstractUnicodeString&static|Symfony\Component\String\UnicodeString&static' is more general than the declared return type 'Symfony\Component\String\UnicodeString&static' for Sy…
• … and 6 more

✅ doctrine-orm: issues +0 −2

Memory 428 → 428 MB (<0.1 %) · Wall 13.0 → 13.2 s (+1.7 %)

Weighted 5 580 → 5 574 (−6, −0.11 %): L3 −2

Issue type	added	removed
LessSpecificReturnStatement	0	1
MoreSpecificReturnType	0	1

Removed (2)

• src/Mapping/ToOneInverseSideMapping.php:35 MoreSpecificReturnType (error): The declared return type 'Doctrine\ORM\Mapping\ToOneInverseSideMapping&static' for Doctrine\ORM\Mapping\ToOneInverseSideMapping::fromMappingArrayAndName is more specific than the inferred return type 'Doctrine\ORM\Mapping\As…
• src/Mapping/ToOneInverseSideMapping.php:50 LessSpecificReturnStatement (error): The type 'Doctrine\ORM\Mapping\AssociationMapping&static' is more general than the declared return type 'Doctrine\ORM\Mapping\ToOneInverseSideMapping&static' for Doctrine\ORM\Mapping\ToOneInverseSideMapping::fromMapping…

✅ drupal: issues +0 −1

Memory 5 196 → 5 196 MB (<0.1 %) · Wall 145.2 → 146.1 s (+0.6 %)

Weighted 161 675 → 161 671 (−4, <0.01 %): L4 −1

Issue type	added	removed
ImplementedReturnTypeMismatch	0	1

Removed (1)

• core/lib/Drupal/Core/Entity/Query/ConditionFundamentals.php:95 ImplementedReturnTypeMismatch (error): The inherited return type 'Drupal\Core\Entity\Query\ConditionInterface&static' for Drupal\Core\Entity\Query\ConditionInterface::setQuery is different to the implemented return type for Drupal\Core\…

_{head 3ee5e5e · base b0cb649 (merge-base) · workflow run}

[jirkace]

Behaviour change (follow-up commit: the failing unit-test shards, 6/8 on all PHP versions)

• ClosureTest::FirstClassCallable:StaticMethod:ClassTemplatesAndStatic expected static::make(...) and self::make(...), called inside Ints extends Collection, to be Closure():(Collection&static), i.e. the declaring class of the native : static return type. That is the behaviour this PR changes: the closure created from the call now returns Ints&static, like the call static::make() itself. The two @psalm-check-type-exact expectations are changed to Closure():(Ints&static); the Ints::make(...)/FinalInts::make(...) assertions of the same case were already right and still pass.
• Verified: that case passes, as does PartialApplication:StaticReturnOfANonFinalReceiver (which already expected P&static for static::create(?)) and the other Static/Partial/FirstClass single cases of ClosureTest and PartialFunctionApplicationTest (the loop printed no failure). Not run: the rest of the shards (CI will tell).

[orklah]

Can you rebase and check interactions with merged #12643

[jirkace]

@orklah Rebased onto master (new head 3ee5e5e). The only conflict was in tests/MethodCallTest.php, where master added mysqliGetWarningsMayReturnFalse at the same spot; both entries are kept.

About #12643: it is not merged yet (still open). It fixes the same behaviour as this PR and #13401. This branch passes the new valid cases of #12643 and #13401, and #12643 passes this PR's case, so only one of them is needed. Merging this with #13401 conflicts only in the test file: both TypeExpander branches give the same result, so they don't conflict. This PR only rewrites a plain TNamedObject return type, so it doesn't touch static<...> generics, which is where #12643 regressed.