[teamleaderleo]

Leo: "But why are there so many tests in Swift? It's Cloud code. Why is it not in Rust?" and "Why are there so many of these tests in the first place?"

What

Step B of the move #32305 started: the NEXT Cloud start's harness login decision is now Rust's, and the Swift copy and its tests go.

• Re-pins CCmuxAppFFI to cmux-app-ffi-2b457e203eac (base's newest FFI release; it carries cmux-cloud-start-core: the Cloud start's harness login decision in Rust over the app FFI (step A) #32305's cloud_harness_login* ops and Hover rules in Rust, step A: the hover card machine, tooltip intent, safe triangle, row card placement and group summary, with their C ABI (no Swift change) #31927's hover ABI, which Swift does not call yet).
• CloudHarnessLoginDecision (CmuxNextCloudCredentials, which now links CCmuxAppFFI) keeps only what the start knows so far. It reads the step, reason, readsMachine, guestTakesLogin, skippedTeamMachine and the record line from cmux-cloud-start-core in one FFI call. loginProvider, the vm-kind probe and its parse are FFI calls too. If the FFI does not answer, the start skips the login with reason ffi-failed, and the record says so, so the skip is never silent.
• CloudStartChat+Login drops its own loginProvider and calls the decision's.
• CloudStartChatLoginTests: its 6 decision tests are deleted; the crate's 10 cargo tests cover every case. One bridge test replaces them. The 3 build-gate tests stay, because the UserDefaults read stays in Swift.

No behavior change: the Rust rules are a 1:1 port, checked case for case in #32305.

Takes back #31305: on purpose. Its decision rules (the team-guest allowance included) now live, unchanged, in Rust (cmux-cloud-start-core, #32305), and this Swift type calls them. The behavior is the same.

Tests

• cmux-ci run --class light --script scripts/ci/run-swift-testing-suites.sh ... CMUX_SWIFT_TEST_SUITES=CloudStartChatLoginTests at a6b1903: 4 tests in 1 suite passed.
• scripts/cmux-next/check-app-ffi-pin.sh --verify-release: the sources match 2b457e2 and the release checksum matches.
• Rust side (from cmux-cloud-start-core: the Cloud start's harness login decision in Rust over the app FFI (step A) #32305): cargo test -p cmux-cloud-start-core gave 10 passed; the FFI cloud_harness_login test gave 4 passed.

Observed / Not run

I ran the suite above and the pin check. I did not run the full CmuxNext suite locally; CI runs it.

Changelog

none

🤖 Generated with Claude Code

Agent session: agent session

[github-actions]

cmux status _{(one comment, edited in place; each line is updated by the job that owns it)}

Review a6b1903: HQ feed card · gallery before/after runs · checks

App of this head (app.zip, kept 1 day)

gh workflow run reload-build.yml -R manaflow-ai/cmux --ref feat-cmux-next -f tag=pr32403 -f ref=lane/cloud-startchat-rust-b -f platform=macos -f backend_mode=local

Shipped in:

• cmux NEXT nightly 1.0.0-nightly.4147272500901 (f4067dc014)
• cmux NEXT nightly 1.0.0-nightly.4147460604401 (87f859ba07)
• cmux NEXT nightly 1.0.0-nightly.4147660705901 (b27c0b3daf)
• cmux NEXT nightly 1.0.0-nightly.4148000752901 (17b75935f0)

[teamleaderleo]

Owned by an agent session · session

[teamleaderleo]

Review (codex, router-review): SHIP at a6b1903

Reviewer: codex gpt-6-astra high, fast tier.

Follow-ups: 1. Each lands as an issue in this repo that links this receipt; the lane that opened the PR fixes it forward.

SHIP at a6b1903

Visual verdict: no images; Leo tries the build

1. [P2] [follow-up] FFI failures are indistinguishable from valid nil results. loginProvider returns nil for both unsupported harnesses and failed calls, while guestKind and guestKindProbe similarly convert failures into normal “no transfer”/“guest unread” behavior (CloudHarnessLoginDecision.swift:89, CloudHarnessLoginDecision.swift:97, CloudHarnessLoginDecision.swift:104). Propagate a distinct FFI failure so the promised ffi-failed record is preserved.

Follow-ups:

• FFI helper failures should preserve the ffi-failed reason.

[cmux-bot]

cmux-bot approves a6b1903: an independent review passed this exact head (router-review receipt by teamleaderleo). A push needs a new review; a later BLOCK at this head dismisses this approval.