From alkblobs (paused on S-01; its ADR-014 is the reviewed downstream
ruling):
- S-01's option-1 sketch (hand the OpenHandler the PublishStream) is
the wrong shape for marked ops — it imports ADR-046's unmarked
call.published semantics into the ADR-047 §2 marker case. The
settled shape: initiator allocates (odd) and writes immediately;
responder adopts (channel id crosses the open invocation, reserved
input key + dispatch-layer strip recommended); wrapper as responder
of record (await the handler's JoinHandle, merge establishment
reply fields + handler result into the single response); zero
call.published on marked ops (nonempty publish chunks rejected
loudly); relay PubTypedOpen lifted.
- Seven items the amendment-4 publish must pin (incl.
publish_schema refused at registration for marked ops; the put
acks-in-result sentence; the client-side stream-then-open
capability absent from ChannelClient today).
- Verification gates list; S-02/S-03/S-04 downstream
non-requirements and remedy preferences; per-crate impact for the
four published consumers (none register Pub-typed openables —
S-01 is capability-additive; semver risk concentrates in any
stop-advertising removals).
Docs-only change.
- version 0.8.0 -> 0.8.1 (bug-fix release; semver-checks 196 pass vs 0.8.0)
- CHANGELOG 0.8.1: the manager_routing-found duplicate adopt/open fix,
the fuzz harness, and the verification summary
- publish exclude gains docs/research/ (internal research notes;
fuzz/ was already excluded — package verified clean of both)
- fuzzing.md §7.9: standing no-hosted-CI policy — corpus replay is the
release-verification fuzz gate; campaigns manual via the detached
runner; OSS-Fuzz out; no workflow files in this repo
- AGENTS.md: corpus replay added to the verification checklist
- lockfiles: alkcall 0.8.1 (root + fuzz)
Verification: 684 tests; clippy -D warnings (main, wasm, fuzz/shared);
fmt clean; doc clean; wasm32-unknown-unknown check+clippy clean;
semver-checks 196 pass; publish --dry-run 116 files (no fuzz/research/
reviews/sdd/AGENTS in the tarball); fuzz corpus replay 5/5 green
\u00a77.8 campaign results: manager_routing 1.06M execs (exact-counter and
parked-bytes invariants held across every adversarial sequence),
envelope_semantic 2.44M execs (coverage saturated, all event kinds
round-trip), spec_parse 10.8M execs (registry compiled every
attacker-shaped schema or rejected cleanly). All exited 0, zero
artifacts.
Also records the two harness-model defects the fuzzer caught in the
interim (per-receiver EOF-sentinel latch; odd/even Adopt range per
ADR-047 \u00a75). Doc-only + harness fixes; no crate changes.
Found by the manager_routing fuzz target (docs/research/fuzzing.md \u00a77.8):
- ChannelManager::open_channel / adopt_channel used
HashMap::insert(...).is_some() as a collision check \u2014 insert REPLACES
the existing entry, so a duplicate adopt/open installed the new state,
dropped the live channel's demux_sender (spurious EOF to its readers,
subsequently routed chunks lost) and still returned
Err(ChannelExists). Fixed with contains_key pre-check; map untouched
on collision. Regression tests: adopt_channel_duplicate_id_leaves_
live_channel_intact, open_channel_duplicate_id_leaves_live_channel_
intact.
New fuzz targets (\u00a77.4 step 3):
- manager_routing: Arbitrary op sequences over ChannelManager; exact
counter models (parked/dropped must equal the manager's monotonic
counters), parked-bytes bound per \u00a76.2-1, clear_all ledger-vs-map
semantics, drainer-byte reconciliation (lossless routing)
- envelope_semantic: constructors -> serde -> write_frame/read_frame
structural round-trip; event-type constants; call.error parse-back
- spec_parse: OpRegisterRequest::from_json -> rebuild -> registry
registration (attacker schemas compile at register, CF-003)
- fuzz/shared/src/arbitrary_value.rs: bounded Arbitrary for
serde_json::Value; 20 spec_parse + 4 manager_routing seeds
- corpus replay for the new targets in fuzz/shared tests
Verification: cargo test 684 passed (682 + 2 regression); clippy
-D warnings clean (main + fuzz/shared); fmt clean (main + fuzz);
cargo fuzz build clean; 20 s smoke on all three new targets clean
(manager_routing 79k, envelope_semantic 141k, spec_parse 517k runs);
crash input replays clean post-fix
chunk_header: saturated (203.9M execs, ~350k exec/s, 33 fork jobs clean);
envelope_frame: 2119 edges/1664 corpus entries, still growing at budget
end, 7.08M execs, 33 fork jobs clean. No crash/OOM/timeout artifacts in
either 10-min detached campaign; exit 0 both. Grown corpus not merged
(per \u00a77.1 corpus policy). Doc-only change.
- fuzz/ workspace (nightly-pinned via rust-toolchain.toml, excluded from
the main workspace and the published package): chunk_header and
envelope_frame targets per fuzzing.md \u00a77.1
- invariant logic in fuzz/shared (stable-toolchain crate): committed
corpus replay as plain cargo test (quinn CI pattern, \u00a77.2 tier 3)
- envelope target adds FrameError shape-partition asserts, exact
consumption accounting, structural write_frame round-trip, serde
key-contract, and a trailing-byte probe (prefix counts body only)
- chunk_header target adds round-trip identity, TooLarge/short error
shape, is_eof, 8-byte consumption, input-never-mutated
- committed seed corpora: 245 deterministic seeds via
fuzz/gen_fuzz_seeds.py (truncations, len=0/MAX+1/u32::MAX, channel 0,
invalid UTF-8, deep nesting); grown corpora + artifacts gitignored
- fuzz/run-detached.sh: \u00a77.6 detached runner (setsid+nohup+log, fork
mode, rss/malloc limits) \u2014 campaigns never share fate with a session
- fuzz/json.dict; README; research doc \u00a77.7 records step-1 status
Verification: cargo fuzz build clean; stable side green (cargo test 682
passed, clippy -D warnings, fmt --check incl. fuzz/shared); corpus
replay 245 seeds green; detached 10-min campaigns on both targets
completed with zero crashes/OOMs/timeouts
- OOM in a fuzz target must cost the fuzzer, never the opencode host
- three defense layers: soft rss/malloc limits, fork-mode blast radius,
setsid+nohup detachment with log-file polling (fuzz/run-detached.sh)
- no ulimit -v with ASAN; corpus replay + CI flag parity notes
- sequencing and summary updated to make the detached runner mandatory
Review 009's coverage debt in full — the paths the review-008 gates
never walk. No wire or API changes.
