Josh Pike

more info: included in a testing.yml in my GitHub actions as a simple bash command: ```node_modules/rome/bin/rome ci client/src/**/*.ts --colors=force ``` GitHub actions: my Mac: This is bonkers. We literally can't...

@ematipico after more experimenting I just realised... the ** star in the glob argument has inconsistent behaviour because it's not behaving recursively consistently. If I run it from Mac terminal,...

OK, that's kind of frustrating, it's the first time I've ever used something that doesn't implement globs. I spent way too much time trying to figure this one out! I...

I think an explicit warning that globs are not supported is needed. We’re in JS/TS tooling, webpack, tsconfig etc… globs are used everywhere and assumed to be supported by anything...

another example: ``` if (!rows || !rows.length || !rows[0]) return []; // no results ``` gets fixed to ``` if (!((rows && rows.length) && rows[0])) return []; // no results...

it's just too risky, I'm going to have to disable auto quick fix until I can be sure that these implementations are tighter.

@Conaclos I've found an interesting problem.. So this is a more detailed story of the problem and how to reproduce. It seems to be a problem specifically when auto fix...

this rule is critical for our projects, it prevents some very nasty situations from occurring, as floating promises can cause sync issues with our data model.

I understand. This rule is actually so critical that I think I'm going to bring back eslint into our CI tests just for that rule alone.

I actually think that this would be the answer, not something a linter should handle as it fundamentally can change the behaviour or the code. Don't really know if it's...