re-baseline the four pending remediation tasks against the ADR-007/008 tree

The tasks were decomposed (d7db6b1) before 49d4432/ac440f3 landed;
both commits touched exactly the areas the tasks reference. Fixes
grounded in verified sources (vendored rustls-acme 0.12.1, rustls-pemfile
2.2.0, rustls-native-certs 0.8.4) and a fresh cargo llvm-cov run:

- acme-event-loop-test: the termination assertion was impossible —
  rustls-acme's Stream for AcmeState never yields None
  (state.rs:407-412, poll_next_infinite + 2^16s backoff); the review's
  U-1 exit-condition premise is withdrawn and corrected in place.
  Replaced with timeout-bounded event collection, a reachable-arm
  inventory (Order warn, AccountCacheStore, Load/Parse error arms,
  DeployedCachedCert/CertCacheStore via deterministic DirCache file
  pre-seeding), and an explicit mark for the full-fake-CA arms.
  server.rs:135 flagged as unreachable dead code (delete or accept).
- coverage-cheap-closes: re-baselined per-line ground truth — original
  groups 1 and 5 are already closed by the handshake suites; group 2's
  TLS 1.3 half is covered, leaving the TLS 1.2 else-arm (client.rs:316);
  new group added for VerifyPresentedCertVerifier::verify_tls12_signature
  (server.rs:349-366, opened by ADR-008; required for the >=98% bar);
  AcceptAnyCertVerifier refs moved to server.rs:468-475 with the stale
  OQ-TLS-09 coordination caveat retired.
- docs-pin-c1-c4-n3-n4: N-4 re-scoped (the mechanism analysis already
  lives in ADR-007 + the resolver doc block; what remains is a short
  server-verifier note covering both verifiers); C-4 updated for
  ADR-007's negotiation-earlier failure point; added the
  FingerprintPinVerifier pop cross-ref update (post-ADR-008 the default
  verifier does verify possession).
- config-validation-and-trivia: added the feature-gate mechanics note
  for the C-3 test (a non-gated test passes vacuously under default
  features); refreshed drifted line refs with a re-grep advisory.
- review 001: Status block records OQ-TLS-09/-10 resolutions; U-1
  carries the termination correction; U-2 carries the supersession
  note. ADR-008 gains the suite-number-to-test-name mapping.

Verification: taskgraph validate 14 tasks; cargo doc --no-deps
warning-free; all edits docs-only (no code paths touched).
This commit is contained in:
2026-09-12 02:42:20 +00:00
parent ac440f3a9a
commit 23893d6236
6 changed files with 299 additions and 121 deletions
+85 -20
View File
@@ -12,17 +12,36 @@ tags: [tests, acme, review-001, u1]
## Description
The ACME event-loop body (src/server.rs:94-133) is unreachable by
The ACME event-loop body (src/server.rs:94-136) is unreachable by
tests: 23 uncovered lines — every `EventOk`/`EventError` arm, the
`debug`/`warn`/`error` mapping, and the "state machine ended" log are
dead code as far as the suite can prove. The `acme_lifecycle` tests
construct the config and assert spawn + ALPN + resolver wiring; the
spawned task runs against a blackhole URL and its events are never
observed. A refactor that drops or mislevels an arm (e.g.
`EventError::Order` warn → error) lands green; the loop's exit
condition (`state.next() == None` → "ACME: state machine ended") is
untested — if upstream changes the stream's termination semantics,
nothing notices.
`debug`/`warn`/`error` mapping. The `acme_lifecycle` tests construct
the config and assert spawn + ALPN + resolver wiring; the spawned task
runs against a blackhole URL and its events are never observed. A
refactor that drops or mislevels an arm (e.g. `EventError::Order` warn
→ error) lands green.
