fix(mobile): keep ordinary offline outbox retries out of warnings

[rynfar]

When a queued mobile message retries during a normal disconnect, the outbox logs console.warn on every attempt. Route known socket and transient environment failures to an opt-in Pylon outbox debug log while keeping server decisions, HTTP/worker RPC failures, protocol decode defects, and unknown failures visible as warnings. Retry and hold decisions are unchanged; the same distinction applies to attachment uploads, and an edit winning delivery cleanup no longer warns.

Selective adaptation of upstream a493946bb42ab16e1d18285dac3e9e4603651330; its generic T3 debug logger and unrelated cloud/terminal logger refactors are excluded. Set globalThis.__PYLON_OUTBOX_DEBUG__ = true in a JS debugger to inspect ordinary retry details. Sixty-one focused outbox tests passed, including non-socket RPC reason regressions; mobile typecheck and scoped lint passed (one existing effect-dependencies warning), as did diff check. No device interaction was authorized.

Model: GPT-6 Sol (low) via the Codex harness, with GPT-6 root oversight.

---

^{Need help on this PR? Tag @codesmith-bot with what you need. Autofix is disabled.}

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment

Project	Deployment	Actions	Updated
pylon-marketing	Ignored	Preview	Sep 25, 2026 5:33am UTC

[github-actions]

Thread transfer impact

✅ Thread transfer remains within every enforced ceiling.

Provider	Metric	Main baseline	This PR	Impact	PR ceiling
Codex	Total thread wire	14.0 KiB	14.0 KiB	−26 B (−0.2%)	15.1 KiB	✅
Codex	Thread snapshot wire	7.2 KiB	7.2 KiB	+11 B (+0.1%)	7.3 KiB	✅
Codex	Live turn WebSocket wire	6.8 KiB	6.7 KiB	−37 B (−0.5%)	7.8 KiB	✅
Codex	Live turn WebSocket decoded	58.0 KiB	58.0 KiB	−44 B (−0.1%)	66.4 KiB	✅
Codex	Live turn messages	10	9	−1 (−10.0%)	21	✅
Claude	Total thread wire	14.0 KiB	14.0 KiB	−6 B (−0.0%)	15.1 KiB	✅
Claude	Thread snapshot wire	7.2 KiB	7.2 KiB	−8 B (−0.1%)	7.3 KiB	✅
Claude	Live turn WebSocket wire	6.7 KiB	6.7 KiB	+2 B (+0.0%)	7.8 KiB	✅
Claude	Live turn WebSocket decoded	58.9 KiB	58.9 KiB	−44 B (−0.1%)	66.4 KiB	✅
Claude	Live turn messages	10	9	−1 (−10.0%)	21	✅

Baseline: 7ed3f71 · PR result: 0b1c461 · Source CI: success

Scenario and decoded snapshot size

10 historical turns, 5 command tools per turn, 878.9 KiB retained MCP result per historical turn, and a 1.05 MiB retained result in the measured turn.

• Codex decoded thread snapshot: 115.7 KiB
• Claude decoded thread snapshot: 116.3 KiB

Updated in place by a trusted workflow. PR artifacts are strictly validated and never executed.

[rynfar]

Adversarial review found that the shared transport-message helper treats a bare RpcClientError as a connection symptom, so a non-socket RPC failure wrapped as an Error could lose its warning. I narrowed this PR’s logging classifier to explicit socket/connection symptoms and added a wrapped StatusCodeError regression. Reconciled with current pylon; 89 focused outbox tests and mobile typecheck pass. New exact-head CI is running.