There is a soundness issue with atty when building on Windows with a
custom allocator.
This PR switches direct dependencies on atty to is-terminal. New semver
compatible versions of clap and snapbox remove atty. #11417 upgrades
env_logger to remove it from there.
Fixes#11415
This removes one path to `atty`.
Others:
- clap: fixed in 4.0.27
- pretty-env-logger: seanmonstar/pretty-env-logger#52 needs to be resolved first
- snapbox: this will be fixed soonish but is also only a test dependency
- direct dependency
This is part of #11416
Expose libgit2-sys/vendored feature as vendored-libgit2
### What does this PR try to resolve?
Compiling cargo on MacOS can be problematic due present libraries, `cargo` already supports vendoring for openssl (and it helps https://github.com/pacak/cargo-show-asm/pull/51 to some users). This pull request attempts to extend vendoring support to libgit2 in order to fix more foreign library issues: https://github.com/pacak/cargo-show-asm/issues/50
While it is possible for me to enable it on my side (https://github.com/pacak/cargo-show-asm/pull/52) a special care needs to keep versions in sync. Exposing this feature in `cargo` directly will make such fixes for my and other crates more robust.
### How should we test and review this PR?
Disabled by default means existing behavior is unchanged, testing with it enabled to confirm if it still works should do the trick.
Improve errors for TOML fields that support workspace inheritance
Fixes#10997
This also addresses the issue with `MaybeWorkspace<VecStringOrBool>` mentioned in #10942:
```
Caused by:
invalid type: string "foo", expected a boolean or vector of strings for key `package.publish`
```
I removed the `maybe_workspace_vec_string` deserializer in 7a50c0c718 because the error message from the inner `Vec<String>` is now surfaced, but I can revert that if it's preferable to keep those changes.
I tried to base the `deserialize` implementation off of the derived impl for an untagged enum from `cargo expand`. This approach [ultimately led me](https://github.com/serde-rs/serde/blob/v1.0.144/serde/src/private/de.rs#L218) to adding the `serde-value` dependency.
Before, we had hacks to intercept raw arguments and to intercept clap
errors and assume what their intention was to be able to implement our
help system.
This flips it around and makes help like any other subcommand,
simplifying cargo initialization.
Bump git2 to 0.15 and libgit2-sys to 0.14
This will allow cargo to avoid vendored builds of git2 in up-to-date
environments going forward, and brings in the [libgit2 1.4.4 CVE fix].
[libgit2 1.4.4 CVE fix]: https://github.com/libgit2/libgit2/releases/tag/v1.4.4
This will allow cargo to avoid vendored builds of git2 in up-to-date
environments going forward, and brings in the [libgit2 1.4.4 CVE fix].
[libgit2 1.4.4 CVE fix]: https://github.com/libgit2/libgit2/releases/tag/v1.4.4
This reverts commit 8345cf5037. Since that
time, there are now multiple calls to get the number of CPUs, to handle
the `-j -NUM` case, so factor out a helper function.
Bump cargo-util version.
#10546 made a semver-incompatible change to the API of `ProcessBuilder::get_args`. Unfortunately we did not catch that until it was published. This bumps the version of cargo-util to 0.2.1 to accommodate that change. Stable will get version 0.2.0 so that the changes on beta can be released as 0.2.1 in their own time.
cc #10803
Retry command invocation with argfile
### What does this PR try to resolve?
Fixes#10381
The goal is to invoke `rustc` and `rustdoc` with ``@path`` arg file as a fallback.
Since the `-Zcheck-cfg-features`[^1] has already been implemented, in the foreseeable future cargo will probably hit the “**command line too big**” more often on Windows. `rustdoc` and `rustc` both support ``@path`` argfile [^2][^3], so we can leverage the feature for the fallback logic.
The idea behind this PR is that we put the **retry logic** in `ProcessBuilder::exec*` functions and reuse them as much as possible.
### How should we test and review this PR?
Review it commit by commit is recommended.
Argfile fallback is hard to tested with integration tests, so I added some unit tests for `cargo_util::ProcessBuilder` and `cargo::ops::fix::FixArgs`.
If you do want to test the support of argfile manually, a new environment variable `__CARGO_TEST_FORCE_ARGFILE` is added for the sake of forcing cargo to use argfile for rustc invocations. For example, `__CARGO_TEST_FORCE_ARGFILE cargo test --test testsuite -- fix::` is usually a good start to test this feature.
[^1]: https://doc.rust-lang.org/beta/cargo/reference/unstable.html?highlight=check#check-cfg-features
[^2]: https://doc.rust-lang.org/rustc/command-line-arguments.html#path-load-command-line-flags-from-a-path
[^3]: https://doc.rust-lang.org/rustdoc/command-line-arguments.html#path-load-command-line-flags-from-a-path