Sweep unrelated message from unnecessary workspace infromation

[weihanglo]

Resolves #8619

Only pass workspace information when the source is from a local crate installation.

[rust-highfive]

r? @alexcrichton

(rust_highfive has picked a reviewer for you, use r? to override)

[weihanglo]

Before 2018 edition, binaries installation from current working directory is deprecated. In 2018 edition without --path it would throw an error. I'll fix it later maybe by identifying editions.

[ehuss]

Thanks, can you maybe add a test for this?

I don't think it is necessary to do any edition checks (which I think would be near impossible anyways).

[weihanglo]

I don't think it is necessary to do any edition checks (which I think would be near impossible anyways).

Yes. Fortunately I did not choose this solution.

Thanks, can you maybe add a test for this?

Seems that installs_from_cwd_by_default and installs_from_cwd_with_2018_warnings already covers the deprecation message tests, I need some direction and advice for what test case should be added. Thanks!

[ehuss]

I would add a new test that creates a package to install from a registry (use Package::new with a simple main.rs and .publish()). Then create a workspace that will trigger a warning (like a patch in a member manifest as described in the original issue). Call something like p.cargo("check") with with_stderr to verify it emits a warning. Then run p.cargo("install reg") to try to install the package created with Package::new and call .with_stderr to ensure the output doesn't include any warnings.

[weihanglo]

I've add a test. Thanks for your mentoring @ehuss!

[ehuss]

Thanks!
@bors r+

[bors]

📌 Commit a527caa has been approved by ehuss

[bors]

⌛ Testing commit a527caa with merge f110fd9...

[bors]

☀️ Test successful - checks-actions
Approved by: ehuss
Pushing f110fd9 to master...