Fix Non-actionable error messages

[ojuschugh1]

What does this PR try to resolve?

This PR fixes issue #15754 by improving error reporting for workspace publish failures.
Previously, when publishing a workspace, if one package failed (e.g., due to an existing version), Cargo did not clearly list which packages failed and which remained unpublished.
This change ensures that the error message now includes both the failed package and any remaining unpublished packages, making it easier for users to understand what went wrong and what still needs to be published.

---

How to test and review this PR?

• A new test (workspace_publish_failure_reports_remaining_packages) was added to publish.rs to demonstrate the previous behavior and verify the fix.
• The test first fails with the old behavior, then passes after the fix is applied.
• All commits are split for clarity: the test is added first, then the implementation is updated.
• To review:
  1. Check that the new test covers the relevant error scenario.
  2. Review the implementation changes in publish.rs for error context propagation.
  3. Confirm that all tests pass (SNAPSHOTS=overwrite cargo test).
  4. Review commit-by-commit for a clear history.

---

How to test locally:

SNAPSHOTS=overwrite cargo test --test testsuite publish::workspace_publish_failure_reports_remaining_packages

or simply

SNAPSHOTS=overwrite cargo test

or

cargo test

to run the full suite.

---

[ojuschugh1]

What does this PR try to resolve?

This PR fixes issue #15754 by improving error reporting for workspace publish failures. Previously, when publishing a workspace, if one package failed (e.g., due to an existing version), Cargo did not clearly list which packages failed and which remained unpublished. This change ensures that the error message now includes both the failed package and any remaining unpublished packages, making it easier for users to understand what went wrong and what still needs to be published.

How to test and review this PR?

* A new test (`workspace_publish_failure_reports_remaining_packages`) was added to publish.rs to demonstrate the previous behavior and verify the fix.

* The test first fails with the old behavior, then passes after the fix is applied.

* All commits are split for clarity: the test is added first, then the implementation is updated.

* To review:
  
  1. Check that the new test covers the relevant error scenario.
  2. Review the implementation changes in publish.rs for error context propagation.
  3. Confirm that all tests pass (`SNAPSHOTS=overwrite cargo test`).
  4. Review commit-by-commit for a clear history.

How to test locally:

SNAPSHOTS=overwrite cargo test --test testsuite publish::workspace_publish_failure_reports_remaining_packages

or simply

SNAPSHOTS=overwrite cargo test

or

cargo test

to run the full suite.

This PR also solves the feedback received on my old fix -: #15791

New Changes include-: 1) All commits are atomic and passed cargo test.
2) Solved the Port Number issue by falling back to [..].
3) Removed verify step runs BEFORE any packages are published since nothing is published at that point so no need for it.

[rustbot]

r? @epage

rustbot has assigned @epage.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[epage]

Same feedback as #15791 (comment)

This commit should be able to pass as-is but it can't because its including test changes for the following commit

[epage]

refactor: clean up publish error context and remove unused variables

Note that a "refactor" means a change to the code with no user-visible impact. If you wish to use Conventional Commits, this would be better term a fix

Also, the commit title talks about cleanup and removing a variable but the commit is about adding extra context to error messages.

[ojuschugh1]

I will fix this commit title , first i wanted to complete the implementation in this pr, I will open a new pr with proper final fixes and commits. So that the git history should be clean

[epage]

#15791 (comment) is not resolved

Why are we reporting two failed to publish messages? Shouldn't we just put all of this information in the one failed to publish message?

[epage]

From @ojuschugh1 at https://github.com/rust-lang/cargo/pull/15799/files#r2252608902

Hi @epage so regarding this https://github.com/rust-lang/cargo/pull/15799/files#r2252499850 , I found out that already in some places like here there are more than one failed to publish messages so i thought to keep like this.

[epage]

Could you clarify what you mean?

[ojuschugh1]

Hi , @epage We show two lines because Cargo prints the error chain: the top-level is a workspace-specific summary (which crate failed + what’s left), and the “Caused by” line is the transport/server detail. This keeps it readable and preserves single-crate behavior. If you prefer one line, I can collapse it and update tests.

[epage]

Can't we extend the original error to include the context? Seems like that would provide much better results.

[ojuschugh1]

Sure, I will try to do this.

[ojuschugh1]

@epage , I have fixed this. Kindly please review and let me know, if anything else you want me to do.

[weihanglo]

Could you edit the title to make it clearer what this is about at first glance?
(Yeah maintainer can surely do that though, just being nice if you want to edit your stuff)

[ojuschugh1]

Could you edit the title to make it clearer what this is about at first glance? (Yeah maintainer can surely do that though, just being nice if you want to edit your stuff)

Sure no issues