Clippy Analysis to CI

[mastrzyz]

Enables Clippy in the CI as a need of the top-level ci-pass job.
Resolution for #2977
Disabled the current violations :

#![allow(
    clippy::needless_borrow,
    clippy::module_inception,
    clippy::useless_format,
    clippy::redundant_clone,
    clippy::manual_non_exhaustive,
    dead_code,
    unused_imports
)]

And explicitly enabled the one requested one :

#![deny(clippy::self_named_module_files)]

Result :

    Checking hyper v1.0.0-dev.0 (/home/runner/work/hyperclip/hyperclip)
    Finished dev [unoptimized + debuginfo] target(s) in 53.42s

If we have a violation , we would see :

error: could not compile `hyper` due to previous error

[seanmonstar]

I was thinking we could add an args: -- -Aclippy::all -Wclippy::specific_lint_name, and thus not need to change the actual source code (like the fmt job above).

[mastrzyz]

Right that's an option but might not be 100% intuitive to someone why their CI Build is failing when locally it isn't?

[mastrzyz]

@seanmonstar ^^ My concern is something will fail in CI and not locally .

[mastrzyz]

I have reverted the change to the main lib file , verified the build is green ! -> https://github.com/mastrzyz/hyperclip/actions/runs/3110424028/jobs/5041594746

[dswij]

My concern is something will fail in CI and not locally .

I agree that changing the top level module would prevent this, but since we're allowing everything except self_named_module_files, this should be fine.

Although now that I think about it, changing the top level module might not be a bad idea at all 😄

[seanmonstar]

Hm, I wonder, does this actually need to be -D and not -W? I suppose just a warning won't stop the CI job from exiting successfully. (I believe there's one instance in the repo that should be failing this lint, hence why I wrote the issue, but seems like the job is happy.)

[mastrzyz]

yeah it's warning but the step is still green here.

[seanmonstar]

Shouldn't we make it -D? We want the CI job to fail, not just yell and then get forgotten.

(I would expect this to trigger a complaint about the ext.rs module.)

[mastrzyz]

Sorry I misread -> Updated.

I verified that -D does in fact "Disallow"

But for the clippy rule we decided on enabling, there isn't any violation reported in ext.rs :/

[dswij]

We can avoid adding these lines directly into src/lib.rs with an argument when we call clippy in CI, as in seanmonstar's comment

[mastrzyz]

Done!

[mastrzyz]

Right that's an option but might not be 100% intuitive to someone why their CI Build is failing when locally it isn't?

[mastrzyz]

@seanmonstar ^^ My concern is something will fail in CI and not locally .

[mastrzyz]

I have reverted the change to the main lib file , verified the build is green ! -> https://github.com/mastrzyz/hyperclip/actions/runs/3110424028/jobs/5041594746