Maks Rafalko

If we manage to make it, I think we can reopen https://github.com/infection/infection/issues/2212 cause we will be safe when develop changes the code but uses stale CC

> 1. One (although I expect minor) glaring flaw is that we re-create `TestLocation` objects, although they are identical. I > do not know the impact but it feels wasteful...

> 2\. Is the outer node `outer node` correct and sufficient? I wonder if we should not select the parent method/function > instead. Thoughts? then I still don't get the...

the we can do something like ```shell --stop-on-escapes= # accumulate 10 and stop --stop-on-escapes=10 # stop on the first escape (default=1) --stop-on-escapes ``` this way - `--show-mutations=` works as before...

> What do you think about adding `--stop-on-escape` that will translate into `--stop-on-escapes=1`? it's fine

can it be https://github.com/CuyZ/Valinor @staabm? I just saw your performance observations there, this is exactly what was proposed by you here :) https://github.com/CuyZ/Valinor/pull/652

1. do you have any example where a similar idea is implemented (phpstan? rector? somewhere else?) I would like to see how it works on GH actions, whether it's executed...

Thank you for the input, very interesting. > As for the AST parsing, we only load one file at a time, and discard this whole file afterwards, so the GC...

Thanks! As an author of `mago`, are you interested in working on this? As I'm not sure when/if I will have time for this.

> Agree that adding a new option is a pain. 👍 It is much less painful than it was before after this https://github.com/infection/infection/pull/2304 I agree PoC would be nice to...