**Correction to this task (2026-09-12, verified against the vendored
rustls-acme 0.12.1 sources) — the review's §U-1 termination premise is
wrong, and so was this task's original work item 3.** `rustls-acme
0.12.1`'s `Stream for AcmeState` never terminates:
`poll_next` is `Poll::Ready(Some(ready!(self.poll_next_infinite(cx))))`
(state.rs:407-412) — it can never yield `None`. Order errors enter an
exponential backoff (`Timer::after(1 << backoff_cnt)`, capped at 2^16
seconds — state.rs:371-375) and retry forever. Two consequences:
1. **There is no "loop ends when the stream ends"** — the
`while let Some(event) = state.next().await` loop in `new_acme`
never exits on its own. Do NOT assert that the `JoinHandle`
resolves; such a test hangs until timeout. The original item 3 is
removed.
2. **`src/server.rs:135` ("ACME: state machine ended") is unreachable
dead code under this dependency.** It cannot be covered by any test
through the real stream. Either delete the line in this task (and
note the removal) or explicitly accept it as uncovered — do not
leave an implementer to burn hours rediscovering that it is
unreachable. (The review's §U-1 framing "if upstream changes the
stream's termination semantics, nothing notices" described a
termination semantics that does not exist.)
The acme feature's only runtime surface is this loop; it deserves one
real integration test.
@@ -33,7 +52,11 @@ real integration test.
directory: a stub HTTP server (tokio, `std::net::TcpListener` on an
ephemeral port — no new deps) that serves a directory JSON with no
usable endpoints, forcing the error path through the *real* event
stream.
stream. (Mechanically verified: rustls-acme's HTTP layer,
`async_web_client` 0.6.3, is plain `async_net::TcpStream` for
`http://` URLs and self-drives via async-io 2's reactor thread —
it works under a tokio runtime, and a plain-HTTP localhost stub is
reachable.)
2. Assert the log events fire. Two shapes, pick one (or both):
- `tracing` test subscriber capturing the `warn!`/`error!` events
(tracing-subscriber with a test layer — add as dev-dependency
@@ -43,16 +66,48 @@ real integration test.
directly (no subscriber needed; cheaper, but the loop body
itself stays uncovered — prefer the subscriber shape if the dev
dep is acceptable).
3. Assert termination: with the fake directory erroring out, the
spawned task's `JoinHandle` resolves (the loop ends when the
stream ends) — pins the exit condition.
4. Keep `TlsError` out of it: per ADR-002/ADR-006, ACME runtime errors
3. **Bound event collection by time, never by stream end**: collect
events under `tokio::time::timeout` (or a bounded channel + a
drop-with-timeout teardown) and assert on what arrived. The
spawned task never resolves its `JoinHandle` — see the correction
above — so any "await the handle" shape hangs. Note: with
`#[tokio::test]` (current-thread runtime) the spawned loop runs on
the test thread, so a thread-local default tracing subscriber is
visible to it; keep the test single-threaded or the subscriber
captures nothing.
4. Which arms are reachable, per rustls-acme 0.12.1's state machine
(state.rs `poll_next_infinite` / `order`):
- `EventError::Order` (warn) — fires against the erroring stub.
- `EventOk::AccountCacheStore` — the first event, generated before
any network I/O (the account key is generated and stored
eagerly, state.rs:377-395); fires even against a dead address.
- `EventError::AccountCacheLoad` / `CertCacheLoad` /
`CachedCertParse` (error) and `EventOk::DeployedCachedCert` /
`CertCacheStore` (ok) — reached by pre-seeding the `DirCache`
directory. The cache file names are deterministic
(`caches/dir.rs:38-53`): `cached_account_{base64url(SHA256(
contact-els…directory_url))}` and `cached_cert_{base64url(
SHA256(domain-els…directory_url))}`. A corrupt pre-seeded file
gives the Load/Parse error arms; a valid PEM (an ECDSA-P256
chain, PKCS#8 key first — `parse_cert` wants
key-then-cert-chain, state.rs:192-214) gives the deployed/stored
ok arms. The PEM can be generated with rcgen in-test (rcgen is
already a dependency on the acme feature path via the server's
self-signed helper — it is a [dependencies] member).
