Fix clippy lints #404
Fix clippy lints #404
Conversation
I don't have this in any of my Rust projects, and I'd rather avoid it if possible. |
P1n3appl3
changed the title
|
Removed it, would you be ok with adding it as an entry in |
|
No opinion on whether to have |
|
Good point, I'll just add it to my global ignore |
|
Closing since people can use a global |
|
Ah, sorry, I didn't pay attention and thought this PR just added |
I saw those clippy warnings before and wasn't yet sure how to handle these. I think we should more carefully look into whether or not those appear in performance critical code. And if yes, whether or not boxing the error type would be beneficial. Otherwise, I'm fine with |
|
Not sure why the doc checks failed, re-running |
There was a problem hiding this comment.
It's possible to add an optional cargo clippy check, by making all other checks required, like CICD / all-jobs in sharkdp/bat#2913.
Even if a new cargo clippy check is optional, it still might encourage contributors to commit fewer linting errors into the codebase when they see a failing clippy check for their PR. clippy usually says how to fix any errors, too.
|
The previous comment is in reference to #355 (comment), but I thought I'd bring it up here, since it's relevant again. |
numbat/src/tokenizer.rs
Outdated
| @@ -694,6 +694,7 @@ pub fn tokenize(input: &str, code_source_id: usize) -> Result<Vec<Token>> { | |||
| } | |||
|
|
|||
| #[cfg(test)] | |||
| #[allow(clippy::type_complexity)] | |||
There was a problem hiding this comment.
Happy to merge this PR if we can remove these excepts. Here and below.
I use a global rustfmt config which cause spurious changes when formatting in this repo. Adding an empty
.rustfmt.tomlin the root will force rustfmt to use the default options for line length/comment reflow/etc.I also wanted to run clippy while making some other changes and cleaned up the existing lints so I could see any new ones I added. The only interesting one was
result_large_errin the typechecker where theOkvariants were around 32 bytes and some of theErrs were a couple hundred bytes. I just allowed it for that whole module but I could go allow the individual functions it fired on, or use boxing to make theResultsmaller.