- C-1: pin the plain-bundle install-failure arm (un-compilable
input_schema; the as-filed duplicate-name route does not fail
registration) — the install task ends before the dispatch loop,
channel 0 never dispatches. ADR-051 §5 gains the loud-install
coverage note: relay-openable + plain-bundle arms pinned,
generic-ops + bootstrap-discovery arms documented as
best-effort-loud (crate-internal specs compile by construction).
- C-2: the HubLegImports filter — filtered closure path + only
partition unit-tested (errata: only was already pinned at filing);
the empty-stash e2e gate (generic ops + discovery only, dropped
ops resolve NOT_FOUND).
- C-3: the batch-form reserved-reply-key rejection pinned (reason
handler_error, teardown, ledger decrement, no pump spawn).
- C-4: open_channel_with_reply's failure path pinned e2e (the typed
error carries the channel:open_failed code + details reason/message).
- C-5: both byte-identical claims golden-pinned — the no-fields reply
against {"channel_id": 2} and the standard-shape wire payload
against the full 9-key literal.
- C-6: derivation edge shapes pinned — channels//sub, channels//direct,
channels → None; the 4-segment strict superset annotated as the
pre-amendment behavior change (errata: actual is Some("x/sub"), the
multi-segment-ALPN rule, not the as-filed Some("alk/x/sub")).
- C-7: builder overwrite pinned last-win (single + batch) with the
doc sentence on with_reply_field.
Verification: 682 tests pass, clippy -D warnings clean, fmt clean,
doc clean, wasm32 check clean.
File: docs/reviews/009 (resolved; errata marked per finding)
Post-landing audit of the 0.7.1 -> 0.8.0 remediation diff: two
hardening guards, one rejection-posture fix, two log/message
corrections, and the deferred coverage debt filed as review 009.
- RelayPlan owns the producer-leg ChannelManager and reclaims the
adopted spoke channel_id via a Drop guard (replaces the pump
handler's post-pump_bidi explicit reclaim). Closes the leak
windows the pump's normal path cannot reach: the wrapper's
establishment bound expiring after the adopt, and the pump
handler's early-return arms (plan absent, downcast failure,
try_unwrap failure, accept_bi failure). The send-half drop still
EOFs the spoke leg via the mux pump's implicit-EOF sentinel, so
the spoke-side cascade is unchanged. ADR-051 §6 documents the
closed post-adopt window (the pre-adopt §6 window and the
inside-adopt_channel cancellation point stay as documented).
- rebuild_spec_for trims and rejects empty/whitespace
channel_open_alpn strings — an empty explicit string previously
overrode a sane name-derived ALPN.
- op_name_is_standard_channel_open_shape applies the same
empty-segment guard as the derivation: channels//sub no longer
serializes boolean-only and then reconstructs unmarked (silent
stub for a marked op); the explicit string rides instead.
- reserved_reply_key_call_error interpolates RESERVED_REPLY_KEY;
the establisher-bug log fires at warn! (programming error).
- Regression tests: the plan drop guard, the empty-ALPN fallback,
the empty-segment shape check (672 tests, 3 new).
- CHANGELOG [Unreleased] entry for the audit fixes.
- docs/reviews/009 — the audit's deferred test-coverage gaps
(template failure arms, filtered/only, batch reserved key,
wire failure path, golden pins, derivation edge shapes, builder
overwrite semantics), each with the test to add and gates.
Verification: cargo test 672 passed; clippy --all-targets -D
warnings clean; fmt --check clean; doc --no-deps clean; wasm32
check clean.
- src/channels/gate2_tests.rs (ADR-051 gates 6/7, review 008 gate 2):
the full consumer -> hub (HubLegTemplate) -> spoke relay e2e —
the open resolves with the hub-allocated channel_id, `bound`
survives the relay, data flows both directions with a fake 8-byte
chunk header riding verbatim (the hub never parses the data plane,
ADR-034/035), spoke-side close cascades to clean reclaim on both
legs; hub-side disconnect (the consumer's duplex end dropped via a
killable transport proxy) tears down both legs with the ledger
decremented; the mid-establishment window (ADR-051 §6) pinned with
its two reclaim signals — the consumer leg reclaims at its own
transport EOF, the spoke channel (allocated before the establisher
replied) is the honest residual, reclaimed when the spoke-leg
transport ends; the channels/tty/sub standard-shape companion pins
no derivation regression.
- src/channels/relay.rs: the relay's adopted producer-leg channel
entry now reclaims when the relayed pump completes (teardown after
pump_bidi) — the consumer-leg wrapper's teardown cannot see the
producer leg's manager; the RelayPlan carries the spoke id for the
reclaim.
- Release bookkeeping: 0.7.1 -> 0.8.0, the CHANGELOG entry covering
Units 1-3 (Establishment reply projection, open_channel_with_reply,
flavor-form discovery derivation, ChannelRelay, HubLegTemplate,
gate-2 harness); review 008 Status -> Resolved with the U-1/U-2
commit refs and the 955->945 errata note; ADR-051 Status ->
all units landed + the §6 mid-establishment residual expanded to
the two-reclaim-signal shape the gate pins; the stale "0.7.2"
version mentions in ADR-047/049 corrected to 0.8.0 (the units land
unreleased).
Verification: cargo test 669 passed / 0 failed; clippy --all-targets
-- -D warnings clean; fmt --check clean; cargo doc --no-deps clean;
cargo check + clippy on wasm32-unknown-unknown clean;
cargo publish --dry-run --allow-dirty passed.
- src/channels/hub_leg.rs: HubLegImports (the Clone discover/stash,
bundles split by the channel_open marker, with the per-consumer
filter) and HubLegTemplate (the hub-leg install hook: per-connection
fork + generic channel ops + plain bundles as-is + marked specs via
ChannelRelay::register_relay_openable + bootstrap discovery closed
over the fork (review 004 F-06) + serving-identity resolution
(CF-005) + the single-stream dispatch loop).