- `EventOk::DeployedNewCert` / `EventError::NewCertParse`
reachable only through a *successful* order, i.e. a full fake
ACME CA (newNonce/newAccount/newOrder/authorizations/finalize/
certificate endpoints). Out of scope at this effort level; mark
explicitly not-covered rather than silent.
5. Keep `TlsError` out of it: per ADR-002/ADR-006, ACME runtime errors
are stream events, not error variants — the test asserts events,
never a `TlsError`.
## Verification
- [ ] server.rs 94-133 covered under `cargo llvm-cov --all-features`
- [ ] server.rs 94-136 covered under `cargo llvm-cov --all-features`
(line 135 either deleted or explicitly accepted as uncovered,
per work item 2's second consequence)
- [ ] The test performs no real network I/O (binds localhost only)
- [ ] `cargo test --features acme`, `--all-features` green; default
build unaffected
@@ -60,17 +115,27 @@ real integration test.
## Acceptance Criteria
- [ ] Every `EventOk`/`EventError` arm is executed by a test
- [ ] The "state machine ended" termination path is asserted
- [ ] Every reachable arm per work item 4 is executed by a test
(`Order` warn, `AccountCacheStore`, the Load/Parse error arms,
`DeployedCachedCert`/`CertCacheStore`); `DeployedNewCert`/
`NewCertParse` are explicitly marked not-covered (full fake CA
needed — out of scope)
- [ ] No test awaits the event loop's `JoinHandle` (it never resolves)
- [ ] The ACME feature's runtime surface is no longer
refactor-fragile
## References
- docs/reviews/001-implementation-review.md §U-1, Part C
- src/server.rs:61-139 (`new_acme` + the spawned loop)
- docs/reviews/001-implementation-review.md §U-1, Part C (§U-1's
termination premise is corrected — see this task's Description)
- src/server.rs:64-145 (`new_acme` + the spawned loop)
- tests/acme_lifecycle.rs (the existing construction-level tests)
- ADR-006 (acme feature layout), ADR-002 (`TlsError` scope boundary)
- rustls-acme 0.12.1 sources (vendored):
`state.rs:407-412` (the never-ending `Stream` impl),
`state.rs:371-375` (order-error backoff + retry),
`state.rs:192-214` (`parse_cert` — key first, then ≥2 PEMs),
`caches/dir.rs:38-53` (the deterministic cache file names)
## Notes
+13 -4
View File
@@ -16,7 +16,7 @@ Four small findings in one pass (all verified in the decomposition
session):
1. **C-2**`acme-tls/1` is appended unconditionally
(src/server.rs:87-89); a caller who already includes it in `alpns`
(src/server.rs:91); a caller who already includes it in `alpns`
gets it twice (probed: `alpn_protocols == [acme-tls/1, acme-tls/1]`).
Harmless to rustls-acme's challenge dispatch today, but the
duplication leaks into the wire config and the ACME/non-ACME
@@ -32,7 +32,13 @@ session):
Err(TlsError::AcmeConfig("empty domain list".into())) }`) at
construction is additive and cheap. Note: empty `contact` stays
legal (RFC 8555 §7.3 allows zero-contact accounts) — do NOT
validate that.
