mirror of
https://github.com/vitorpamplona/amethyst.git
synced 2026-10-05 19:28:25 +00:00
Claude
c90851610a
Stage 0 of the Marmot resync: get a live reference back, so the later stages are written against real bytes instead of a careful reading of the spec. The vector generator pinned stock crates.io openmls 0.8. MDK builds against erskingardner/openmls with the `extensions-draft` feature, and the whole current Marmot profile is expressed in terms of what that feature adds — app_data_dictionary (0x0006), app_components (0x0001), safe_aad (0x0002), app_data_update (0x0008). Vectors from the published crate cannot reach any of it. Pinned to MDK's exact rev instead. Adds `marmot-profile-gen`, which builds a group the way cgka-engine does: required capabilities of extension 0x0006 plus proposal 0x0008; GroupContext dictionary carrying the required-component list, group profile, admin policy, Nostr routing and lifecycle; per-leaf dictionaries carrying the supported list, an empty safe_aad list and the 104-byte account-identity-proof v2 component; last resort as the empty-data 0x0004 component in the KeyPackage dictionary, not an extension type; PublicMessage handshakes. It emits the Add commit, the Welcome, and exporter KATs for both group-event and the conformance commitment. The identity-proof encoder is hand-rolled from the spec rather than lifted from MDK, and asserts itself against the fixture published in account-identity-proof-v2.md before emitting anything — so if the generator runs at all, the kind-450 canonical serialization, its id, the BIP-340 signature and the component layout are known to match. The interop harness cloned marmot-protocol/whitenoise-rs, which was archived on 2026-08-05 pinned to mdk-core 0.8.0: it was testing us against a frozen MIP-era client, which is part of how the drift went unnoticed. Repointed at marmot-protocol/mdk, building -p wn-cli. Both source patches are dropped — mock-keyring is replaced by MDK's native --secret-store file, and skip-unprocessable-retry targeted a path MDK does not have. The daemon socket is now pinned via wnd --socket rather than guessed from a derived default. The harness changes are read off MDK's DaemonArgs and wn-cli manifest, not off a passing run; building MDK's workspace needs its pinned toolchain and a local relay. A human run of marmot-interop-headless.sh is the acceptance test. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_016kCuA6tc4JQzHPCDd39GHq
2.7 KiB
2.7 KiB
Summary
Test plan
- Ran
./gradlew spotlessApply— repo is formatted - Ran
./gradlew test(or the relevant module's tests) - Manually exercised the change (see notes below)
Notes / screenshots:
Interop suites
- N/A — change can't affect wire bytes / decoded audio / MLS state / DM envelopes
- Marmot / MLS —
cli/tests/marmot/marmot-interop-headless.sh(Marmot / MDKwn) - NIP-17 DM —
cli/tests/dm/dm-interop-headless.sh - Audio rooms manual —
cli/tests/nests/nests-interop.sh(Amethyst ↔ nostrnests.com) - MoQ-lite hang-tier —
:nestsClient:jvmTest -DnestsHangInterop=true - MoQ-lite browser-tier —
:nestsClient:jvmTest -DnestsBrowserInterop=true - QUIC interop-runner —
quic/interop/run-matrix.sh -s {aioquic,picoquic,quic-go,quinn}
AI assistance
- Drafted with AI assistance, manually reviewed and tested
- Written by hand
If "Drafted with AI assistance" is ticked, also read
CONTRIBUTING-WITH-AI.md for the additional
gates that apply to AI-authored PRs.
License
- By submitting this PR, I agree to license my contribution under the MIT license. Any code I did not author personally carries its original license header.