Arne Blankerts

I just spent some time looking into this. While I cannot yet say why the described error occurs, I can shed some light as to why we - as in...

I spent some more time trying to reproduce the issue and so far failed. So I guess, if we want to address this, we indeed need a bit of a...

I created a snapshot phar from the current main branch with an added debug dump before the exception is thrown. It does *not* address the issue in any other way:...

> If it helps anything, I introduced caching when I upgraded to phpunit 10, to order tests by the first failed previously to be run first on the next PR...

As we cannot see any way this can happen in our code, we assume it has to be something infrastructure related. We have some theories, but decided there's no value...

> > I created a snapshot phar from the current main branch with an added debug dump before the exception is thrown. It does _not_ address the issue in any...

Dealing with extensions isn't easy - as that opens the door to actual dependency handling. See #88 for details. PHPStan, if i recall correctly, is particularly problematic as it uses...

PHPUnit is actually one of the main reasons phive exits: As it's a tool with runtime impact when executing tests, collisions in packages are a nightmare. I fail to see...

Partially. > you might get rate limited by the GitHub API I don't know why Github rate limits their own action subsystem and not auto-grants the same API access levels...

I totally see that we need to find a better automated/CI support. I'm just not sure yet what approach is feasible without sacrificing security ;) Phive has quite some things...