- The loud assembly posture holds at install: a Pub-typed (or
unregistrable) marked spec ends the leg's install task — channel 0
never dispatches, never a silent stub (ADR-051 §6).
- Assembly rule surfaced by the gate e2e: from_call discovers the
spoke's bootstrap discovery ops like any plain op; the template
skips BOOTSTRAP_DISCOVERY_OPS (new pub const, discovery.rs) when
re-registering plain bundles — the template's own install, closed
over the fork, supersedes the imported copies.
- Gate tests (7): stash split + filter, the full template e2e
(services/list shows the re-exposed ops, a plain imported op
round-trips through the hub, the marked op opens through the relay
with bound surviving, data both directions), serving-identity
precedence (scope-bearing override resolves, scope-less fails
closed), the loud Pub-typed posture, and the filtered-stash subset
run (NOT_FOUND for the filtered-out marked op).
- Docs: ADR-051 status (3a + 3b implemented, 3c remains); review 008
Unit 3b landed note (the bootstrap-op skip rule).
Verification: cargo test (665 passed), clippy --all-targets
-D warnings, fmt --check, doc --no-deps — all clean.
Design session outcome for the relay unit. The hub/spoke family (hub
or spoke may relay; the hub re-exposes spoke services by ACL without
binding ports) needs the call-half support too, so the unit's scope
grew beyond the as-pinned sketch and is now three sub-units.
- ADR-051 (new): the relay is the wrapper composition — translate hop
= the ADR-049 establisher shape, byte-forward hop = pump_bidi per
ADR-050; the relay holds Arc<CallConnection> + ChannelManager per
producer leg (not a ChannelClient — take_call_connection's detach
is wrong for a hub with three CallConnection claimants); the reason
mapping preserves the spoke's code+message (timeout is the one
non-1:1 case, mapping to dial_failed); the registration seam is
two-phase (discover/stash → per-connection fork-register — the
ADR-047 §4 fork mechanism makes a one-call import impossible); the
ADR-042 relay map dissolves (implicit per-channel mapping) and
channel/close needs no translation surface (EOF cascade propagates
with correct ledger accounting on both legs); Pub-typed marked
specs are a loud assembly error; establishment bounds compound per
hop (noted, no fix); the ACL layering note is pinned (the spoke's
AccessControl sees only the hub identity; forwarded_for is never
checked).
- ADR-042 amended: the §Scope note is revised (implementation is an
alkcall export; hub crates compose it) and the two mechanism
supersessions are recorded — the contract and auth-model rationale
unchanged.
- ADR-047 amendment 3's forward reference and the README ADR index
updated (001..051).
- Review 008 remediation plan: Unit 3 split into 3a (ChannelRelay
component + gates), 3b (hub-leg install template — the call-half
support), 3c (gate-2 e2e incl. the mid-establishment disconnect
window); sequencing note updated; adoption note records the
session's decisions.
Verified: cargo doc --no-deps (markdown-only change).
Flavor-form open op ids (channels/tunnel/direct, channels/tunnel/
forwarded — alktunnels ADR-007/008) now survive discovery + the
op/register announce path: the marker reconstructs, so a hub consuming
through discovery wraps the op with relay machinery instead of the
silent plain-forwarding-stub failure.
Per the review-008 plan's combined (b)+(c) shape:
- derive_alpn_from_op_name generalizes from strip-/sub|/pub to
strip-last-segment (rsplit_once) — a strict superset: standard
two-suffix shapes derive identically, multi-segment ALPNs survive
the same way, and the flavor form derives. The boolean marker
remains the gate (consulted only for marked ops in
rebuild_spec_for) — a plain op named channels/tty/query is
unaffected.
- spec_to_json_pub emits channel_open_alpn (the explicit string)
beside the boolean when the op name is not the standard
channels/<seg>/(sub|pub) shape; standard shapes stay byte-identical
to the pre-amendment payload. The advertised operation_spec_schema
documents the optional ["string","null"] property.
- rebuild_spec_for prefers the explicit string, else boolean →
generalized derivation. Both wire consumers (from_call import,
op/register announce) parse through the same parser, so one change
covers both. A channel_open_alpn without the boolean never marks an
op (the boolean is the dispatch hint).
Residual ambiguity pinned in the ADR: an ALPN segment colliding with
a flavor name (channels/x/direct → alk/x vs alk/x/direct) is
undecidable from the name alone; the explicit string is the
disambiguator.
Tests (review gates 1 + 3; gate 2 lands with the Unit 3 relay):
- channels/tunnel/direct reconstructs WITH channel_open = alk/tunnel,
both via the explicit string and via the boolean alone (the
deployment-skew case: old producer, new consumer)
- standard shapes (channels/tty/sub, multi-segment
channels/custom/proto/sub) round-trip byte-stable — no new key
- explicit string overrides a colliding derivation; string without
boolean is ignored
- op/register announced flavor-form spec round-trips with the marker
- derivation unit tests: flavor form, non-op-type suffix not
special-cased, prior shapes unchanged
Docs: ADR-047 amendment 3 — the convention sentence (mirroring
alktunnels ADR-002 Amendment 1: flavors are new op ids, additive, the
.../sub op is never reused), the wire shape, the preference order,
the residual-ambiguity note, one-way-door timing.
Verification: cargo test 649 passed (12 new); clippy -D warnings
clean; cargo fmt --check clean; cargo doc --no-deps clean.
The establisher can now contribute to the open-op success reply — the
establisher → opener direction amendment 2 left empty. First consumer:
alktunnels ADR-008's bind-first listen establisher, whose observed
OS-chosen bound address rides the reply as an additive `bound` field
(the SOCKS5 BIND reply#1 BND.ADDR fidelity ask).
- Establishment gains reply_fields (private, builder-constructed):
with_reply_field / with_reply_fields / reply_fields. The map shape
generalizes beyond `bound` without a fourth amendment; the
#[non_exhaustive] carrier (amendment 2) makes the extension additive
— no construction-site break.
- run_open_wrapper merges contributed fields into the success output
after reserving `channel_id`. The key is wrapper-owned: an
establisher supplying it fails the open loudly
(channel:open_failed, reason handler_error, message naming the
reserved key), tearing the just-allocated channel down — never
shadowing. Absent fields leave the reply byte-identical to the
pre-amendment shape (asserted exactly).
- ChannelClient::open_channel_with_reply returns
(channel_id, reply, send, recv) — the full success output — so
consumers read `bound` first-class; open_channel delegates and
discards the fields, signature unchanged.
Tests (all gates from the review + plan):
- projection produces { channel_id, bound } on the wire; no-fields
replies are byte-identical (establisher and no-establisher shapes)
- reserved-key establisher fails with reason handler_error; channel
torn down, ledger decremented, pump handler never spawned
- registry-level register_openable_with_establisher projects fields
- e2e over a real channels connection: bound reaches
open_channel_with_reply; open_channel unchanged against the same
accept side
- existing establishment tests (establisher-success round-trip,
plan-flow, no-establisher compat) pass unchanged
Docs: ADR-049 amendment 3 (projection, reservation, read path,
compatibility posture, door type).
Verification: cargo test 637 passed; clippy -D warnings clean;
cargo fmt --check clean; cargo doc --no-deps clean.
Both findings verified against HEAD 3b36b40 (0.7.1); all citations
check out (one line drift: U-2 reply construction is operations.rs:945,
errata noted rather than silently edited).
Plan decisions recorded above the as-filed option lists:
- U-1 shape: combined (b)+(c) — explicit channel_open_alpn string for
non-standard shapes; derivation generalized to strip-last-segment,
gated on the boolean marker. As-filed option (a) is unsound (the
derivation runs consumer-side; a registration-threaded allowlist has
no path to rebuild_spec_for).
- U-1 gate 2 escalated to an in-tree relay component (src/channels/
relay.rs), amending ADR-042's downstream-implementation scope note —
the re-produce/relay shape is general (alkhttp fallback hub, alknodes)
and a real consumer now requires it.
- U-2 adds open_channel_with_reply (ergonomic read path for `bound`);
open_channel unchanged.
Units, sequenced 1 -> 2 -> 3, one minor release:
1. U-2 reply projection (ADR-049 am. 3) — lands first; the relay's
establisher projects spoke reply fields through it.
2. U-1 derivation + wire field (ADR-047 am.) — standard shapes
byte-stable; round-trip gates for flavor-form and boolean shapes.
3. Relay component + gate-2 e2e (ADR-042 am.) — establisher = translate
hop (forwarded_for, adopt_channel, reason-code mapping, bound
projection), OpenHandler = inline-await byte-forward hop;
consumer -> hub -> producer e2e asserting bound survives the relay.
As-filed findings retained untouched above the divider (the review-007
pattern).
Verification: docs-only change; no build/lint/test impact.
Filed from the alktunnels graduation spec work (ADRs 007/008):
U-1 extends ADR-047's op-name → ALPN derivation to flavor-form open
op ids (channels/tunnel/direct, channels/tunnel/forwarded) so the
marker survives discovery and hub relay wraps them as channels, not
plain forwarding stubs; U-2 (ADR-049 amendment 3 sketch) gives the
establisher a path to contribute additive reply fields (the bind-first
listen 'bound' address — BND.ADDR fidelity). Both additive; data
plane untouched. alktunnels implementation is sequenced after these
land.
Batch-fixes every remaining open alkcall-side finding from downstream
consumers so 0.7.0 is the only release they need to absorb.
- CF-006 (the CF-005 corollary): run_open_wrapper derives a per-call
AuthContext — the opener's dispatch-resolved identity (the same
identity the ACL gate and cap check saw) overlaid onto the
install-time context — and passes it to both the establisher and
the pump handler. Identity-less calls keep the install-time
identity (no synthetic-anonymous rewrite); transport-truthful
fields are never rewritten. Signatures unchanged — behavior-only;
identical on per-connection registries, hub-forwarded opens now
show the end client. Gates:
open_wrapper_overlays_per_call_identity_on_install_time_auth +
open_wrapper_keeps_install_time_identity_when_call_identityless.
- CF-007 (alkhttp review 006 Part C doc drift): ADR-016 amended to
the eight-code list — ALREADY_EXISTS + CONNECTION_CLOSED in the
Context, §3 table, and from_openapi collision rule; new §2a
documents the undelivered-vs-ambiguous write-failure distinction.
- ChannelPlan type doc now states the Send + Sync payload constraint
(alktunnels POC F-1 re-derived it by compiler error).
- Ledger: CF-006 and CF-007 filed + resolved; CF-005's corollary
note points at CF-006; the Open section is empty. ADR-049 gains
the per-call-identity note. Changelog 0.7.0 covers the batch.
Sweep result (all downstream reviews): no other open alkcall items —
alktunnels W3/F-2 are downstream-by-design, alktty R3/P14 are closed
constraints, alknet has none; OQ-24/37/39/40/41 stay
deferred-by-design (no consumer pull yet).
Verification: cargo test (631) + --all-features (648), clippy
(all-targets, all-features, -D warnings), fmt --check, doc
--no-deps, wasm32 check, semver-checks (no update required),
publish --dry-run.
Verifies and fixes CF-005 (alktunnels reverse-flow POC W1): the
connect-side serving path built channel 0 internally and never set an
identity, so a scope-gated serving op could only be satisfied via the
payload auth_token. Token is now the fallback (hub-forwarding /
browser path); transport/key-based identity is the primary path.
- ServingConfig gains identity: Option<Identity> — the explicit
override (remediation a). Semver-relevant struct-literal change →
0.7.0 (minor bump at 0.x, wire surface unchanged).
- from_connection_with_serving propagates the transport
Connection::identity() to the channel-0 connection via set_identity
before the serving loop starts (remediation b) — mirrors the accept
side's install-hook set_identity; process-local, nothing new on the
wire.
- Dispatch identity precedence on the serving loop: payload
auth_token → identity_provider, then ServingConfig.identity, then
transport identity; identity-less dispatch still fails closed
(FORBIDDEN).
- Public core::auth::NoopIdentityProvider (resolves nothing; the
ServingConfig::default() provider — three private test copies
existed).
- Regression gates: four cf005_* e2e tests (transport propagation,
override precedence, identity-less denial, token fallback +
precedence).
