Impl warn for locked install without Cargo.lock

[CPerezz]

If we're installing in --locked mode and there's no Cargo.lock published
ie. the bin was published before #7026
the cargo install errors were not stating that it was due to the lack of
the Cargo.lock file. Instead, the error seemed completely unrelated.

Therefore, this tries to address this by adding a warn in the stderr
output.

Closes #9106

I will need some help on the testing side (assuming the code I added for the warning is correct).
It looks to me that the publish function implemented for testing purposes does not publish Cargo.lock which is the actual convention. Should this be updated too? See #7026

[rust-highfive]

r? @alexcrichton

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

[alexcrichton]

Thanks for the PR, but this is a new feature of Cargo which unfortunately isn't following our guidelines on adding new features. Existence of an issue isn't typically enough to warrant a PR to change Cargo, but rather typically discussion is needed first. We've explicitly decided in the past to not warn about this, for example, but there hasn't even been an opportunity to discuss this before this PR was created.

[CPerezz]

Thanks for the PR, but this is a new feature of Cargo which unfortunately isn't following our guidelines on adding new features. Existence of an issue isn't typically enough to warrant a PR to change Cargo, but rather typically discussion is needed first. We've explicitly decided in the past to not warn about this, for example, but there hasn't even been an opportunity to discuss this before this PR was created.

I completely understand. Apologies. I'll comment on the issue whether this decision that you took stills being the same. If that's the case, I'll close the PR. If otherways you consider that it should be closed now, feel free to do so. Thanks!

[dekellum]

Ship it.

[alexcrichton]

Gah sorry I completely misinterpreted this PR. This is a much more minor feature than I thought so this should be fine to land. I agree this is reasonable to land, and it just looks like there's some test fixes that need to happen.

[CPerezz]

@alexcrichton Now that we agree on the landing part, I asked in the PR a question:

I will need some help on the testing side (assuming the code I added for the warning is correct).
It looks to me that the publish function implemented for testing purposes does not publish Cargo.lock which is the actual convention. Should this be updated too? See #7026

This is the help I need in order to understand whether or not the testing has to go on the direction I'm moving towards.

[ehuss]

The Package::publish function doesn't automatically build a Cargo.lock file, and I don't think it should. For tests that specifically need a Cargo.lock, that can be included manually (here is an example).

For testing that cargo publish actually uploads a Cargo.lock, those tests actually call cargo publish.

For this PR, I think you can just take out the call to fs::remove_file.

[alexcrichton]

Apologies again, my reading skills clearly need to improve...

[CPerezz]

No worries at all @alexcrichton !

The PR is ready for review btw! 😃

[alexcrichton]

@bors: r+

Thanks! Sorry again for the confusion @CPerezz :(

[bors]

📌 Commit b526fad has been approved by alexcrichton

[bors]

⌛ Testing commit b526fad with merge e099df2...

[bors]

☀️ Test successful - checks-actions
Approved by: alexcrichton
Pushing e099df2 to master...