validate that. **Mechanics note (2026-09-12 re-check): the new
test must be `#[cfg(feature = "acme")]`-gated like the existing
ACME tests — under default features the Acme arm already errors
with `AcmeConfig` for the *missing-feature* reason (pinned by
`new_acme_identity_without_feature_returns_config_error`,
server.rs), so a non-gated empty-domains test would pass
vacuously and prove nothing.**
3. **N-6**`tasks/*.md` and `docs/architecture/**` ship in the
published package (`cargo package --list` re-verified). One-line
`exclude` addition: `"tasks/`, `"docs/architecture/` per the
@@ -80,10 +86,13 @@ session):
## References
- docs/reviews/001-implementation-review.md §C-2, §C-3, §N-6, §N-7
- src/server.rs:87-89 (ALPN append), src/server.rs:62-76
- src/server.rs:91 (the ALPN append), src/server.rs:64-145
(`new_acme` — where the domains validation goes),
src/identity.rs:64-66 (`AcmeDirectory::Custom`), Cargo.toml:11
src/identity.rs:65 (`AcmeDirectory::Custom`), Cargo.toml:11
(exclude list)
- (line refs re-checked 2026-09-12 against the post-ADR-007/008 tree;
the file has shifted since the decomposition — re-grep before
editing rather than trusting these absolutely)
## Notes
+73 -41
View File
@@ -1,6 +1,6 @@
---
id: coverage-cheap-closes
name: Cheap coverage closes — resolver resolve() calls, non-Ed25519 pin arm, PEM parse-error arm, fallback seam (U-2)
name: Cheap coverage closes — non-Ed25519 pin arms, PEM parse-error arm, fallback seam, escape-hatch methods (U-2, re-baselined)
status: pending
depends_on: []
scope: narrow
@@ -12,56 +12,84 @@ tags: [tests, coverage, review-001, u2]
## Description
The coverage inventory (review 001 Part C, re-verified in the
decomposition session — llvm-cov matches every listed range) shows
several small, cheap-to-close gaps. One pass over unit + integration
tests closes them all. Per group:
The coverage inventory (review 001 Part C) — **re-baselined
2026-09-12 by a fresh `cargo llvm-cov --all-features` run** after
ADR-007/ADR-008 landed (commits `49d4432`/`ac440f3`): the new
handshake and impersonation suites already closed two of the original
groups, and ADR-008's new verifier opened one new gap. Current
per-line ground truth (all line numbers re-verified against the
current tree):
1. **client.rs 183-189, 212-218**`RawKeyClientCertResolver::resolve`
and `NoClientCertResolver::resolve` are never *called* by a test.
One-line tests: `resolve` returns `Some(key)` for the raw-key
resolver and `None` for `NoClientCertResolver`.
2. **client.rs 297, 316-322**`FingerprintPinVerifier`'s
non-Ed25519 TLS 1.3 signature arm
(`rustls::crypto::verify_tls13_signature`) is only covered by the
deleted review probes. Pinned test: an ECDSA-P256 rcgen cert +
`dss_with_scheme` (P256/SHA256), assert ok, then a forged sig →
err. Mirrors the existing Ed25519 routing pin
(`fingerprint_pin_verifier_routes_ed25519_spki_tls13_signature_through_raw_key_path`).
3. **pem.rs 30**the `Err(e) => Err(io::Error::other(e))` arm of
1. **client.rs:316** — `FingerprintPinVerifier::
verify_tls12_signature`'s non-Ed25519 else-arm
(`rustls::crypto::verify_tls12_signature`) is never executed. (The
original group-2 text pointed at the TLS 1.3 twin at 297/316-322 —
that TLS 1.3 arm is now covered by the handshake suite's pin
suites; only the TLS 1.2 else-arm remains.) Pinned test: an
ECDSA-P256 rcgen cert + `dss_with_scheme` (P256/SHA256) against
`verify_tls12_signature`, assert ok, then a forged sig → err.
Mirrors the existing Ed25519 TLS 1.3 routing pin
(`fingerprint_pin_verifier_routes_ed25519_spki_tls13_signature_
through_raw_key_path`, client.rs:472).
2. **server.rs:349-366** — `VerifyPresentedCertVerifier::
verify_tls12_signature`'s body (ADR-008's default verifier; did
not exist at decomposition time) is entirely uncovered: no TLS 1.2
handshake exists in the suite. Same pop-routing shape as the pin
verifier (Ed25519-SPKI → `verify_tls13_signature_with_raw_key`,
else `verify_tls12_signature`). A direct unit-call test (Ed25519
SPKI + valid sig → ok; wrong-key or mismatched-message dss → err)
closes it. **This group must be included or the ≥98% acceptance
bar is unreachable while the ACME loop stays open** (owned by
`acme-event-loop-test`).
3. **pem.rs:30** — the `Err(e) => Err(io::Error::other(e))` arm of
`load_private_key` (a *parse* failure, distinct from "no key
found") has no test: garbage-but-keyed file
(`b"-----BEGIN PRIVATE KEY-----\n!!!\n-----END PRIVATE KEY-----\n"`)
exercises it.
4. **client.rs 144-147** — the webpki-roots fallback *push* loop is
exercises it. Verified against rustls-pemfile 2.2.0: malformed
base64 in a keyed section is a real `Error::Base64Decoding`-class
`Err` from `private_key` — the prescribed input works.
4. **client.rs:145-148** — the webpki-roots fallback *push* loop is
covered only nondeterministically (passes vacuously when the
platform store is non-empty). Deterministic remediation: a
`#[cfg(test)]`-visible helper taking the "native certs" as a
parameter (or an injectable `load_native_certs` seam) so the
empty-platform case is testable without root. This is a
load-bearing invariant whose fallback branch has no deterministic
test.
5. **server.rs 341-346**`RawKeyCertResolver::resolve` is never
called (only `only_raw_public_keys()` is). A one-line assert
(`resolve(hello).is_some()`) covers it — a `ClientHello` can be
synthesized via `rustls::server::test_client_hello`-style helpers;
the end-to-end raw-key handshake in `handshake-tests` also covers
it, but the one-liner keeps this task independent of that one.
6. **server.rs 287-303**`AcceptAnyCertVerifier`'s two
signature-assertion methods are never called by a test. Two-line
test each (call, assert `Ok`), making the no-pop posture explicit
in the suite. Note: this pins the *current* behavior; if
OQ-TLS-09's resolution changes the verifier, these tests change
with it (coordinate with `fix-accept-any-cert-verifier-posture`).
7. **fingerprint.rs 67** — the second disjunct
test. (`rustls_native_certs::CertificateResult` has public `certs`
/ `errors` fields, so the seam shape is a plain refactor — see
Work item 2.) The `tracing::warn!` for native-cert errors at
client.rs:139 sits in the same uncovered pocket.
5. **server.rs:468-475** — `AcceptAnyCertVerifier` (now the
*escape-hatch* verifier post-ADR-008, moved to server.rs:429) is
only partially covered by the impersonation suite's escape-hatch
tests: `verify_tls12_signature`'s body and
`supported_verify_schemes()` (468-475) are never called. Two-line
test each (call, assert `Ok` / assert the nine-scheme list),
making the no-pop posture explicit in the suite. (The original
group-6 range 287-303 is now `VerifyPresentedCertVerifier`'s code
— the file shifted under the ADR-008 commit.) Note: the original
"coordinate with OQ-TLS-09's resolution" caveat is stale —
OQ-TLS-09 is resolved (ADR-008), the escape-hatch posture is
permanent and pinned by `tests/impersonation_posture.rs`.
6. **fingerprint.rs:67** — the second disjunct
(`len() != 33 || [0] != 0x00`) matrix case (34-byte bit-string +
unused-bits ≠ 0 vs len ≠ 33). Cosmetic.
**Already closed by the ADR-007/ADR-008 suites (do not redo):** the
client resolvers' `resolve` calls (original group 1 —
`RawKeyClientCertResolver::resolve` and `NoClientCertResolver::resolve`
are exercised by the handshake suites) and `RawKeyCertResolver::resolve`
(original group 5 — exercised end-to-end by
`tests/handshake_behavior.rs`, e.g.
`raw_key_client_presents_spki_and_server_extracts_fingerprint`).
`FingerprintPinVerifier::verify_tls13_signature`'s non-Ed25519 arm
(original group 2's TLS 1.3 half) is also covered.
## Work
1. Write the unit tests (in-module `#[cfg(test)]` where the items are
private-visible, tests/ where public API suffices).
