Niels Keurentjes

If the application stops working first stop should be to test whether your application can receive multicast UDP packets. In my experience most virtual machines will not pick them up...

On a related note it seems this one is also wrong: https://phpstan.org/r/5815a7c7-2d34-47af-8bd3-b264adfd3552, as phpredis happily casts integer keys to string keys internally.

I backtracked all the recent versions, ran into it in a project that I hadn't touched in weeks, and it did start in the same version as the `hMSet` issue....

That is not the correct solution, or at least not the full one. * The `has_side_effects` flag in that file is highly misleading, and its comments already mark it as...

I understand the rationale, is the default behavior currently based on method calls starting with get, has or is? If so I would consider doing so, as it means only...

That is not really an improvement 😆

Running into the same issue with long lived Messenger worker processes after upgrade to ORM 3. The lazy ghosts are reset between messages and reinstating them causes it to crash....

This PR is important, also for older PHP versions. Consider this code: ```php public function foo(): iterable { yield 'foo'; } public function bar(): void { dump(iterator_to_array($this->foo())); } ``` PHP...

@alcohol think you're referring to https://github.com/composer/composer/issues/6108, it had a different cause in the end but it also pointed to issues with shortest path detection. I can't really reproduce it as...

Just tested some things with `findShortestPath` and it doesn't seem to be causing this. Next possible culprit would be https://github.com/composer/composer/blob/master/src/Composer/Autoload/AutoloadGenerator.php#L866: ``` if (empty($installPath)) { $installPath = strtr(getcwd(), '\\', '/'); }...