- Ledger CF-005 → resolved; ADR-022 §connect-side-serving amended;
README example updated; changelog 0.7.0.
Verification: cargo test (629) + --all-features (646), clippy
(all-targets, all-features, -D warnings), fmt --check, doc
--no-deps, wasm32 check, semver-checks (no update required at
0.7.0), publish --dry-run.
- architecture README: index rows for ADR-049 (establishment phase) and
ADR-050 (pump_bidi); stale ADR-001..045 range labels corrected
- channel-operations.md: design-decision table + references list the
two new ADRs
- CHANGELOG: missing link references for 0.4.0/0.4.1/0.5.0/0.6.0
- AGENTS.md: ADR range 001..050
Status → Resolved with the three unit commits; both deviations from
the review's sketches (typed-opaque ChannelPlan over Option<Value>;
(u64, u64) over io::Result) recorded up top with their rationale, as
the ADRs carry them.
Extracts the two-pump data-plane helper alknet ADR-078 deferred until
the shapes converged (they have: alktunnels POC pump_halves + alktty's
channels session). Purely additive.
- channels::pump::pump_bidi(channel, peer_read, peer_write) -> (u64, u64):
two joined pumps, shutdown-on-completion per direction; copy counts
for observability. The channel side is a single AsyncRead +
AsyncWrite value (the accept_bi BiStream); the peer side takes split
halves — the establisher's natural dial result (into_split).
- Return (u64, u64), not the review sketch's io::Result<(u64, u64)>:
both pumps swallow copy errors by contract (mid-stream error =
abrupt close, no error channel mid-stream per ADR-049 §6), so an
Err state would be dead code. Deviation recorded in ADR-050.
- alktty's three-pump session does not fit (exit future as a third
signal) and stays as-is, per the review's scope.
- Tests reproduce the POC's two-pump semantics through the helper:
bidirectional flow with exact counts, EOF-from-one-side completes
the other's shutdown (clean EOF at the far end), dead-source =
EOF-shaped teardown.
- ADR-050 records the decision, deviations, and two-way door type.
Verification: cargo test (625 passed, +2), clippy -D warnings, fmt
--check, doc clean, test --all-features clean, wasm32-unknown-unknown
check clean.
Implements ADR-049 amendment 2 — the reserved Establishment payload is
filled, and the OpenHandler lifetime contract is documented.
- Establishment { plan: Option<ChannelPlan> } with ChannelPlan =
Arc<dyn Any + Send + Sync>: typed-opaque, because the payload an
establisher hands the pump handler is a live handle (dialed socket,
TTY handle), not JSON — the review's Option<Value> sketch could not
satisfy its own verification gate. #[non_exhaustive] keeps a future
carrier change from being another break. Construction:
Establishment::new(plan) / Establishment::default().
- OpenHandler gains the plan parameter:
Fn(Value, Option<ChannelPlan>, Connection, AuthContext) ->
JoinHandle<()>. Separate parameter (not merged into input) — a
typed payload cannot ride the JSON input; no schema collision.
Wire surface unchanged: the plan is process-local (establisher ->
wrapper -> handler).
- run_open_wrapper threads establishment.plan to the handler; None
when no establisher is registered. Kills the alktunnels-POC
side-channel handoff (resource-keyed slot + poll loop) whose
concurrent same-resource race is now unreachable — each open's
establisher result flows to its own handler.
- Lifetime contract documented (R-02, doc-only half): the returned
JoinHandle must track the data-plane lifetime — the wrapper awaits
it and its completion triggers teardown; early return = teardown
at birth. Noted on the OpenHandler type docs and both registration
entry points.
- Breaking at 0.6.0 (the point of landing it before alktunnels
Phase 1): Ok(Establishment {}) sites become
Ok(Establishment::default()) mechanically.
Verification: cargo test (621 passed, +4: plan-flows-to-handler,
concurrent same-resource opens get distinct plans, no-establisher
None plan, Establishment construction), clippy -D warnings, fmt
--check, doc clean, test --all-features clean.
Findings filed to prevent a second fix->publish->update-dependents
cycle; each was reached by building working code against 0.5.0:
- R-01 [major] — Establishment is payloadless but the channel plan
is exactly what establishers need to hand to the pump handler
(ADR-049's own 'reserved for a channel plan' note). Costs verified
in two consumers: alktunnels POC side-channel handoff (same-resource
opens race the slot), alktty forced to keep backend allocate
post-open in-band (allocate_failed stays an in-band frame — the
shape ADR-049 eliminates, alive one layer down). Ask: fill the
reserved field (plan: Option<Value>, process-local, wire unchanged)
in a 0.6.0 sweep; the break is mechanical (Establishment::default).
- R-02 [minor] — the OpenHandler JoinHandle lifetime contract is
undocumented and load-bearing: the wrapper's await of the returned
handle IS the teardown trigger; a handler that returns before its
pumps finish tears the channel down at birth (the POC found this
empirically — every tunnel EOF'd instantly). Doc note + ADR-049
amendment; optional debug warning.
- R-03 [minor, optional] — the ADR-078 two-pump helper's convergence
test is satisfied (POC gives both shapes); extracting pump_bidi
now is additive (no break) and pins the contract upstream. Decide
in the same 0.6 sweep.
Non-findings: reverse-flow (-R) needs no upstream mechanism
(from_connection_with_serving + serving-side allocation, verified by
trace); establisher receives registry-validated input; typed
establishment errors complete on the wire; early-arrival cap
unchanged; EstablishmentError reason set sufficient.
Goal stated in the review: 0.6 is the last breaking sweep forced by
known work. alkcall tests: 617 passed (docs-only change; baseline
check).
- E-03: the open wrapper's handler-exit teardown no longer discards
UnknownChannel silently — debug log + benign-race pinning comment
(ledger take is the atomic gate; no double-decrement)
- E-04: consumer-facing doc note on the 64-parked-chunks observable
bound (channel-client.md + EARLY_ARRIVAL_CAP const doc) for
tunnel-style push-first producers
- N-2: ChannelManager::early_arrival_count() accessor (the
observability choice over removing the write-only counter),
documented monotonic, with a park/adopt-drain monotonicity test
- review 006: Unit 3 marked implemented in Status and remediation plan
Verification: 617 tests pass; clippy -D warnings clean (host +
wasm32 check); fmt clean; doc clean
Implements ADR-049 Unit 1 — the open-op wrapper gains an awaited,
bounded establishment phase, and the client stops erasing the error.