2. Group (4) is the only one touching non-test code: extract a
2. Item 4 is the only one touching non-test code: extract a
test-visible seam for the fallback loop. Keep the seam
`#[cfg(test)]`-visible or behind a plain `pub(crate)` fn — do not
grow the public API.
@@ -69,24 +97,28 @@ tests closes them all. Per group:
## Verification
- [ ] `cargo llvm-cov --all-features` shows the seven groups covered
(client.rs 183-189/212-218/297/316-322, pem.rs 30, server.rs
287-303/341-346, fingerprint.rs 67)
- [ ] `cargo llvm-cov --all-features` shows the six groups covered
(client.rs 316/145-148, pem.rs 30, server.rs 349-366/468-475,
fingerprint.rs 67)
- [ ] The fallback test deterministically exercises the push loop
(platform-store-independent)
- [ ] `cargo test`, `cargo test --all-features`, clippy, fmt green
## Acceptance Criteria
- [ ] Line coverage ≥ 98% (from 95.32%) with every load-bearing
uncovered group closed
- [ ] Line coverage ≥ 98% (from 95.81% currently; 95.32% at
decomposition) with every load-bearing uncovered group closed
- [ ] No public-API growth
## References
- docs/reviews/001-implementation-review.md §U-2, Part C (the inventory)
- docs/reviews/001-implementation-review.md §U-2, Part C (the
inventory rows for the closed groups are stale — this task's
Description is the current baseline)
- src/client.rs, src/pem.rs, src/server.rs, src/fingerprint.rs
- tasks/handshake-tests.md (the overlap note for group 5)
- tests/handshake_behavior.rs (the suites that closed original groups
1 and 5), tests/impersonation_posture.rs (the escape-hatch pins —
group 5's `verify_tls13_signature` coverage comes from here)
## Notes
+84 -56
View File
@@ -25,7 +25,7 @@ question):
tighter validity is ever wanted, `not_before`/`not_after` are
additive params.
2. **C-4**`RemoteIdentity::fingerprint` accepts any string and the
pin comparison is exact-string (src/client.rs:272): uppercase-hex
pin comparison is exact-string (src/client.rs:293): uppercase-hex
pins, `sha256:` lowercase prefix, and cross-prefix pins
(`SHA256:`-of-an-ed25519-remote) construct fine and reject at
handshake — fail-closed, never a downgrade, but a config-author
@@ -33,7 +33,19 @@ question):
`FingerprintPinVerifier`): pins must be produced by
`fingerprint_from_cert_der` — case- and format-exact
(`ed25519:<lowercase hex>` / `SHA256:<lowercase hex>`); malformed
pins fail closed at handshake.
pins fail closed at handshake. **Update for ADR-007 (commit
`49d4432`): the pin format is no longer only a verification
selector — it also selects the cert-type offer (`ed25519:` → offer
`[RawPublicKey]`, `SHA256:` → default X.509 offer), and a
cross-format mismatch now fails earlier, at cert-type negotiation,
before the pin compare is even reached (pinned by
`tests/handshake_behavior.rs`:
`ed25519_pin_against_x509_server_fails_closed_at_negotiation`).
The review's original C-4 mechanism ("rejected at pin comparison")
is superseded — same fail-closed verdict, earlier failure point.
Write the doc line from that chain: mismatched *format* fails at
negotiation (a config error, ADR-007); same-format-but-wrong pins
(case, garbage) construct and fail at the pin compare.**
3. **N-3**`TlsServerConfig` is not `Clone` even under
`default = []` where the struct has no `JoinHandle` field. The
all-features posture is what the API freeze pins (conditional
@@ -41,53 +53,59 @@ question):
configurations). Decision: keep non-Clone for v1; record the
reasoning in the type's doc (one sentence) and close the question
— revisit only with a concrete consumer demanding it.
