Improve resolver message to include dependency requirements

[weihanglo]

Resolves #6199.

Thanks for previous efforts: #5452, #6374, #6665, which are great but somehow outdated, so I tweak them and create this PR. This will also be obsolete if we ship pubgrub-rs with cargo in the future 😃 But before that happens, IMO these changes are still helpful.

---

This PR changes the resolver error message from

cargo/tests/testsuite/build.rs

Lines 1104 to 1106 in 216f915

	previously selected package `bad v1.0.0`
	... which is depended on by `baz v0.1.0`
	... which is depended on by `foo v0.0.1 ([..])`

to

cargo/tests/testsuite/build.rs

Lines 1104 to 1106 in 0afd40b

	previously selected package `bad v1.0.0`
	... which satisfies dependency `bad = \"=1.0.0\"` of package `baz v0.1.0`
	... which satisfies dependency `baz = \"^0.1.0\"` of package `foo v0.0.1 ([..])`

Also provide different message for different source kinds, such like:

cargo/tests/testsuite/build.rs

Lines 2810 to 2812 in 0afd40b

	package `a v0.0.1 ([CWD]/a)`
	... which satisfies path dependency `a` of package `foo v0.0.1 ([CWD])`
	... which satisfies path dependency `foo` of package `a v0.0.1 ([..])`",

TODO?

From #5452 (comment), there shall be at least one task left behind:

3. Special case pind by a lock file and not a "=1.1.2" in a dependency. Also add a "note: try cargo update" to the end.

In this PR, validate_links also faces this issue that a dependency requirement is locked into a precise version =0.1.0.

cargo/tests/testsuite/build_script.rs

Lines 1002 to 1004 in a5f8bc9

	package `bar v0.1.0`
	... which satisfies dependency `bar = \"=0.1.0\"` of package `foo v0.1.0 ([..]foo)`
	links to native library `a`

I am uncertain about how to resolve this. Besides the functionvalidate_links, is this problem really a thing that may happen? If not, since validate_links only handles old validation logic, it may be ok to drop the commit a5f8bc9 and leave it as is.

[rust-highfive]

r? @alexcrichton

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

[weihanglo]

r? @Eh2406
I guess you would be interested in resolver stuff but no pressure. 🙂

[Eh2406]

Thank you so much! This is very exciting!

Handling lock files looking like "=1.1.2" can be left for a follow up. Or we can flag all "=" reqs with something like "maybe from a lockfile" for now. (I do have a plan for a long term fix, if you want to take it on.)

Why is validate_links getting =? But overall it should be rare, and I don't much care if it is confusing. We can leave it with less info or have a consistent look with confusing reqs, whichever.

[weihanglo]

Why is validate_links getting =?

Because it is called after ops::resolve_ws_with_opts? But yeah I guess the way to address the issue is similar to addressing lockfile. Anyway, if the current status of this PR is acceptable, then I think it's ready for review. Thanks!

[Eh2406]

@bors r+

Thank you!

[bors]

📌 Commit a5f8bc9 has been approved by Eh2406

[bors]

⌛ Testing commit a5f8bc9 with merge 77a0379...

[bors]

☀️ Test successful - checks-actions
Approved by: Eh2406
Pushing 77a0379 to master...