- OpenEstablisher hook + Establishment/EstablishmentError types:
register_openable_with_establisher awaits the establisher bounded
(earlier of dispatch deadline and per-registration timeout, else
ESTABLISHMENT_TIMEOUT = 10s) after allocation, before the reply and
before the pump handler is spawned (ADR-049 §1/§2). Implementation
note: the establisher takes (input, auth) only — the channel's
yield-once BiStream belongs exclusively to the pump handler
(amendment recorded in ADR-049).
- Establishment failure: teardown_channel + opener-ledger take +
policy.on_close un-increment (allocation and teardown balance;
the ledger take is the atomic gate, ADR-047 §7), reply
channel:open_failed with details {reason, message} — reason ∈
dial_failed / unknown_resource / resource_shortage / handler_error
/ timeout (ADR-049 §3). SSH contract consumer-visible: a failed
open never returns a channel_id.
- register_openable unchanged (no establisher = always-OK; existing
registrations compile and behave identically — compat gate test).
- ChannelClient::open_channel returns ChannelOpenError (breaking at
0.5.0): CallFailed { error: CallError } carries the wire error
verbatim (establishment_reason() branches on details.reason);
MissingChannelId / AdoptFailed cover the local-only shapes
(ADR-049 §4, review 006 N-1).
- Tests cover all four verification gates from the review: e2e
establisher failure through a real channels connection (typed
reason + no-channel + ledger un-increment), bounded timeout,
no-establisher compat, establisher-success pump round-trip; plus
reason-vocabulary mapping and bound arithmetic.
- Bump to 0.5.0 (open_channel error-type change is semver-relevant).
Verification: cargo test (608 passed), clippy --all-targets -D
warnings, fmt --check, doc --no-deps, wasm32 check — all clean.
Design review from the alktunnels Phase 0 research pass, verified
against tree a22b2b8 (0.4.1). Findings numbered E-01..E-04:
- E-01 [major] — the open op cannot fail after allocation: the
wrapper replies {channel_id} the moment the OpenHandler is spawned;
establishment failures (params-valid-but-rejected, backend lookup
failure, target dial failure) present to the consumer as a
successful open followed by an instant, indistinguishable clean
EOF (implicit-EOF mux path + unified poll_read EOF arms). SSH
semantics (RFC 4254 §5.1 open-failure reply with reason codes;
channel never exists opener-side), SOCKS5 reply codes, and
udpgw's opaque ERR bit (counterexample) surveyed in
alktunnels/docs/research/ssh-socks5-survey.md. alktty's in-band
error-frame mechanism (send_negotiation_error, 0x00-peek) is the
per-crate workaround this upstream establisher obsoletes for the
channels path. Proposed shape: an awaited establishment hook
(OpenEstablisher) or await-and-inspect OpenHandler, tearing down on
failure and replying channel:open_failed with SSH-four reason codes
in ADR-016 details. Remediation sketch + verification gates
included.
- E-02 [minor] — services/list discloses no per-op metadata; OQ-40
(channel/resources/subscribe) becomes load-bearing for the first
time via the alktunnels discovery resolution (OQ-TN-08).
- E-03 [minor] — OpenHandler-exit vs channel/close teardown race is
benign (ledger take is the gate) but the let _ = discard at
operations.rs:509 is silent; recommend log-or-comment.
- E-04 [minor] — early-arrival park cap (64) is an observable bound
for push-first producers under slow adopters; no change requested,
filed so the constraint is visible to the next consumer.
Non-findings recorded: open-op ACL path complete across all three
dispatch entry points; input_schema enforcement covers open params;
EOF arms unified; channel_open marker + resource_id_path wire
round-trip intact; opener-ledger decrement atomic at every call site.
alkcall tests: 597 passed (docs-only change; baseline check).
Surfaced by alkhttp review 006 (UP-03): services/list-peers showed
every peer with an empty operations array. PeerCompositeEnv overrode
peer_ids only, so peer_operations fell to the trait default
(Vec::new()) and the ADR-022 amendment's "announced op is discoverable
via services/list-peers" promise never resolved on the wire. ADR-030
prescribed the fix but it had never been ported into alkcall. The
existing list-peers unit tests passed because they mock
peer_operations with hand-rolled envs.
Implements ADR-030 as specified:
- OperationEnv gains list_operation_names (default Vec::new(),
back-compat for all existing implementors)
- OverlayOperationEnv overrides it with its overlay's registered names
- PeerCompositeEnv::peer_operations delegates to the peer overlay's
list_operation_names; PeerCompositeEnv::list_operation_names
aggregates session + connections + base (mirrors its contains())
- LocalOperationEnv enumerates its registry; ChannelsSessionEnv
delegates to base
Gate: announced_op_is_discoverable_via_services_list_peers in
src/registry/op_register.rs — announces an op through op/register,
then asserts both the direct peer_operations probe and the
services/list-peers wire shape attribute the announced op to the peer,
over the exact compose_root_env shape (PeerCompositeEnv + attached
connection overlay). Verified load-bearing: reverting the
peer_operations override fails the gate.
ADR-030 status Proposed -> Accepted with the UP-03 provenance note.
Verification: 590 default / 607 all-features tests, clippy
(all-targets, all-features, wasm32) clean, fmt clean, doc clean,
semver-checks 196 pass against v0.3.0 (defaulted trait method is
non-breaking).
- resource_id_path rides both halves of the spec wire round-trip:
spec_to_json_pub serializes it (optional string key), rebuild_spec_for
parses it. Additive optional field - absent stays absent. Previously
an announced (or from_call-imported) op declaring ownership-scoped
resource extraction silently rebuilt with resource_id: None, so ACL
checks ran without the resource ID.
- Gates: spec_round_trips_resource_id_path (serialize -> parse ->
field intact) + spec_without_resource_id_path_stays_absent (additive
field breaks no consumer).
