Skip to content

fix: accept reviewed set-aside intents as complete review scope - #338

Merged
vinceblock99 merged 1 commit into
fix/triage-reviewfrom
fix/review-scope-set-aside
Nov 15, 2026
Merged

vinceblock99 merged 1 commit into
fix/triage-reviewfrom
fix/review-scope-set-aside

Conversation

@vinceblock99

@vinceblock99 vinceblock99 commented Nov 13, 2026 •

Copy link
Copy Markdown
Contributor

Stacked on #336 (fix/triage-review).

Problem

The shape:Subject scope-completeness rule only recognizes includesIntent, but the resolver pins icebox intents under includesSetAside. So a reviewed change whose retained receipts observe an intent that was later set aside fails validation with

Include the observed work intent in the original review intent's reviews references…

even when the review intent explicitly reviews that intent. There is no way to satisfy the rule from the review side.

Reproduction (now the regression test reviewed_set_aside_intent_observed_by_a_change_does_not_block): one change, one retained receipt attesting two work intents by the same author; the second intent is later set to icebox; the review intent reviews both. Before this change validation fails with the message above; after it conforms.

Fix

  • review-1.0.0.ttl: the scope-completeness FILTER NOT EXISTS accepts includesIntent or includesSetAside, and the message says so.
  • Managed-change coverage is untouched and still requires includesIntent, so a set-aside intent cannot stand in as implementation evidence. The existing retired_intent_cannot_supply_managed_change_coverage test still passes, and the new test also asserts that dropping the set-aside intent from the review scope is rejected.
  • Test fixture gains Fixture::with_second_observed_intent() (one receipt node attesting several intents in a single multi-command invocation); existing receipt_node callers are unchanged.
  • One sentence added to docs/retrospective-publication-review.md.

Note: the shapes file participates in profile_hash, so reviews signed before this change will read as stale and need a superseding signature. That is the documented behaviour for any shape change.

Tests

  • cargo test -p atomic-repository --features review-shacl --lib publication_review: 40 passed (39 existing + new)
  • cargo test -p atomic-repository --features review-shacl: 1903 passed, 0 failed
  • cargo test --workspace: 13339 passed; the only 2 failures are the pre-existing local failpoint tests that need --features adoption-test-injection (git_shim_cb13d_test::crash_during_effect_recovers_without_partial_bytes, graph_only_export_cb9b_test::graph_only_export_matrix_from_recorded_graph), unrelated to this change
  • cargo test -p atomic-shacl, atomic-core --test kg_ontology, atomic-repository --test it intent_graph: all green

@vinceblock99
vinceblock99 marked this pull request as ready for review November 13, 2026 20:42
@vinceblock99
vinceblock99 force-pushed the fix/review-scope-set-aside branch from 98201f2 to f4f7256 Compare November 13, 2026 20:49
The subject scope-completeness rule only recognized includesIntent, but
the resolver pins icebox intents under includesSetAside. A change whose
retained receipts observe an intent that was later set aside could never
pass validation, even when the review intent explicitly reviews it.

Accept either relation for scope completeness. Managed-change coverage
still requires includesIntent, so a set-aside intent cannot stand in as
implementation evidence (covered by the existing
retired_intent_cannot_supply_managed_change_coverage test).

Note: the shapes file participates in profile_hash, so previously signed
reviews will read as stale and need a superseding signature.
@vinceblock99
vinceblock99 force-pushed the fix/review-scope-set-aside branch from f4f7256 to b6f14cc Compare November 15, 2026 17:55
@vinceblock99
vinceblock99 merged commit e0b0d17 into fix/triage-review Nov 15, 2026
@vinceblock99
vinceblock99 deleted the fix/review-scope-set-aside branch November 15, 2026 18:04
vinceblock99 added a commit that referenced this pull request Nov 15, 2026
* fix the review process with new ontology

* fix: accept reviewed set-aside intents as complete review scope (#338)

The subject scope-completeness rule only recognized includesIntent, but
the resolver pins icebox intents under includesSetAside. A change whose
retained receipts observe an intent that was later set aside could never
pass validation, even when the review intent explicitly reviews it.

Accept either relation for scope completeness. Managed-change coverage
still requires includesIntent, so a set-aside intent cannot stand in as
implementation evidence (covered by the existing
retired_intent_cannot_supply_managed_change_coverage test).

Note: the shapes file participates in profile_hash, so previously signed
reviews will read as stale and need a superseding signature.

---------

Co-authored-by: Vincent <vince@atomic.dev>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant