feat(cli): add --check flag to rustup update for semantic exit codes

[gkamathe]

Summary

This adds a --check flag to rustup update that causes the command to set its exit code based on whether any toolchains were actually updated:

• 0 — no updates were applied; everything was already up to date
• 100 — at least one toolchain was updated
• 1 — an error occurred

Without --check, the behavior is unchanged (exit 0 on success regardless of whether updates occurred). This preserves backward compatibility for existing workflows.

The exit codes align with rustup check (#4681).

Closes #4987

Changes

• Added --check flag to the rustup update subcommand
• Modified update() to track whether any toolchain's UpdateStatus was
  Installed or Updated, and return ExitCode::UPDATES_AVAILABLE (100)
  when --check is set and updates occurred
• Modified update_all_channels() to return whether any channel was updated
  alongside the existing exit code
• Added exit status documentation to update_help() in help.rs
• Added 4 integration tests covering: no updates (exit 0), with updates
  (exit 100), without --check (always exit 0), and all-channels update

Test plan

• cargo fmt --all --check passes
• cargo clippy --all-targets --all-features -- -D warnings passes
• cargo test --features=test passes
• New tests update_check_no_updates, update_check_with_updates,
  update_without_check_always_succeeds, and update_check_all_channels
  all pass

[rami3l]

Hello!

I have reasons to believe that this PR is low-effort and has violated our contributing guidelines (LLM usage is explicitly banned for E-easy unless agreed otherwise):

rustup/doc/dev-guide/src/index.md

Lines 64 to 70 in 60ae6e6

	If you are opening a pull request, you are expected to be able to explain the proposed changes in
	your own words. This includes the pull request body and responses to questions. Make sure you have
	reviewed the PR yourself before submitting it for review to the maintainers. Do not copy responses
	from the AI when replying to questions from maintainers. As an exception, issues marked as `E-easy`
	are meant for new contributors as a learning opportunity; the use of an LLM when submitting PR for
	such issues is disallowed except without explicit permission from the team. Failure to comply may
	result in the PR being closed directly without further notice.

As such, I'm closing it now.