Théo FIDRY

>we need to see what other mutation testing tools mean under MSI Universally means `nbr of mutants killed / total nbr of mutants`. Free of interpretation if you mutate uncovered...

You can check the draft #1065 to see what it entails: IMO it simplifies a lot of things besides the performance improvements

I asked for mule they don't. >Google seems to ignore uncovered code in their report on mutation testing. Google is a special case I wouldn't base myself on: they have...

>I think this ability for Infection to evaluate total number of escaping mutations, be it due to missing coverage or faulty tests, is a valuable feature. The fact that other...

As mentioned [here](https://github.com/infection/infection/pull/1065#issuecomment-589649613) this is not an issue. Right now your "single metric", when including non-covered, is both fairly expansive (your reported a 14% which is far from being nothing...

Did a PoC at #1066: getting the code coverage score to then be able to calculate what we wish looks relatively simple. So I'm confident in doing a PR in...

In what way does it currently do that? On Sat 22 Feb 2020 at 13:04, Alexey Kopytko wrote: > This computed metric disregards actual density of complexity per line of...

I think there is plenty of other metrics that would fit better the bill there, Cyclomatic complexity or CRAP for example. You can have a single array assignment and have...

@localheinz @maks-rafalko @BackEndTea would love your input there as well :)

@maks-rafalko I would take that information from mutant with a grain of salt. Let me explain, to use mutant, you rely on declaring what test test what. Pretty much what...