Allow returning null from a PHPDoc void union

[herndlm]

Bug report

See phpstan/phpstan-src#2778 (comment) or WordPress/WordPress-Coding-Standards#2416 / https://github.com/WordPress/WordPress-Coding-Standards/blob/5f0c8a7ebdde0314ea5e30f76375a10c7d6f3910/WordPress/Sniffs/WhiteSpace/ObjectOperatorSpacingSniff.php

just to minimize the risk I forget..

fyi @jrfnl (also a big phpcs fan here 😊, thx for your work!)

Code snippet that reproduces the problem

https://phpstan.org/r/12a90bd8-64db-4620-855c-23579d0085f6

Expected output

No errors.

Did PHPStan help you today? Did it make you happy in any way?

No response

[staabm]

Maybe thats something we can error about in bleeding edge, but no longer error for non-bleeding-edge?

Or maybe move the whole void-to-null transformation behind bleeding edge entirely

[jrfnl]

Thanks @herndlm !

I've updated the example a little and added a second related code sample too (though that doesn't seem to error in the sandbox):
https://phpstan.org/r/20bb00b6-236f-4c0a-bb72-81415f5e9127

[ondrejmirtes]

This fixes the new issue introduced in the latest release: phpstan/phpstan-src@40c8fb2

I'm not sure what to do about Result of method HelloWorld::myrand() (void) is used. because that's been reported for a long time and didn't change recently. We'd have to do some kind of exception when void is typehinted only in PHPDoc, because otherwise it's not really valid code.

Also - I added WP CS into integration tests here so we're gonna now in the future when we'd break it with our changes :)

[jrfnl]

@ondrejmirtes Thank you for the swift action!

I'm not sure what to do about Result of method HelloWorld::myrand() (void) is used. because that's been reported for a long time and didn't change recently.

I wouldn't worry about that one, we made a choice to ignore that (for the previously mentioned stability reasons) before.

I added WP CS into integration tests here so we're gonna now in the future when we'd break it with our changes :)

That's great to hear. In the future, I'm looking to see if I can raise the level for PHPCS itself too (currently at level 0). That will likely result in more reports 😉

[ondrejmirtes]

Added PHPCS too :) c615993

[Brenneisen]

@ondrejmirtes I don't quite understand the change yet, but it seems that this is now causing problems in Larastan's unit tests. For example https://github.com/larastan/larastan/actions/runs/7197697590/job/19605436384?pr=1792 (larastan/larastan#1792).

[ondrejmirtes]

I'd say these results are better and more precise and the test expectations need to be adjusted on Larastan's master.

[github-actions]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.