Markus Staab

another idea which came to my mind, how infection could be faster: would it be possible to spread the load of killing mutants across more CPUs? the process as I...

> just to be clear: are you talking about GH only (or better to say - not locally)? > > Because we already spread processes across CPU cores thanks to...

that looks useful indeed

> I would have loved to be run Infection on a specific test on my previous project: doesn't `infection -s --filter=tests/MyExampleText.php` work like that?

I agree that I was also confused about infection `--filter` vs. PHPUnit `--filter`. ... and a command like `--test-framework-options="--filter='tests/MyExampleText.php'"` looks hard to remember and todo right without having to lookup...

Changing the implementation is a good improvement for the infection codebase. I agree that making the mutator smarter is likely not a good idea. Makeing infection smarter in mutations which...

I am not sure I agree. I see value in these mutations beeing separately categorized, because timed-out mutations are very inefficient and as I understood it they should be prevented/tackled...

> I think there was time when I purposefully fixed all "runaway" mutations on my projects, and it made MT really fast. So another way make MT faster is to...

after https://github.com/infection/infection/pull/2300 lands, timeout-mutants waste waay less time, so "give timeout-mutants the focus they deserve" no longer needs to be a priority

hopefully this benchmark makes sense: I have [commented out this `XdebugHandler::check()`](https://github.com/infection/infection/blob/247c6172d923c073e1d03509f044278ccea5bbad/src/Command/RunCommand.php#L587) to make sure no xdebug related restarts happen and the extension is loaded all the time. ran `php bin/infection...