Ignore errors when declaring a registry on path dependency

[w4]

When depending on crate A from crate B, if crate A declares a registry on a path local dependency, cargo would attempt to resolve the registry name from crate B's config.toml, even though it would never be used and likely doesn't exist there.

It may be wishful thinking to open a PR for this in its current form, but it fixes the exact issue outlined in #12271. I'm happy to modify it if there's a more "correct" fix out there.

Environment

CMX2XM in cargo-issue-repro
❯ \cat ws/Cargo.toml
[workspace]
resolver = "2"
members = ["ws1", "ws2"]

CMX2XM in cargo-issue-repro
❯ \cat ws/ws1/Cargo.toml
[package]
name = "ws1"
version = "0.1.0"
edition = "2021"

# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html

[dependencies]

CMX2XM in cargo-issue-repro
❯ \cat ws/ws2/Cargo.toml
[package]
name = "ws2"
version = "0.1.0"
edition = "2021"

# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html

[dependencies]
ws1 = { path = "../ws1", registry = "a" }

CMX2XM in cargo-issue-repro
❯ \cat leaf/Cargo.toml
[package]
name = "leaf"
version = "0.1.0"
edition = "2021"

# See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html

[dependencies]
ws2 = { git = "file:///Users/jordan/Code/cargo-issue-repro/ws" }

Before

❯ CARGO_LOG=trace cargo check
...
2023-10-11T20:47:59.716458Z  INFO cargo::ops::cargo_read_manifest: skipping malformed package found at `/Users/jordan/.cargo/git/checkouts/ws-a3d765f02ebb7e68/6b8487d/ws2`
...
error: no matching package found
searched package name: `ws2`
perhaps you meant:      ws1
location searched: file:///Users/jordan/Code/cargo-issue-repro/ws#6b8487d6
required by package `leaf v0.1.0 (/Users/jordan/Code/cargo-issue-repro/leaf)`

After

❯ /Users/jordan/Code/cargo/target/debug/cargo check
    Finished dev [unoptimized + debuginfo] target(s) in 0.17s

[rustbot]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @weihanglo (or someone else) soon.

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

[weihanglo]

Thanks for being interested in this.

For direct dependencies, if the registry configuration is missing, Cargo errors out. It seems a bit unfortunate that the behavior of transitive and direct dependencies can't be consistent after this changes.

I'll propose this to the Cargo team for discussion. Before that, if you have any more thoughts please comment on the issue instead.

[epage]

btw you link to #12271 but I think you meant #12272

[weihanglo]

Close as the current behaviour is expected. See #12271 (comment) and #12272 (comment) for more.

Do let us know if there is something impossible to be resolved by those approaches mentioned. Thank you.

[weihanglo]

Re-opened for #12272. To move forward, we need some more tests. This is how people usually do :

• The first commit will have tests demonstrating the current bad behavior.
• The following commits will fix the bad behavior and the tests.

That make the intention clear when reviewing or reading the Git commit history.

You might want to add some new tests around this registry_incompatible_with_git test. Feel free to ping me if you have questions.

[bors]

☔ The latest upstream changes (presumably #13080) made this pull request unmergeable. Please resolve the merge conflicts.

[weihanglo]

@rustbot author

[rustbot]

☔ The latest upstream changes (possibly 081d7ba) made this pull request unmergeable. Please resolve the merge conflicts.

[epage]

As this has been waiting-on-author for quiet a while, I'm going to go ahead and close this. Feel free to resurrect these changes!