Markus Schirp
Markus Schirp
@vaporyhumo I agree and: I think we should generally forbid to recurse into `T` constants. This is very easy to patch. Wanna take a stab at it?
Super high level: Tests for mutation operators `meta/` directory. Code: Find the emitter that handles send nodes, and prevent any mutation in case the receiver is the constant `T`. ```...
@vaporyhumo I recommend you do this increments (in separate PRs): 1. Ignore all mutations on `T.` via early exit. 2. Turn the early exit into a delegate to a sorbet...
@vaporyhumo this can be done now via user specific settings, see: #1341. I agree mutant should do this by default, but having user control is a step up.
Yes this is "not" the final solution the real solution is to run sorbet post mutation to use its erroing as pointer to not remove the `T.must` and friends. I...
> I plan to build this as a third party extension for mutant when I have time. I can tell you now though that @mbj is not interested in any...
I like this idea. The existing infrastructure should be able to support this.
@mvz I believe this is a duplicate of an existing issue we agreed upon fixing. Till I've the time to reference the original I'll keep it open. I agree on...
> If it showed up in your github notifications again it is probably because I mistakenly commented on it last night (and then deleted my comment when I realized it...
@joneshf This was discussed multiple times in the past. The conclusion is: * Yes I support this * The required APIs internally exist already, so production of the results is...