ship/u8 docs m1 gate#5643
Merged
Merged
Conversation
matthewevans
commented
Jul 12, 2026
matthewevans
commented
Member
- docs(parser): describe the Milestone-1 source-addressed document IR (U8)
- fix(tooling): point cargo export-cards at the oracle-gen binary
Update the oracle-parser skill — the parser's single source of truth — to describe the architecture U3a/U3b/U4/U5/U6/U7/U20 actually landed, replacing the pre-remediation category-ordered pipeline description. Finishes the WIP draft (b2ad9b256c) and resolves its TODO(U6-C5/C6) fence, whose three blocked facts are now settled on the tree. - Pipeline: OracleDocIr is source-ordered items carrying a stable OracleItemId + OracleUnitSource, plus closed DocumentRelationIr relations keyed by exact producer/consumer IDs. DocEmitter is the single emission authority — every preprocessor, Class included. - SpanPrecision table: now Exact + ChainRelative only. WholeDocument is retired (its sole producer, the Class façade, is gone). Records that ChainRelative is the last non-card-absolute tier and the only one a renderer must refuse to print as a card position. - Documents the ordering authority explicitly: OracleDocBuilder keys items by (first_line, ordinal) and RE-SORTS, so emission order is irrelevant. To perturb item order you must perturb the span, never the sequence of emit calls — a parity probe that reorders calls instead of spans comes back falsely green. - Assembly (TODO fence resolved): the arena is keyed by node INDEX, not ClauseId — a continuation can push an output node no clause owns, so a ClauseId-only key cannot address every amendable node. AssemblyEnv carries a typed role registry per amendable class (6 of them), as LISTS not last-only slots, because a guarded selector may walk past a candidate that fails its guard. Bindings are (AntecedentRole, AntecedentSelector). Roles are deliberately narrow and non-nested: DigOrRevealUntil is a different set from DigOrMill even though both contain Dig, because widening to a superset re-opens the nearest-match trap for every card in the difference. - lower_effect_chain_ir survives as a one-line delegator; the ~30 pub(super) clause-lowering helpers did NOT relocate out of lower.rs. Both stated as fact. "Known Milestone-1 debt" re-derived against the current tree rather than carried forward: the Class shim bullet is deleted (retired this unit), PreLowered* is re-counted (26 refs) and re-scoped to ITEM granularity, TriggerBody::PreLowered points at the bring-up plan, and two items are ADDED — ChainRelative as the last imprecise span tier, and the 28 undeclared positional bindings that remain outside the arena. Carries the U20 structural finding as a standing warning: byte-identity is STRUCTURALLY BLIND to a narrowed binding (every walk-back site sees exactly one candidate on today's pool, so all nearest-match rules are output-identical). Never sign off a binding migration on byte-identity alone — use a forced diagonal. Adds a "Retired in Milestone 1 (do not re-introduce)" list so the deleted scaffolding cannot be read back in as the target architecture. scripts/check-skill-doc.sh passes.
The card-data-export binary was removed; the alias silently failed for every agent invoking it. oracle-gen (with --features cli) is the export entry point, matching the semantic-audit alias's invocation.
Contributor
|
Warning You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again! |
matthewevans
enabled auto-merge
July 12, 2026 05:59
export-json and migrate-abilities point at binaries with no [[bin]] declaration anywhere in the workspace (same cluster as the retired card-data-export). A dead alias is worse than none: it fails with a confusing 'no bin target' instead of 'no such command'.
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.