[Astro-Han]

Summary

maka desktop fails when the launcher is started through a symlink, which is the usual way to put a source build on PATH:

$ ln -s "$PWD/target/release/maka" ~/.local/bin/maka
$ maka desktop
/Users/me/.local/bin/Maka.app/Contents/MacOS/maka-desktop-shell: No such file or directory (os error 2)

The launcher looks for the shell beside std::env::current_exe(), and on macOS that is the path it was started through, not the installed executable. This PR resolves the launcher path first, as deployment already does when it compares the running executable with the saved deployment. The shell is then found beside the real maka, and MAKA_LAUNCHER passes the same resolved path to the shell, which uses it to start the Root's next Host.

Verification

• cargo nextest run -p maka-cli -p maka-desktop-shell --test integration -E 'test(/^desktop::/)': 7 passed. maka-desktop-shell is selected so the build unifies its runtime-shaders feature; -p maka-cli alone needs Xcode's metal compiler.
• The new desktop_launched_through_a_symlink_runs_the_shell_beside_its_target fails without the fix with the error above.
• cargo fmt --all -- --check passes. cargo clippy -p maka-cli -p maka-desktop-shell --all-targets --no-deps -- -D warnings passes. I did not run just check.
• Manual check on macOS arm64: with a release build, maka desktop started through ~/.local/bin/maka opens the Maka window, and stopping the launcher reaps the shell.
• scripts/asf-license-headers.mjs, scripts/rust/dependencies.mjs and scripts/rust/notices.mjs checks pass.

Unrelated CI fix

a9aae72, pushed directly to rust-preview, moved the Cua dependency to a new revision and added sysinfo, ntapi and the windows 0.57 crates without regenerating the license inventories, so the dependency inventory check fails for every pull request against rust-preview. The second commit regenerates DEPENDENCIES.rust.tsv, THIRD_PARTY_NOTICES.txt and notices.json; no new license texts are added.

AI use

Select exactly one:

• No generative tool made a substantive contribution
• Generative tooling made a substantive contribution

Tool(s) and scope: Claude Code diagnosed the failure, wrote the fix and the regression test, and ran the checks above.

Checklist

• Tests cover the change and fail without it
• Lint, format, typecheck and the affected suites pass locally

Does this PR entail a change in behavior?

• Yes — described under Summary above
• No

[M4n5ter]

The path resolution fixes symlinked desktop launches with a minimal change. The regression test exercises the actual launcher and reuses the existing fixture. No blocking issues found.