fix(hardware): detect Apple Silicon under Rosetta 2 in GPU probe (BE-3435)

[mattmillerai]

ELI-5

If you run an Intel-flavored Python on a modern Apple Mac (via Apple's Rosetta 2 translator), the machine lies and says it's an old Intel chip. Our hardware probe believed the lie and reported "no GPU" — even though the Mac has a perfectly capable Apple GPU. This teaches the probe to ask macOS one extra question ("am I being translated?") so it correctly recognizes the Apple chip underneath and reports its unified-memory GPU.

What & why

comfy_cli/hardware.py:_detect_gpu keyed Apple Silicon unified-memory detection on machine == 'arm64'. An x86_64 Python running under Rosetta 2 on an Apple Silicon Mac reports platform.machine() == 'x86_64', so the unified-memory block was skipped and the machine fell through to the NVIDIA/AMD probes (which find nothing) — a capable GPU reported as gpu: null.

Fix: a small _is_apple_silicon(machine) helper keeps the native arm64 fast path and, when machine != 'arm64', falls back to sysctl -n sysctl.proc_translated (via the existing _run helper). A value of '1' means the process is translated, which only happens on Apple Silicon, so the apple/unified block is used.

Notes for the reviewer

• Chesterton's Fence — this deliberately broadens an existing check. The original machine == 'arm64' guard was introduced (intentionally minimal) in the base PR feat: add cross-platform hardware block to comfy env --json (BE-3399) #562 (BE-3399); this ticket (BE-3435) is the scoped follow-up to broaden it. The old and new conditions differ only in the newly-matched state — machine != 'arm64' and proc_translated == '1', i.e. exactly Apple Silicon under Rosetta. On a genuine Intel Mac the sysctl.proc_translated OID is absent, _run returns None, None == '1' is False, so Intel behavior is unchanged (the existing test_intel_mac_has_no_apple_gpu still passes).
• Never-raise contract preserved: _is_apple_silicon only calls _run (which never raises and is timeout-bounded) plus a string compare, and it runs inside _detect_gpu's existing try/except.
• Under Rosetta, arch is still honestly reported as x86_64 (that's what the process is); only the gpu block is corrected. The model comes from machdep.cpu.brand_string, which returns the real Apple chip name even under translation.
• New unit test mocks _run to return '1' for the sysctl.proc_translated key and asserts the unified block.

Stacking

Opened stacked on matt/be-3399-hardware-block (#562) — hardware.py does not exist on main yet; it lands with #562. GitHub will retarget this PR to main when #562 merges; the diff here is just the net Rosetta change.

Testing

• pytest tests/comfy_cli/test_hardware.py → 20 passed
• ruff check + ruff format --check on the touched files → clean

[coderabbitai]

Warning

Review limit reached

You’ve reached a temporary PR review limit under our Fair Usage Limits Policy.

Your recent review volume is higher than typical usage, so adaptive limits are currently applied.

Next review available in: 54 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Organization UI

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: 3b923c3b-91aa-48f5-8ca5-33f3423a88a4

📥 Commits

Reviewing files that changed from the base of the PR and between 98325c9 and e475dca.

📒 Files selected for processing (2)

• comfy_cli/hardware.py
• tests/comfy_cli/test_hardware.py

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch matt/be-3435-rosetta-gpu-probe

✨ Simplify code

• Create PR with simplified code
• Commit simplified code in branch matt/be-3435-rosetta-gpu-probe

---

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

🔍 Cursor Review — Consolidated panel

Triggered by @mattmillerai.

✅ No high-signal findings.

Panel: 6/8 reviewers contributed findings.

Reviewers that did not contribute: kimi-k2.5:adversarial (empty), kimi-k2.5:edge-case (empty)

[bigcat88]

This PR currently conflicts with main — GitHub reports mergeable: CONFLICTING, so it needs a rebase before I can review it and I'm skipping it in the current review sweep.

Please rebase (or merge main in) and I'll pick it up on the next pass. main moved a fair bit in the last day, including #614 (ANSI sanitisation across the pretty-print call sites) and #628 (the duplicate server_died error-code fix that had main red), so a refresh may also clear unrelated CI noise on this branch.