Split command implementations and shared support out of uv

[zaniebot]

Command implementations and shared project workflows are concentrated in uv. Move every command family into a dedicated *-command crate and give shared operations, project logic, CLI settings, output, and process helpers separate owners. uv retains initialization, dispatch, logging, and the central error-hint registry.

Selected dependencies:

uv (initialization and dispatch)
|
+-- uv-*-command (25 command crates)
|   |
|   +-- uv-project ----------> uv-operations
|   `-- uv-operations              |
|                                 `-- existing domain crates
+-- uv-cli-settings ---------> uv-cli-arguments
`-- uv-cli-arguments              (formerly uv-cli)

Shared command support:
  uv-cli-types       exit status and command-selection types
  uv-cli-output      terminal output and progress reporters
  uv-cli-report      installation reports
  uv-cli-error       command error classification
  uv-child-process   process execution and environment files

Command crates:
  uv-pip-command          uv pip ...
  uv-lock-command         uv lock
  uv-sync-command         uv sync
  uv-add-command          uv add
  uv-remove-command       uv remove
  uv-run-command          uv run
  uv-init-command         uv init
  uv-export-command       uv export
  uv-tree-command         uv tree
  uv-format-command       uv format
  uv-check-command        uv check
  uv-upgrade-command      uv upgrade
  uv-version-command      uv version
  uv-audit-command        uv audit
  uv-workspace-command    uv workspace ...
  uv-python-command       uv python ...
  uv-tool-command         uv tool ... and uvx
  uv-auth-command         uv auth ...
  uv-cache-command        uv cache ...
  uv-build-command        uv build and build-backend entrypoints
  uv-publish-command      uv publish
  uv-venv-command         uv venv
  uv-help-command         uv help
  uv-self-update-command  uv self update (optional)
  uv-pylock-command       shared pylock command support

Callers import directly from owning modules. Existing domain crates retain their responsibilities, and the split keeps command behavior, error classification, and application metadata intact. self-update remains optional.

[codspeed]

Merging this PR will improve performance by ×9.2

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 11 improved benchmarks
⏩ 78 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	unzip_wheel_many_files	371 ms	12.4 ms	×30
⚡	install_wheel_many_files	385 ms	13.6 ms	×28
⚡	prepare_wheel_many_files	383.4 ms	15 ms	×25
⚡	unpack_sdist_many_files	317.2 ms	36.8 ms	×8.6
⚡	resolve_warm_jupyter_universal	87.6 ms	11.6 ms	×7.6
⚡	resolve_warm_jupyter	17.4 ms	2.5 ms	×7.1
⚡	discover_workspace_from_all_members_with_excludes	52.7 ms	8.2 ms	×6.4
⚡	discover_workspace_from_all_members	50.5 ms	7.9 ms	×6.4
⚡	run_python_version_synthetic_workspace	268.4 ms	42.5 ms	×6.3
⚡	resolve_warm_airflow	233.2 ms	41.6 ms	×5.6
⚡	hash_sha256	1,165.6 µs	404 µs	×2.9

Tip

Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing zb/split-uv-crate (8283b76) with main (b7c8c62)²}

Footnotes

1. 78 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

2. No successful run was found on main (1f85ea1) during the generation of this report, so b7c8c62 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[astral-automations-bot]

[P3] Remove the superseded command and support crates

The new command crates copy implementations while leaving uv-pip-commands, uv-python-commands, uv-tool-commands, and uv-workspace-commands intact. Cargo metadata shows those four crates have no consumers, and their old support crates form a dependency graph unreachable from uv. Because workspace membership uses crates/*, workspace checks still maintain both implementations. Remove the superseded crates and their workspace dependency entries so each workflow has one owner.

[astral-automations-bot]

[P3] Keep snapshot dependencies limited to test builds

insta is used only inside the #[cfg(test)] module in installation.rs and is already declared under [dev-dependencies]. Adding it here unnecessarily brings snapshot infrastructure into ordinary uv builds. The new uv-publish-command manifest also moves its test-only insta dependency into production dependencies. Keep both declarations under [dev-dependencies].

[astral-automations-bot]

[P3] Avoid duplicating the complete build entry point

build_impl is private and has only this caller; it still performs CLI reporting and error rendering. The wrapper forwards the entire argument list and converts a new two-variant BuildResult directly into ExitStatus, so every future build option must now be maintained across two signatures and another forwarding call without establishing a separate reusable operation. Keep the implementation in build_frontend and return ExitStatus directly.

[astral-automations-bot]

[P3] Retain derived serialization for project version output

Replacing ProjectVersionInfo's Serialize implementation with separately serialized fields and a handwritten JSON object creates a second definition of the output schema. Field additions or removals must now be synchronized with this format string, although the previous serde_json::to_string_pretty(&final_version) already produced the output covered by version_get_json. Restore the derive and serialize the struct directly in the new crate.