Closed nfriedly closed 7 years ago
I don't understand why the exit code for --auto-correct
would matter. I think that option should only be used when running manually versus as part of a chain where the exit code matters (like on CI).
Well, I am running it in a chain, in my case I had this in my npm scripts:
"autofix": "eslint . --fix; dependency-lint --auto-correct",
The exit code doesn't block it, but it does cause npm to spew out lots of debugging junk.
Also, I just think it should report the current state, not what the state was before the tool ran. It's not a big deal, though. If you prefer it the current way, then just close this ticket.
Auhh. Yeah the npm part is a bit annoying but thats more an npm problem than anything else. Also I don't think it makes sense to have those two in a chain as you won't always need to fix both.
Fair enough.
I can see an argument for the current behavior, but I think that dependency-lint --auto-correct should work similar to eslint . --fix in that fixed errors do not count against the total errors, and if all errors were successfully fixed, then it should return 0 (no error).
Rather then ending with "✖ 1 error", it could say
✓ 1 correction ✓ 0 errors
For the moment, I'm currently putting dependency-lint --auto-correct || dependency-lint in my npm script.
BTW, this is a great tool either way, and I'm definitely going to be using it in more projects.