fix(nargo): resolve local dependencies relative to root of depending package

[ax0]

Related issue(s)

Resolves #

Description

When running nargo in a subdirectory of a package whose Nargo.toml file contains a local dependency expressed as a relative path (e.g. pkg = { path = "../../lib" }), the dependency is parsed relative to the subdirectory, resoluting in an error of the form

Error: cannot find src directory in path ../../lib

Location:
    crates/nargo/src/cli/mod.rs:69:5

This PR sets the current directory to the project root before executing nargo commands.

Summary of changes

• A call to std::env::set_current_dir after determining the project root.

Dependency additions / changes

Test additions / changes

Checklist

• I have tested the changes locally.
• I have formatted the changes with Prettier and/or cargo fmt with default settings.
• I have linked this PR to the issue(s) that it resolves.
• I have reviewed the changes on GitHub, line by line.
• I have ensured all changes are covered in the description.

Documentation needs

• This PR requires documentation updates when merged.

Additional context

BEGIN_COMMIT_OVERRIDE
fix(nargo): resolve local dependencies relative to root of depending package
END_COMMIT_OVERRIDE

[TomAFrench]

Thanks for flagging up this edge case, I'll make sure to add some tests to run commands away from the package root at some point. WRT this PR though, I think of this as just another occurrence of #1013.

My preference for this is that we pass the package root directory (i.e. manifest_path.parent()) into resolver.resolve_manifest so that local paths can be resolved relative to this. We'd avoid magical linkages from code in start_cli fixing issues in the resolver (i.e. we shouldn't rely on current_dir being set correctly) and also support chained local dependencies at the same time for free.

[jfecher]

LGTM.

@TomAFrench what do you think?

[TomAFrench]

LGTM however I haven't tested it locally.

[TomAFrench]

Tested this locally and we're for a dependency chain A->B->C, we're now properly resolving C relative to B rather than A