Commit Graph
3 Commits
Author SHA1 Message Date
glm-5.3-flash 65fb69db61 feat: local-socket-halves — the local feature (TCP/UDP/unix halves functions)
- src/local/mod.rs (feature = "local" -> tokio/net): dial_tcp (5s
  timeout), connect_udp (ephemeral bind + connect + the FRAMED UdpHalf
  adapter), dial_unix (OQ-TN-14 in v1), bind_tcp +
  TcpListenerHalves::accept_loop (the listen shape's assembly half),
  local_dial() (one DialFn covering all three substrates).
- UdpHalf truncation fails loud (OQ-TN-13): poll_read recvs into a
  scratch buffer first and size-checks against the caller's buffer —
  an over-size datagram is io::ErrorKind::InvalidData, never a silent
  truncation (tokio's poll_recv-into-ReadBuf would truncate).
- No socket type appears outside src/local/ (convention 16); the
  default crate stays wasm-clean; stdio bridging is NOT here (alktty
  owns process stdio).
- Tests (tests/local_halves.rs, 7, feature-gated): TCP/UDP/unix dial
  round-trips through the real establisher path, the 1400-byte MTU
  datagram, the truncation fail-loud probe, dial-refusal ->
  dial_failed, listen producer over a real TCP listener end-to-end.

Verified: cargo test green (default + --all-features, 44 + 7),
clippy -D warnings (default + all-features + wasm32), fmt clean,
cargo check --target wasm32-unknown-unknown passes.
2026-09-08 09:44:23 +00:00
glm-5.3-flash 3a447273a8 docs: resolve OQ-TN-14 — unix in via local, stdio out (alktty owns process stdio)
Split the last open substrate-placement question:

- Unix: ships with the local feature v1 (dial_unix — same halves
  shape as TCP; the wire enum already carried unix per ADR-001;
  the params task's schema list includes all three values)
- Stdio: OUT of scope — a spawned process's stdin/stdout/stderr IS
  alktty's pipe mode (LocalTtyBackend + tokio::process + Stdio::piped,
  alktty tty-local.md): three multiplexed logical streams + the
  exit-code control chunk (alktty ADR-004) + signal forwarding
  (REQ-TTY-02). A stdio bridge here would be alktty's runner mode
  with the terminal stripped out — a strictly worse duplicate that
  also drops the semantics that matter (a byte tunnel has neither
  exit codes nor signals). Remote command execution composes via
  alktty on the same channels substrate.

Updated: open-questions.md OQ-TN-14 (resolved), overview.md feature
gate + deps + OQ summary, producer.md OQ ref, OQ-TN-10 promotion
(#3 split), phase-0-findings + both POC summaries' resolution notes,
params task (schema enum includes unix), local-socket-halves task
(unix ships, stdio does NOT — with the composition rationale),
oq-tn-14-tracker task repurposed (boundary-maintenance: re-opens only
if a consumer needs stdio-without-process-semantics — which would
need its own ADR, or if the alktty/alktunnels boundary needs
sharpening).

Verified: taskgraph valid (12 tasks, no cycles)
2026-09-07 19:44:03 +00:00
glm-5.3-flash 69498b79cc tasks: Phase 2 decomposition — 12-task dependency graph for v1
tasks/architecture/:
- oq-promotion-sync (planning): back-pointers from the phase-0 ledger
  + AGENTS.md to the promoted OQ tracker (the convergence checklist's
  final half)
- oq-tn-14-tracker: the Safe-Exit external-trigger tracker task for
  OQ-TN-14 (unix/stdio placement; [external-trigger, deferred-oq],
  risk trivial, level research per the two-halves rule)

tasks/tunnels/ (the implementation graph, 8 generations):
- crate-init: module skeleton per overview.md's module map
- params: TunnelParams + open-op spec (ADR-001 wire-stable surface)
- wire-codec: frame_datagram/DatagramReader + the 8 POC-pinned test
  families (ADR-003)
- producer-open-op: establisher (dial, plan flow R-01) + pump handler
  (pump_bidi inline R-02) + registration; POC-ported integration tests
- consumer-session: TunnelSession (open/adopt, data planes, teardown
  matrix — ADR-005); generalizes the reverse POC's ReverseTunnel
- producer-listen: the listen establisher + AcceptQueue contract
  (ADR-004 shape 2)
- local-socket-halves: the local feature (TCP/UDP/unix halves
  functions; truncation fail-loud per OQ-TN-13; unix ships per
  OQ-TN-14's lean-yes, stdio deferred)
- review-core-crates: review-injection point before the downstream
  tasks build on the high-risk producer/consumer shapes
- end-to-end-suite: 6 suites / >=20 tests consolidating both POC
  suites against the public API (the spec's executable form)
- review-impl: the phase-gate review (wire/API/conventions/docs sync;
  findings doc per the alkhttp/alkcall house pattern)

Graph verified with taskgraph: 12 tasks valid, no cycles, 8
generations; critical path = oq-promotion-sync -> crate-init ->
params -> wire-codec -> producer-open-op -> consumer-session ->
review-core-crates -> review-impl; risk concentrated in the two
session tasks (both POC-validated); parallel groups available at
generations 1 and 6
2026-09-07 19:07:20 +00:00