- ADR-022 amendment: bootstrap-op set gains services/list-peers with a
dated G-05 note (the installer has registered it since the amendment
landed; the doc lagged the code). Set remains closed at four.
Verification: cargo test 589 / --all-features 606, clippy
(all-targets, all-features, wasm32) clean, fmt clean, doc clean.
Refs docs/reviews/005-...md (G-04, G-05; all findings closed).
- op_register_handler takes the serving registry alongside the
connection and rejects announced names that collide with the serving
side's own registrations (ALREADY_EXISTS regardless of replace).
Peer-announced ops may collide with peer-announced ops (replace
governs, the reconnect path) but never shadow the deployment's own
ops: the connection overlay resolves before base in PeerCompositeEnv,
so an unscreened same-name announce would silently rewrite what a
wire-dispatched handler's ctx.env.invoke resolves. Composition
authority (ADR-018) stays with the deployer.
- ADR-022 amendment (2026-09-04): collision policy recorded in the
2026-09-03 amendment's op/register section (rationale + visibility
irrelevance); status line notes the sub-amendment.
- Gates: base-External collision rejected even with replace (overlay
stays clean, serving registration untouched); Internal base op
equally protected; overlay/overlay collisions still follow replace;
nested composition of a base op resolves the serving side's own op
after an unrelated announce (real compose_root_env env shape).
- PeerCompositeEnv resolution order deliberately unchanged.
Verification: cargo test 587 / --all-features 604, clippy
(all-targets, all-features, wasm32) clean, fmt clean, doc clean.
Refs docs/reviews/005-...md (G-03; Unit 3 open).
- Split dispatch() into dispatch_start() (sync prefix) + spawned
invocation: both single-stream loops (serve_single_stream and the
accept-side run_loop_single_stream) spawn Once invocations, Sub
pumps, and sink response writers; only the Pub sink start stays
inline (chunk_tx must register before the next call.published).
Inline dispatch deadlocked same-connection nested composition: the
read loop awaited the parent handler, which awaited a nested call
whose response only the same read loop could resolve (resolved only
via the 30s sweeper). Spawned handles tracked + aborted at loop exit;
in_flight_sinks behind an Arc<parking_lot::Mutex> with guards dropped
before awaits.
- run_loop_single_stream gains the pending-resolution arms
(RESPONDED/COMPLETED/ERROR): the accept side previously served only
and had no loop resolving its own outbound pendings in single-stream
mode — the latent accept-side imported-op composition hazard is
mechanized shut.
- Write-failure in the spawned Once path warns instead of closing the
loop (matches the Sink arm; dying transport still surfaces via
ConnectionClosed on the next read).
- G-02 gate: hub_handler_composes_peer_announced_op_via_nested_composition
— announce -> consumer calls hub/compose -> hub's serving loop
wire-dispatches it -> handler composes via ctx.env -> forwarding
stub's nested call crosses back to the consumer. The F-05 gate
bypassed this path entirely.
- Interleaved-directions gate: outbound_call_resolves_while_inbound_
subscription_is_being_served — consumer serves a live Sub while a
wire-dispatched hub handler issues an outbound call on the same
connection.
- Both gates verified load-bearing: run against the pre-fix loop each
reproduces the G-01 hang (no progress, bounded-timeout failure);
post-fix both resolve in <0.2s, no sweeper evictions.
Verification: cargo test 583 / --all-features 600, clippy
(all-targets, all-features, wasm32) clean, fmt clean, doc clean.
Refs docs/reviews/005-...md (G-01, G-02; Units 2-3 open).
Post-remediation review of f84d214 (review 004 Units 1-3). Five
findings, verified in source and (for G-01) empirically via a probe
test that was added, run, and removed:
- G-01 [major]: serve_single_stream awaits dispatch inline; a
wire-dispatched handler composing a peer-announced op (or a
from_call import) over the same connection deadlocks — the nested
call resolves only via the 30s sweeper (probe: TIMEOUT at 30.0007s).
- G-02 [major]: the F-05 e2e gate calls the announced op directly,
bypassing the forwarding stub — the one path G-01 breaks.
- G-03 [major]: op/register's collision gate is overlay-only;
PeerCompositeEnv resolves connections before base, so an announced
op can shadow the serving side's own ops in nested composition.
- G-04 [minor]: resource_id_path does not survive the spec wire
round-trip (pre-existing shape, load-bearing for op/register).
- G-05 [minor]: install_bootstrap_discovery registers
services/list-peers; ADR-022's bootstrap set doesn't name it.
Non-findings bound the re-review: fork surface lock discipline,
bootstrap discovery closure, frame-arm equivalence of the composed
loop, unchanged pure-consumer default, alkhttp cross-repo claims,
CJK sweep (none), all gates reproduce (581/598, clippy, fmt, wasm,
doc).
Remediation plan: Unit 1 (concurrent serving loop + stub-exercising
gate) gates Unit 4 downstream; Unit 2 (collision policy); Unit 3
(round-trip completeness + doc alignment).
Verification: cargo doc --no-deps clean; tree unchanged apart from
this review doc.
Focused design-mismatch review found via alkhttp's WS data-channel
drill-down (alkhttp review 003 WS-24/WS-25). The channel machinery is
done and proven; the gaps are the dispatch-resolution mechanism and
the connect-side serving half — both upstream of any transport.
Findings:
- F-01 [major]: top-level dispatch consults only the dispatcher's
base registry — ops registered per the ADR-047 §4 amendment's
overlay mechanism resolve NOT_FOUND on the wire; the only proven
shape (per-session registry as the dispatcher's base) differs from
the ADR's wording
- F-02 [major]: no Clone/fork surface on OperationRegistry or
HandlerRegistration — every inner payload type IS Clone-able
(verified type-by-type, incl. jsonschema::Validator and
Capabilities), so the fork is a small addition
- F-03 [minor]: fork must carry handlers + validators, not just specs
- F-04 [major]: connect-side channel-0 read pump resolves responses
only; inbound call.requested frames are silently dropped — no
serving half on the single-stream shape (ADR-022/AGENTS §8
bidirectionality unreachable from the connect side)
- F-05 [major]: no wire mechanism announces client-side ops; the
six call.* kinds are closed. Resolution candidate: bootstrap op
(op/register) served per-session, handler writes into the
connection-local overlay; discovery rides services/list-peers
- F-06 [minor]: per-session fork must carry bootstrap discovery ops
for per-session openables to be discoverable
Includes a non-findings section (e2e reference shape,
register_openable completeness, channel-id split, envelope-kind
closure, wasm-cleanliness) and a 4-unit plan: ADR decisions (Unit 1)
-> fork surface (Unit 2) -> client serving + bootstrap op (Unit 3)
-> alkhttp wiring downstream (Unit 4, tracked in alkhttp review 003).
