feat(protocol): support AdCP 3.1.8 schemas - #992
Merged
Conversation
![github-code-quality[bot]](https://cdn.github.com/u/9919?s=60&v=4){kind=small}
| await asyncio.wait_for(_wait_for_call_count(lambda: call_count, 2), timeout=1.0) | ||
| finally: | ||
| await sync.stop() | ||
| await task |
Contributor
|
The automated review encountered an issue (possibly reached max turns, timed out, or failed to post the final This is an automated message from the Argus AI review workflow. |
bokelley
enabled auto-merge (squash)
July 29, 2026 10:51
KonstantinMirin
added a commit
to KonstantinMirin/adcp-client-python
that referenced
this pull request
Jul 29, 2026
Review follow-up. main moved to 3.1.8 (adcontextprotocol#992) while this branch pinned 3.1.1, so test_manifest_pins_the_spec_version_the_sdk_targets failed once the two met. Only spec_version and source change: the 3.1.1 and 3.1.8 dist/compliance/.../test-vectors/request-signing trees are byte-identical, and regenerating the manifest reproduces all 43 file hashes unchanged. That is the pin doing its job -- it forced an explicit re-vendor decision on a spec bump rather than letting the old set ride silently. Note for the next bump: `python -m tests.conformance.signing.vectors --write` regenerates `files` only and carries spec_version/source forward, so those two fields need editing before the regen, not by it. Refs adcontextprotocol#975.
bokelley
pushed a commit
that referenced
this pull request
Jul 29, 2026
…set and pin it against drift (#980) * test(signing): vendor the complete 3.1.1 request-signing vector set and pin it The vendored request-signing conformance vectors were an incomplete copy of the spec set, and the loaders could not tell. Both graders globbed the vector directory and asserted only that the result was non-empty (test_verifier_vectors.py, test_canonicalization.py), so any non-empty subset passed and a missing vector was indistinguishable from a complete set. Against AdCP 3.1.1 -- the version src/adcp/ADCP_VERSION already targets -- the vendored copy was missing: - 4 of 12 positive vectors (009/010 percent-encoding edges, 011/012 IPv6 authority) - 8 of 28 negative vectors (021-028) - canonicalization.json entirely, despite the vendored README documenting it and devoting a section to it Three further vectors (016, 017, 020) had silently drifted from their upstream contents. 017 is the notable one: it had been vendored at a revision that revoked test-ed25519-2026, the primary signing key, rather than the dedicated test-revoked-2026 keypair the spec now uses -- and keys.json had never been refreshed to carry that key. Changes: - Vendor all 12 missing files plus canonicalization.json, and re-sync the three drifted vectors, keys.json and README.md. The tree is now byte-identical to dist/compliance/3.1.1/test-vectors/request-signing/. - Add tests/conformance/signing/vectors.py: one shared loader whose parametrization is gated on a completeness pin, replacing the per-module glob-and-assert-non-empty helpers. - Add vector_manifest.json, pinning the SHA-256 of every vendored file. It fails on a missing file, an unexpected extra file, and on content drift. A filename-only manifest would not have caught what happened to 016/017/020, so the pin is over content, not names. - Add test_vector_completeness.py, which also ties the pin to get_adcp_spec_version(), so bumping the targeted spec version without re-vendoring fails loudly instead of leaving the SDK graded against a stale contract. - Wire canonicalization.json into test_canonicalization.py. It earns its own grader: it exercises canonicalize_target_uri/canonicalize_authority over 31 cases including 6 malformed-authority rejections that no signed vector covers at all. The newly graded vectors surface 12 real conformance failures, marked strict-xfail with the gap each one tracks so a fix XPASSes and forces the entry to be retired. They fall into two groups: - The verifier has no step-1 strict-parse stage, so malformed or ambiguous input flows on to a later check and is rejected under the wrong error code (negative 021, 022, 023, 026). Conformance grades the code byte-for-byte, so "rejected anyway" is not a pass. - @target-uri canonicalization applies no UTS-46 A-label conversion, drops a trailing empty query, and accepts five malformed authority shapes the spec requires be rejected. adcp.signing ships canonicalize_host(), but canonical.py never calls it. Each group is filed separately; this change deliberately adds no production behavior, only the vectors and the guard that keeps them complete. * test(signing): cite tracking issues in the strict-xfail ledgers * test(signing): avoid private _pytest import in vector parametrization * test(signing): note why _write_manifest replaces files instead of mutating * test(signing): assert the reject code, not just the refusal The reject branch asserted a bare ValueError -- but urlsplit raises exactly that for `https://[::1/p`, so `malformed-ipv6-missing-closing-bracket` passed on someone else's exception while five sibling cases sat in the xfail ledger. Six reject cases ship in the data; only five were recorded as gaps. No case asserted `expected_error_code` at all, so the vector set could not detect its own incompleteness -- the defect this PR exists to close. Capture the exception and assert the code the vector ships. All six reject cases now xfail together under #978 and retire together when the typed error lands. * test(signing): re-pin the vendored vectors at AdCP 3.1.8 Review follow-up. main moved to 3.1.8 (#992) while this branch pinned 3.1.1, so test_manifest_pins_the_spec_version_the_sdk_targets failed once the two met. Only spec_version and source change: the 3.1.1 and 3.1.8 dist/compliance/.../test-vectors/request-signing trees are byte-identical, and regenerating the manifest reproduces all 43 file hashes unchanged. That is the pin doing its job -- it forced an explicit re-vendor decision on a spec bump rather than letting the old set ride silently. Note for the next bump: `python -m tests.conformance.signing.vectors --write` regenerates `files` only and carries spec_version/source forward, so those two fields need editing before the regen, not by it. Refs #975.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Bumps the packaged ADCP spec pin from 3.1.1 to 3.1.8 and regenerates the bundled schema cache, generated Pydantic types, skills, and docs metadata.
Updates compatibility glue for the 3.1.8 list_creatives response split and adjusts canonical registry tests for registry v1.2.0 with 29 mappings.
Also hardens a timing-sensitive RegistrySync retry test so it waits for the observed second poll instead of assuming a fixed sleep is enough.
Validation: make regenerate-schemas, pre-commit hooks, mypy, ruff, black --check, git diff --check, focused regression tests, and full pytest (5893 passed, 41 skipped, 9 deselected, 1 xfailed).