Update testing

this likely needs to be split up in more steps, we should also either in those issues or here mark the validation jobs that are fixed as allow failure false.

to void conflicts between diffrent phpstan jobs, phpstan on previous major should be disabled.

Depends on the result. phpcbf includes several fixes that I think are pointless, such as adding empty/placeholder docblocks just to get rid of missing docblock complaints, I will not accept that, that's not an improvement. comment stuff in general can be challenging to review.

I'm not requesting to have a separate issue for every single distinct issue.

A compromise might be an issue that does phpcbf but excludes anything that changes comments, at least wording part. Although something like adding a . to a weird non-sentence comment isn't really going to fix it, so I think an easy rule is just ignore anything comment related in a first pass.

another challenging topic with phpcbf is changing types in functions which technically is a BC break, such as adding a arrays to variables documented as array. That's kind of still allowed as this is still in beta, but I also don't think it's worth possibly breaking someone's code over that.