Verification: cargo test (565), clippy --all-targets -D warnings,
fmt, doc --no-deps — all clean at c16b069. No source changes.
- CF-004: services_schema_handler now applies the same visibility +
AccessControl gates as invoke() (identity resolution mirrors invoke:
handler_identity under internal). Restricted ops return spec-404 NOT_FOUND
— matches "restricted ops don't exist" and leaks nothing about the
restricted surface. Closes the unauthenticated /call-path disclosure.
- CF-003: publish_schema compiled at registration time (both
OperationRegistry::register and OperationRegistryBuilder::store);
un-compilable schemas are a registration error — an unvalidated ingest
path can no longer be constructed. Compiled validator cached per-op
(publish_validator) and consumed by dispatch; per-request compile gone.
BEHAVIOR CHANGE: register/builder reject un-compilable publish_schema.
- CF-002: demux TooLarge skip streams through a fixed 64 KiB buffer
instead of allocating the peer-declared length (u32, up to ~4 GiB);
cumulative 256 MiB skipped-bytes budget tears down dribbling peers.
Existing resync test passes unchanged.
- CF-001: new retryable CallError::connection_closed (CONNECTION_CLOSED)
applied only where the call is provably undelivered — request-frame
write failures on all consumer paths (call/subscribe/publish, both
stream modes; publish pump tags write stages). Mid-publish failures and
producer-side fail_all stay non-retryable INTERNAL (delivery ambiguous).
New code string is additive; retryable flag is the machine-readable
signal.
Verification: cargo test (558 pass, 15 new), clippy --all-targets -D
warnings, fmt --check, wasm32-unknown-unknown check.
Found via alkhttp Review 002 (PRJ-16): the services/schema handler
does a bare registry.registration(name) with no Visibility and no
AccessControl check, so POST /call (and the MCP call tool) can fetch
any Internal op's complete spec unauthenticated. The GET /schema route
and MCP schema tool enforce the pre-checks; the /call path is the
hole.
Found while fixing alkhttp's HTTP-side instance
(review-001-publish-schema-validation-robust): the identical
warn-and-skip pattern exists at src/protocol/dispatch.rs:352-366 — a
compile failure of a Pub op's publish_schema proceeds with
validator: None, so arbitrary unvalidated JSON reaches the sink
handler over the wire. Suggested direction: fail-closed + per-
registration validator cache (worked shape in alkhttp 1572a9d,
src/gateway/schema_cache.rs) + registration-time schema compilation.
Single compile site verified (both pump arms consume the one
InFlightSink.publish_validator).
Cross-crate finding from alkhttp Review 001 (WS-12), filed here per the
ledger's purpose (consumer-surfaced alkcall findings).
src/channels/adapter.rs:144: the ChunkError::TooLarge arm allocates
vec![0u8; length] from the peer's untrusted 8-byte header before
reading; length is u32, so ~4 GiB can be pinned per connection and held
indefinitely by a dribbling peer. Reachable via the alkhttp WS path by
any authenticated browser. Skip/resync logic is correct; the memory
shape is wrong — stream-skip with a bounded buffer instead.
The normal payload arm (:159) is safe (parse_header bounds it at
MAX_CHUNK_LEN); only the TooLarge arm is unbounded.
- Remove alktype from Cargo.toml (its only usage was a thin wrapper
over jsonschema::options().build())
- Replace alktype::validation::build_validator with direct jsonschema
in dispatch.rs
- Add docs/architecture/chunk-header.bast.json — the chunk header's
BAST (Binary Abstract Syntax Tree) machine-readable wire spec
- Embed as channels::wire::CHUNK_HEADER_BAST via include_str! so
downstream Rust crates can consume it without a file lookup
- Add test asserting the embedded BAST doc is valid JSON and matches
the wire format
- Update AGENTS.md §10 and implementation-specialist.md: BAST docs
are the contract; trivial/hot-path formats stay hand-rolled,
complex formats use the alktype engine or codegen
Verification:
- cargo test: 543 passed, 0 failed
- cargo clippy --all-targets -- -D warnings: clean
- cargo fmt --check: clean
- cargo doc --no-deps: clean
- BAST doc compiles + round-trips against alktype v0.2.0 engine
The R-02 acceptance gate was a line-level grep
(`grep -v 'mod tests'`) that cannot return zero while any test uses
`.expect()` — `grep -v` filters per-line, not per-scope, so every
`.expect(` inside a `#[cfg(test)] mod tests {}` block whose
individual line lacks the literal "mod tests" still matches. The suite
has ~269 such test-only `expect` calls, making the literal gate
unsatisfiable (an impossible-goal spec bug, surfaced during Unit 1
remediation).
Replace both copies of the gate (the R-02 finding gate and the Unit 1
acceptance gate) with a scope-aware verification: either (a) a targeted
grep for the specific removed strings, or (b) file-by-file confirmation
that every remaining `.expect(` is within a `#[cfg(test)] mod tests`
boundary. The intent — no `expect` in library (non-test) code — is
unchanged; only the verification method is corrected.
- docs/architecture/README.md: add Roles and Composition section with
the four roles (producer, consumer, hub, spoke), dependency layering
diagram, protocol crate pattern, and the two-path adapter model
- docs/architecture/channels-overview.md: fix stale alknet ADR numbers
(071/072/073/074/075/076/077/078/079/080/081 -> 034-044), update
relationship section for post-extraction world, update crate
dependencies to reflect single-crate alkcall
- src/lib.rs: add Downstream composition section with role table and
protocol crate pattern, linking to the architecture README
Verification: cargo test (483 passed), cargo clippy (clean),
cargo fmt (clean), cargo doc (no warnings)
glm-5.3-flash
glm-5.2
deepseek-v4-pro