Markus Staab

let me re-phrase what the actual problem is what I am trying to solve here. we discussed at several places about `--skip-initial-tests` and the pre-condition to have a green test-suite...

yes. I think if infection users could do that, that would ease integration of infection into their projects and simplify the setup -> reduce hurdles. it would simplify ci configs...

> please note that 310 here is not the exact value of how many tests were executed. ouch, this broke my assumptions ;-). maybe we should invoke phpunit in a...

> could you please elaborate on this? how? what does it mean? my thinking was: we don't produce output in infection for the initial test-run, but just invoke phpunit and...

thank you! --- created isolated repro: https://github.com/staabm/phpstorm-bug-unreachable-statement reported to Jetbrains in https://youtrack.jetbrains.com/issue/WI-81787/Unreachable-statement-false-positive

Random would be possible. As soon as we support levels I think we should report those 20 with the lowest levels

looking at a counter, I can see the PR is removing 131 PublicVisibility mutations. how many time this saves depends on how slow the test-suite is and how fast/effectively phpunit...

@maks-rafalko I think the PR works now as expected - its still not faster though. it seems in the infection/infection repo the unit test-suite is optimized for infection using CoversClass...

interessting that on my local machine infection reports > "killed with AST: 149" while in [CI I see](https://github.com/infection/infection/actions/runs/15559978698/job/43809604107?pr=2163) > "killed with AST: 74"

> is it the same PHP version as you use locally? yes tried on 8.2.x and 8.4.x and both report the same thing locally > For BetterReflection thanks for testing....