chore: rename path package and change path

chore: inline composer.json using codeBlock

Adds a php require of ^8.1 and a config.platform.php of 8.1.99 everywhere as it exists in the root composer.json already. This basically tells renovate to use PHP 8.1 instead of the latest reachable one (e.g. 8.2) which could otherwise result in invalid update PRs like https://github.com/phpstan/phpstan-src/pull/2285

Ok, understood. There are other ways to configure renovate properly if it bothers you with weird PRs :)

chore: add dependency with a proper skipReason instead of completely ignoring it

Adds a php require of ^8.1 and a platform.php of 8.1.99 everywhere as it exists in the root composer.json already. This basically tells renovate to use PHP 8.1 instead of the latest reachable one (e.g. 8.2).

Bleeding edge - check invalid PHPDocs in previously unchecked statement types

Bleeding edge - InvalidPHPStanDocTagRule in StubValidator

InvalidPHPStanDocTagRule - added forgotten handling of ClassConst

Fix wrong positives about templates in conditional types

Update PHP 8 stubs

Update PHP 8 stubs

Regression tests

Closes https://github.com/phpstan/phpstan/issues/8408 Closes https://github.com/phpstan/phpstan/issues/7310

Fix build

Backed enums - dynamic return type extension for from() and tryFrom()

Fix build

Fix XMLReader::XML return type

TMP - always commit PHAR

PHAR - sign as PHPStan Bot

Revert "TMP - always commit PHAR"

This reverts commit 42e33e67b44c29f7084968092735f35f6348ec60.

I guess this could be seen as a fix even too? The actual feature is the path-based datasource support I suppose 🤔

Changes

Skips composer path repo dependencies in the extraction phase. Previously these were looked up on packagist.org instead which does either not work or might be problematic even.

Context

This is the first step for a path-based datasource as it was described in https://github.com/renovatebot/renovate/issues/11674#issuecomment-926065715

Documentation (please check one with an [x])

• [ ] I have updated the documentation, or
• [x] No documentation update is required

How I've tested my work (please select one)

I have verified these changes via:

• [ ] Code inspection only, or
• [x] Newly added/modified unit tests, or
• [ ] No unit tests but ran on a real repository, or
• [ ] Both unit tests + ran on a real repository

test: cover more cases by unit tests (#19086)

fix(cache): skip save repo cache on dry-run (#19094)

fix(nuget): disable msbuild node reuse (#19095)

docs(readme): update supported platforms (#19098)

feat(nuget): skip restore non-leaf projects (#19097)

chore: Disable consistent-return eslint rule (#19101)

fix(presets): pass through tag to platform for local (#19102)

feat(manager/gomod): dynamic golang install 🎉 (#19104)

chore(deps): update github/codeql-action action to v2.1.34 (#19108)

Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>

chore(deps): update dependency @openpgp/web-stream-tools to v0.0.13 (#19107)

Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>

build(deps): update dependency @opentelemetry/instrumentation-bunyan to v0.31.0 (#19109)

Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>

feat(manager/cargo): support dynamic install 🎉 (#19106)

chore(deps): update node.js to ba8317a (#19114)

Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>

chore(deps): update dependency eslint to v8.28.0 (#19115)

Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>

feat(branch/update): add edited/blocked PR comment on modified branches (#18895)

Co-authored-by: Rhys Arkins rhys@arkins.net Closes https://github.com/renovatebot/renovate/issues/18639

feat(manager/gradle): add support for alternative Gradle plugin syntax (#19113)

chore(deps): update dependency memfs to v3.4.12 (#19118)

Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>

fix(data): automatic update of static data (#19119)

Co-authored-by: Renovate Bot renovate@whitesourcesoftware.com

fix(gomod): non-0.0.0 digest matching (#19085)

fix(gomod): Revert "fix(gomod): non-0.0.0 digest matching" (#19120)

and dynamic return type extensions

thx. I was unaware of that :) that explains the behaviour here of course, because StrictComparisonOfDifferentTypesRule is doing exactly that (using getNativeType() instead of getType())

Ignore Nop nodes in NodeScopeResolver when creating UnreachableStatementNode

Remove @var Type where possible

Report narrowing PHPStan\Type\Type interface via @var

Fix build

I'd argue the other way that the expected output is also a reported error when treatPhpdocTypeAsCertain is disabled. Because in your current example the class indeed is always existing, so false is never returned. And PHPStan has this already figured out, but for some other reasons the check is not properly triggered 🤔

https://phpstan.org/r/8289b254-dc9f-4688-8882-6e55f83ae0a2

makes sense, thx for the inputs! I thought here it might be okay, because nobody ever should call these functions differently. I think it's a bit different compared to the linked PRs, but I get the general problem and point :)

Ignore Nop nodes in NodeScopeResolver when creating UnreachableStatementNode

Closes https://github.com/phpstan/phpstan/issues/8966

This is a weird one, there is indeed special handling for Nop statements in

https://github.com/phpstan/phpstan-src/blob/04edd0483f927e015c8ed1a3cd138046cd81d477/src/Rules/DeadCode/UnreachableStatementRule.php#L24-L26

but this seems to be dealing with comments to avoid triggering the rule for code like the following

public function doLorem()
	{
		return;
		// this is why...
	}

interestingly the ; seems to be not even a statement and therefore no Nop, it's just nothing. but the comment in your example is considered as Nop and if you remove it, the rule reports properly: https://phpstan.org/r/d687f1c8-c782-4df2-81f0-8f71e8d5ea24

not sure how to deal with this tbh, weird edge case

I don't know generics well enough, but it looks like the core problem is not related to array_filter, see https://phpstan.org/r/af2cb5f6-0b79-44fb-9b5f-7439ef3a781f

interestingly the dumped type is looking correct with bleeding edge enabled 🤔

Update dependency symfony/console to v6.2.7

Update compiler

Update PHP 8 stubs

Update nikic/php-parser

Fix value-of for multiple enums

Fix build

Fix typo

Fix ->value on unions of enums

Update BetterReflection

Regression test

Closes https://github.com/phpstan/phpstan/issues/8889

Fix changelog-generator

Process enum case expression

Fixed missing processing of UseUse and GroupUse

Fixed missing processing of Scalar

Fix interface accepting a Closure

Drop inline @var

Fix type aliases in method-level template types

Fix type aliases in traits

Check local type aliases above traits

Update PHP 8 stubs

Well, array_combine() never returns false in PHP 7.

your example could lead to a fatal error on 7.4 because of the array return type hint, see https://3v4l.org/H3sNV#v7.4.33, but it would indeed return false. also mentioned in the docs further down

8.0.0 array_combine() will now throw a ValueError if the number of elements for each array is not equal; previously this function returned false instead.