Silence the warning about forgetting the vendoring

[jneem]

When sparse crates.io is used, dependencies from crates.io always come from ReplacedSources, as the non-sparse crates.io gets replaced by the sparse one. As far as diagnostics go, this case shouldn't "count" as a replaced source.

I had trouble adding a test for this change, because the test harness replaces everything with a dummy registry.

Fixes #12802

[rustbot]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @ehuss (or someone else) some time within the next two weeks.

Please see the contribution instructions for more information. Namely, in order to ensure the minimum review times lag, PR authors and assigned reviewers should ensure that the review label (S-waiting-on-review and S-waiting-on-author) stays updated, invoking these commands when appropriate:

• @rustbot author: the review is finished, PR author should check the comments and take action accordingly
• @rustbot review: the author is ready for a review, this PR will be queued again in the reviewer's queue

[ehuss]

Thanks!

I believe you can use the replace_crates_io() function to replace crates.io without using the typical source replacement. See the test builtin_source_replacement as an example. Can you give that a try and see if it works?

Also, can you update the docstring on Registry::is_replaced to mention that it is false for built-in replacements?

Perhaps @arlosi could give this a second set of eyes to determine if this is the best strategy to fix the issue. I think it should be fine. This method is only used for this one error message. I just have a slight concern that it could cause confusion in the future (returning false for something that is replaced internally).

[jneem]

Thanks for the hint! I managed to get a test working (which fails before this PR, and passes after).

I just have a slight concern that it could cause confusion in the future (returning false for something that is replaced internally).

I was a little worried about this too. I think the usages of ReplacedSource::is_builtin_replacement suggests that it's often necessary to distinguish between builtin and non-builtin replacements. So if it ever becomes necessary to do so on the Source trait, we could replace is_replaced by a method returning a three-value enum (NotReplaced, BuiltinReplaced, OtherReplaced)

[arlosi]

I think this is OK:

• The fact that Cargo is using source replacement to handle sparse registry access to crates.io is an implementation detail.
• Consumers of the Source trait should not see it as a source replacement since the user did not configure source replacement here.

Thanks for this fix!

@bors r+

[bors]

📌 Commit 2f97a3b has been approved by arlosi

It is now in the queue for this repository.

[bors]

⌛ Testing commit 2f97a3b with merge 34a6a87...

[bors]

☀️ Test successful - checks-actions
Approved by: arlosi
Pushing 34a6a87 to master...