Closes Issue4903

[xiongmao86]

fixes #4903.

Check list:

• Followed lint naming conventions
• Added passing UI tests (including committed .stderr file)
• cargo test passes locally
• Executed ./util/dev update_lints
• Added lint documentation
• Run ./util/dev fmt

---

changelog: implement lint that warn about single component path imports(use ident;).

[bors]

☔ The latest upstream changes (presumably #4945) made this pull request unmergeable. Please resolve the merge conflicts.

[flip1995]

Should we allow pub use ident; statements? Can this be useful? Maybe if you want to reexport a dependency?

[xiongmao86]

I never think of reexport case. Reexport should be allowed. I will update the lint logic.

[flip1995]

Allowing pub imports is as easy as

rust-clippy/clippy_lints/src/wildcard_imports.rs

Lines 73 to 75 in 719526f

	if item.vis.node.is_pub() || item.vis.node.is_pub_restricted() {
	return;
	}

[xiongmao86]

I figure out item.vis.node. is_pub(), what case does item.vis.node.is_pub_restricted() apply to?

[xiongmao86]

The latest nightly doesn't have is_pub_restricted()? I update a couple minutes ago. Wait to see what ci results.

[flip1995]

The latest nightly doesn't have is_pub_restricted()

Oh, just realized, that only rustc_hir::VisibilityKind has this function. is_pub_restricted are things like pub(crate), pub(super), crate, pub(my_module), ...

[flip1995]

But I guess pub(crate) krate is also useless, since you can access krate directly anyway. So allowing is_pub should be enough.

[flip1995]

LGTM. Can you squash some commitsß After that, this should be ready to merge.

[xiongmao86]

@flip1995, I notice you have being emphasizing squash commits, I am not very sure what makes commits looks good. Are there anything I can read?

[flip1995]

Oh I just meant that you squash the commits on this branch together with git rebase -i upstream/master¹. For example the two commits 6db5b20 and 0861f0b can be squashed together, since the first commit adds extern crate .. to the test file and the second one removes it again.

What I normally do is 1 commit for the lint implementation/fix, 1 commit for everything tests, and maybe some additional commits for unrelated things (like ./util/dev {fmt,update_lints}, refactors, ...)

---

¹: This opens your editor where you can pick commits you want to keep and squash commits you want to include in the commit on the line above.

pick fe07ad7a Small refactor of mutable_debug_assertions
squash 8d93e826 Don't trigger `debug_assert_with_mut_call` on `.await`
pick ac3ca610 Add test for `await` in `debug_assert!(..)`

for example would include the second commit in the first commit and keep the 1st and 3rd commit.

[flip1995]

I guess the easiest thing you could do would be

git reset upstream/master
git add clippy_lints/src/single_component_path_imports.rs
git commit -m "..."
git add tests
git commit -m "..."
git add clippy_lints/src/methods clippy_lints/src/types.rs
git commit -m "Run .util/dev fmt"
git add -u
git commit -m "Update lints"

Before you push, do a sanity check with git diff origin/issue4903, which shouldn't have any output

[xiongmao86]

Squashed. Thank you for helping me with this pr. @flip1995.

[xiongmao86]

Oh the history didn't disappear?

[flip1995]

Squashed. Thank you for helping me with this pr. @flip1995.

Helping recurring contributors is always a pleasure :)

Oh the history didn't disappear?

GitHub quirk. You have to reload the page 😉

@bors r+

[bors]

📌 Commit ac9f019 has been approved by flip1995

[bors]

⌛ Testing commit ac9f019 with merge f69835b...

[bors]

☀️ Test successful - checks-travis, status-appveyor
Approved by: flip1995
Pushing f69835b to master...