Disabling a non-existant check should not be a fatal error about go-critic HOT 4 OPEN

warnings tend to get ignored and some tools end up having 10+ warnings printed out for no reason.

I like the idea that requesting something that does not exist is an error. We can disagree on that one if you would like.

For me, the main problem is the checker being removed. We usually don't do that.

And if we do, it's usually a checker that was marked with an experimental tag.
If users rely on experimental checkers stay forever - it's outside of the backward compatibility promises.

404c56b#diff-389d96f065f3108bd8fbfd7a928e9c9b81f4f5f308a021fc539fc09a49345763L789

The checker you've mentioned was experimental.
If think it's fair to say that experimental stuff can break.

Unless we remove the experimental tag, the checker is basically in its public testing phase.
You may use it if you find it useful, but we still have some freedom in making changes and/or removing it.

from go-critic.

Yeah, kinda make sense. Sorry for that. Wanna make a PR ? 😉

from go-critic.

How can I reproduce this fault @natefinch?

$ ./bin/gocritic check -disable 'ehsandar,offBy1,dummyName,#dummyTag' -enable 'offBy1' ./...
[offBy1]
[ehsandar offBy1 dummyName #dummyTag]
init checkers: empty checkers set selected

from go-critic.

I was running gocritic via golangci-lint, so I am not sure how to do it via gocritic itself, but I just had sloppyTestFuncName configured to be disabled. Ironically, I realize now that we had also removed experimental checkers from our list of enabled categories, so it was redundant to have it in the disabled list.

from go-critic.