4. **N-4**the cert-type negotiation interaction needs an explicit
doc note on `AcceptAnyCertVerifier`. **Read the corrected mechanism
first** — the review's §N-4 original parenthetical ("alknet's own
client resolver offers both types, which is why production works")
is inaccurate; the review's §Status block and §N-4 now carry the
correction (found during the S-1 remediation, commit `e86b8ba`).
The accurate chain, from the rustls sources (0.23.41 AND 0.23.44):
- rustls sends `client_certificate_types = [RawPublicKey]` (only)
iff the *resolver's* `only_raw_public_keys()` is true
(client/hs.rs:331-334 in 0.23.41, :355-358 in 0.23.44) — no
"offers both types" behavior exists in either version.
- Against `AcceptAnyCertVerifier`
(`requires_raw_public_keys() == false`), a client offering only
`[RawPublicKey]` hits the `(false, true, false)` arm of
`process_cert_type_extension` (server/hs.rs:274-288 in 0.23.44)
`IncorrectCertificateTypeExtension` → handshake failure.
- A client with an X.509-typed resolver (or no resolver) offers no
cert-type extension / `[X509]` → the `(false, _, true)` or
`(false, false, false)` arm → default X.509 → works, and the
server passes presented bytes through unparsed to a custom
`verify_client_cert`.
- So `requires_raw_public_keys()` must stay `false` (do not "fix"
it to `true` — that would break X.509 clients), and the doc note
must state that a raw-key *client* resolver makes rustls offer
`[RawPublicKey]` only, which this verifier rejects
(fail-closed, not a downgrade). Note the S-1 doc on the same
type (landed, commit `e86b8ba`) already carries the no-pop
section; add the negotiation note as a sibling section, and note
that a raw-key *server* presentation (`RawKeyCertResolver`,
`only_raw_public_keys() == true` on the server cert side) is a
different knob (the `server_certificate_types` extension) that
is unaffected. Without the note, the next reader may "fix" the
verifier to `true` and break X.509 clients, or mis-diagnose the
raw-only-client failure as a bug.
4. **N-4****re-scoped 2026-09-12: the mechanism analysis this item
originally asked for is already documented.** ADR-007
(`docs/architecture/decisions/007-cert-type-negotiation.md`,
commit `49d4432`) records the whole negotiation chain, and
`RawKeyClientCertResolver`'s doc block (src/client.rs:152-171)
carries the client-side mechanism note (`only_raw_public_keys() ==
false` → X.509 offer; the `(false, true, false)` arm reasoning).
Do not re-derive or duplicate that analysis. What remains is a
short server-verifier note (a sibling section in the
`AcceptAnyCertVerifier` and `VerifyPresentedCertVerifier` docs, or
one shared paragraph referenced from both):
- `requires_raw_public_keys()` stays `false` on **both** server
verifiers — do not "fix" it to `true` (that would reject X.509
clients; the request-but-don't-require shape accepts both cert
types).
- Post-ADR-007, the crate's own `RawKeyClientCertResolver` presents
the SPKI under the **X.509 offer** unconditionally
(`client.rs:187`, `raw_public_keys: false`), so a raw-key client
against this crate's servers never sends a raw-only offer. The
`IncorrectCertificateTypeExtension` rejection of a raw-only
client offer (`(false, true, false)` arm) can now only arise
from a *foreign* rustls resolver that sets
`only_raw_public_keys() == true` — state it as an interop
boundary of the request-not-require shape, fail-closed, not a
downgrade.
- A raw-key *server* presentation (`RawKeyCertResolver`,
`only_raw_public_keys() == true` on the server-cert side) is the
other knob (`server_certificate_types`) and is unaffected.
Without the note, the next reader may "fix" the verifier to `true`
and break X.509 clients, or mis-diagnose a foreign raw-only client
failure as a bug.
## Work
1. C-1 doc lines (`generate_self_signed_cert` + `SelfSignedCert`).
2. C-4 doc lines (`RemoteIdentity::fingerprint`,
`FingerprintPinVerifier` — format-exactness + fail-closed).
