clj-kondo

https://github.com/clj-kondo/clj-kondo
Helins 2021-05-30T09:56:44.099400Z

I get this error:

gen.cljc:1:57: error: Invalid digit ` in unicode character.
Which is weird since I use ` a lot and it has never complained + there is actually nothing on 1:57

Helins 2021-05-30T09:58:44.100Z

And it complains even if I delete all ` from the namespace

borkdude 2021-05-30T10:02:24.100300Z

never seen that error before

borkdude 2021-05-30T10:02:44.100600Z

perhaps you're loading hooks?

borkdude 2021-05-30T10:03:03.101Z

set aside your clj-kondo config to be sure it's nothing related to that

Helins 2021-05-30T10:06:47.101900Z

Not it... I don't know when it appeared. I must have changed something and it somehow confused the linter. But the namespace compiles and works well

borkdude 2021-05-30T10:10:06.102600Z

maybe try setting aside your entire .clj-kondo dir including the cache

Helins 2021-05-30T10:11:28.103900Z

Alright, I got it I think. On a completely different line I have this docstring:

"Any valid character for hexadecimal notation, from `\0` to  `\f`."

Helins 2021-05-30T10:12:19.104700Z

I mean, this is problematic:

`\0
`

borkdude 2021-05-30T10:13:43.105100Z

do you think clj-kondo should accept it, hence a false positive?

Helins 2021-05-30T10:16:36.106700Z

It is a false positive, I was trying to highlight \0 in markdown for documentation but it thinks I am trying to write a unicode char \0`

Helins 2021-05-30T10:16:55.107200Z

But it's very specific, I wouldn't worry too much about it, it's just puzzling when it happens

borkdude 2021-05-30T10:17:15.107600Z

feel free to post an issue, it might be a rewrite-clj specific thing

1👍
borkdude 2021-05-30T10:17:20.107800Z

or tools.reader

Helins 2021-05-30T10:34:39.109400Z

Hmmm, I am sure I am right with that \0 unicode thing (since error goes away) yet I can't reproduce it in a minimal ns. Must be some kind of weird interaction. I can't share publicly this namespace. But it's probably an exceptional case, so no big deal at all.