Adds a new lint that checks if there is a semicolon on the last block statement if it returns nothing

[1c3t3a]

changelog: Added a new lint: SEMICOLON_IF_NOTHING_RETURNED
fixes #6467
Adds the SEMICOLON_IF_NOTHING_RETURNED lint and therefore closes #6467.

[rust-highfive]

r? @phansch

(rust-highfive has picked a reviewer for you, use r? to override)

[1c3t3a]

The dogfood tests are currently failing, because the lint fires in several occasions in the clippy codebase. Should we maybe change the lines in the codebase? Otherwise pedantic is maybe the wrong lint level. The lint also fires on for loops as I saw, which is not desired behavior.

[xFrednet]

pedantic should be the correct lint group. The lint is triggered on Clippy because dogfood enables clippy::pedantic on the Clippy code itself.

It can happen that a new lint triggers on Clippy code. The lint implementer usually just updates the code to conform with the new lint. It's basically a self improving machine where each lint also improves Clippy itself. Your observation with if_chain! would for instance be too much and the suggested fix would look weird. @_flip1995 already mentioned a potential solution on Zulip (Sorry for splitting the discussion into two channels. I think we can keep being on Zulip, but I wanted to answer the questions you had here 🙃 ).

---

I suggest that you first try his suggestion with checking the snippet's last character for an }. Then take a look at this false positive where Clippy lints on if expressions:

44 |                 for &ext in &["stdout", "stderr", "fixed"] {
   |                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ help: add `;`: `&["stdout", "stderr", "fixed"];`

Afterwards there should only be a few reports left. Examples like this following one are exactly what the lint is made for and can be fixed by adding a ; at the end. (If the amount is reasonable which it should be).

 error: add `;` to terminate block
  --> clippy_lints/src/verbose_file_reads.rs:48:13
   |
48 | /             span_lint_and_help(
49 | |                 cx,
50 | |                 VERBOSE_FILE_READS,
51 | |                 expr.span,
...  |
54 | |                 "consider using `fs::read_to_string` instead",
55 | |             )
   | |_____________^

---

One more side note: This new lint should be included in the Change log currently you have changelog: none in your PR description. Maybe change it so something like:

changelog: Added a new lint: `SEMICOLON_IF_NOTHING_RETURNED`

This makes it easier for the one writing the change log for the next release 🙃

[1c3t3a]

Oke, so the lint is now working. In an earlier version the lint was (falsely) triggered on those lines, but after resolving the snippet in another way, this did not come while testing. Thanks again to @xFrednet and @flip1995 for the great support!

[phansch]

LGTM, thanks!

@bors r=xFrednet,flip1995,phansch

[bors]

📌 Commit 6b4789d has been approved by xFrednet,flip1995,phansch

[bors]

⌛ Testing commit 6b4789d with merge 001185d...

[bors]

☀️ Test successful - checks-action_dev_test, checks-action_remark_test, checks-action_test
Approved by: xFrednet,flip1995,phansch
Pushing 001185d to master...