`FingerprintPinVerifier` — format-exactness + fail-closed, with
the ADR-007 negotiation-earlier failure point per the Description).
3. N-3: the one-sentence decision note on `TlsServerConfig`'s doc;
close N-3 in the review's finding list (no code change).
4. N-4: the negotiation note on `AcceptAnyCertVerifier`'s doc. If
desired, add a pinned unit test asserting
`requires_raw_public_keys() == false` stays default (it's a trait
default — the test just documents the choice).
5. Cross-check the review's finding numbering so each doc change
4. N-4: the short server-verifier negotiation note per the re-scoped
item 4 in the Description (a sibling section on
`AcceptAnyCertVerifier` + `VerifyPresentedCertVerifier`, or one
shared paragraph referenced from both). If desired, add a pinned
unit test asserting `requires_raw_public_keys() == false` stays
default (it's a trait default — the test just documents the
choice).
5. While in these docs: `FingerprintPinVerifier`'s pop cross-ref
(src/client.rs:253-255) currently says only "the **server-side**
`AcceptAnyCertVerifier` does not [verify possession]" — post-ADR-008
the *default* server verifier `VerifyPresentedCertVerifier` does;
mention both (the default verifies, the escape hatch does not).
6. Cross-check the review's finding numbering so each doc change
cites its finding.
## Verification
@@ -102,24 +120,34 @@ question):
- [ ] A config author reading the rustdoc cannot mis-case a pin, pin
the wrong prefix, or expect a never-expiring dev cert to expire
- [ ] The raw-only-client negotiation trap is documented before a
consumer hits it
- [ ] The raw-only-client interop boundary is documented before a
consumer hits it (short note per the re-scope; the full
mechanism lives in ADR-007)
## References
- docs/reviews/001-implementation-review.md §C-1, §C-4, §N-3, §N-4
(§N-4 — read the correction block; the original parenthetical is
inaccurate), and the §Status block (the correction summary)
- src/server.rs:226-256, src/client.rs:225-236, src/credentials.rs:35-39,
src/identity.rs (SelfSigned doc)
- rustls 0.23.44 `server/hs.rs` `process_cert_type_extension`
(server/hs.rs:263-288 — the `(requires_raw_keys, offers_raw,
offers_x509)` negotiation table N-4 documents) and
`client/hs.rs:350-358` (the `[RawPublicKey]`-only offer rule) —
both pinned in rustls 0.23.41 too (client/hs.rs:326-334)
- tests/impersonation_posture.rs (the S-1 probe — its fixed resolver
pins `only_raw_public_keys() == false` for the X.509-typed offer;
do not confuse that with a raw-key client resolver)
(§N-4 — the correction blocks are historical context now; ADR-007
is the authoritative record of the negotiation mechanism), and the
§Status block
- src/server.rs:223-257 (the self-signed helper + `SelfSignedCert`),
src/server.rs:18-24 (`TlsServerConfig` — where the N-3 note goes),
src/client.rs:253-255 (the pop cross-ref), src/credentials.rs:36-38
(`RemoteIdentity::fingerprint`), src/identity.rs (SelfSigned doc)
- ADR-007 (`docs/architecture/decisions/007-cert-type-negotiation.md`)
— the authoritative negotiation record; ADR-008 (the verifier
default change that makes `VerifyPresentedCertVerifier` part of the
N-4 note)
- rustls 0.23.44 `server/hs.rs::process_cert_type_extension` (the
negotiation table) and `client/hs.rs` (the `[RawPublicKey]`-only
offer rule) — consult only if extending the analysis beyond what
ADR-007 records
- tests/impersonation_posture.rs (its fixed resolver pins
`only_raw_public_keys() == false` — post-ADR-007 that is also the
crate resolver's unconditional shape),
tests/handshake_behavior.rs
(`ed25519_pin_against_x509_server_fails_closed_at_negotiation`
the negotiation fail-closed pin)
## Notes