417847daab
The change of wai/conform-output, on current main, without its wai doctor: main's doctor, its tests, its catalogue entry, its MCP tool and its changelog entry stand as they are, and this change's copies of them are gone, with the lines dev.rs held for them. wai/README.md, which had no line for main's doctor, gains one in its first commands and `toolchain_doctor` in the core tools. - wai conform --format text|json|junit|html, reports made from the signed records. - --independent beside a kind holds the corpus's independent verifier to the same cases; alone, it runs every corpus's verifier, a record each, from one table. - wai-rs/build.rs reads that table from the `# independent:` lines of the step files (.github/wai-conformance/verifiers/), which take the verifiers' command lines out of the steps' bodies. scripts/wai-conformance-lists.py reads and checks the same lines. - wai-conformance.yml runs the union as wai conform --independent --format junit. - wai mcp's conformance tools read the table. - wai-release.yml runs the quickstart from each archive as a new user runs it (clean-machine.sh, its PowerShell twin clean-machine.ps1), held by check-wai-release-quickstart.py. clean-machine.ps1 keeps clean-machine.sh's contents step and its check of WAI Studio. Each leg keeps main's archive_doctor.py check; the step's own wai hello, verify and dev --check are clean-machine.sh's now. - check-wai-ops-doors.py reads VERBS and NEAR_SYNONYMS from catalogue.rs and keeps no list, and the verb-list test reads what that check reads. - The catalogue's export and Studio's actions hold the new operations. Co-Authored-By: Claude Opus 5.5 <[email protected]>
31 lines
2.2 KiB
Shell
31 lines
2.2 KiB
Shell
# step: route record · wai-rs and route_record_verify.py
|
|
# order: 450
|
|
# features: route_record
|
|
# paths: wai/extensions/route-record.md
|
|
# independent: route-record live needs=cryptography python3 -I live-conformance/route_record_verify.py live-conformance/routes
|
|
#
|
|
# wai.route.record (extensions/route-record.md): the reference module's unit tests (a
|
|
# document read strictly and held to the subset, §2's encodings, an estimate past the subset
|
|
# unknown, a null ranked last, the expected figures per successful job and when they are
|
|
# unknown, the outputs' commitment, §8 listing exactly the reference's codes) and
|
|
# tests/route_record.rs (vectors.json regenerates byte for byte from its seeds and no other
|
|
# file is there; wai-rs gives every vector its stated outcome; every reason of §8 and every
|
|
# reused code refuses a vector; the accepted vectors are §10's, of version 1 and of version 2,
|
|
# every limit removing a candidate; a record is bound to each of its documents; `wai route
|
|
# verify` prints the outcome and exits by it); and, in the step of every independent verifier
|
|
# above, the corpus's second implementation,
|
|
# route_record_verify.py, written from the extension's text and sharing no code with wai-rs,
|
|
# which checks the router's signature with WAI's strict Ed25519 rule and the hop receipts
|
|
# with the OER/2 reference verifier (the `cryptography` package, installed above), computes
|
|
# the outputs' commitment with claim-conformance/_blake3.py, recomputes every candidate,
|
|
# removal, expected figure, rank, order, decision and attempt, requires every mutated copy
|
|
# of the corpus to fail, and drops each of its own checks in turn and requires the corpus to
|
|
# catch the weakened verifier; run isolated (-I).
|
|
#
|
|
# The paths this step brings to the workflow's trigger, and why:
|
|
# wai/extensions/route-record.md: tests/route_record.rs and the module's tests require this
|
|
# extension's §8 to list the codes the reference reports, route_record_verify.py reads that
|
|
# table, and the registry's wire identifiers test reads its sections.
|
|
set -euo pipefail
|
|
(cd wai-rs && cargo test --no-default-features --features "$WAI_FEATURES" --lib route_record::)
|
|
(cd wai-rs && cargo test --no-default-features --features "$WAI_FEATURES" --test route_record)
|