Skip to content

Affected-set de-fork: close function_edited gap in floor disposition (partial #5994) - #6019

Closed
gunbai-bot[bot] wants to merge 10 commits into
mainfrom
session/cool-fox-169
Closed

gunbai-bot[bot] wants to merge 10 commits into
mainfrom
session/cool-fox-169

Conversation

@gunbai-bot

@gunbai-bot gunbai-bot Bot commented Jun 30, 2026 •

Copy link
Copy Markdown
Contributor

Summary

Partial Step-3 prep only — does NOT close witness (a) PROVE gate.

Closes the function_edited gap in floor_witness_run_disposition (design #5994 Step 4 prerequisite):

  • Axis (i) node-frontier and axis (ii) function-edited both feed the .dag skip disposition (touches_frontier || function_edited → run; both false → skip).
  • Floor-enrolled witness affected_set_disposition_both_axes_witness_holds exercises the disposition model in isolation.
  • Rust floor_disposition_kernel_alignment test checks floor_kernel_would_skip agrees with the live Rust skip predicate when both sides share Rust-computed touches_frontier / function_edited — disposition-kernel alignment, not impl-vs-impl.

Witness (a) still open (operator sequence): independent .dag affected_set_closure vs NodeFrontierSeeds on a real origin/main...HEAD diff, superset assertion, full run/skip on both axes, RED perturbation control. No MIGRATE/DELETE until that gate is green.

Also: nested-cargo compile guards retry on transient runner temp-dir/sccache errors (CI flake hardening).

Test plan

  • cargo test -p v1-compiler --lib disposition_kernel_aligns — pass
  • cargo test -p v1-compiler-tests floor_skip — 9/9 pass
  • gunbc run --claim-run … affected_set_disposition_both_axes_witness_holds → true

@gunbai-bot gunbai-bot Bot changed the title Implement affected-set CI de-fork (signed design #5994 / docs/plans/affected-set-precompute-pruning.md), AUDIT-FIRST: prove witness (a) full-predicate equivalence (.dag affected-set closure superset of Rust NodeFrontierSeeds on a REAL diff, BOTH skip axes node-frontier AND edited_test_fns, green-by- Affected-set de-fork: witness (a) full-predicate equivalence (#5994) Jun 30, 2026
@gunbai-bot
gunbai-bot Bot marked this pull request as ready for review June 30, 2026 18:31
@gunbai-bot

gunbai-bot Bot commented Jun 30, 2026

Copy link
Copy Markdown
Contributor Author

Review feedback (cursor/composer-2.5 APPROVE): Verified — no code changes required. The approval is valid for the partial Step-3 landing (disposition model + disposition-level equivalence test). Operator sequence (stern-moth-225) correctly notes PROVE gate gaps remain: impl-vs-impl on a real diff, superset assertion, and RED control are follow-on before MIGRATE/DELETE.

CI failure (rust_tests, 2 failing): Investigated — both failures are runner infra flakes, not code regressions:

  • bootstrap::stage0_cargo_check — couldn't create a temp dir: No such file or directory during parallel cargo check on the runner
  • v1_compiler_lib_tests_compile_green — same temp-dir error during cargo test --release compile

Both pass locally (stage0_cargo_check ok in 31s, v1_compiler_lib_tests_compile_green ok in 73s). Pushed 8e8a0e2→HEAD to retrigger CI.

— sent from cool-fox-169

@gunbai-bot gunbai-bot Bot changed the title Affected-set de-fork: witness (a) full-predicate equivalence (#5994) Affected-set de-fork: close function_edited gap in floor disposition (partial #5994) Jun 30, 2026
@gunbai-bot

gunbai-bot Bot commented Jun 30, 2026

Copy link
Copy Markdown
Contributor Author

cursor/composer-2.5 REQUEST_CHANGES — accepted (valid finding).

Verified against current code: the review is correct. At cli_run.rs the Rust harness feeds Rust-computed entry_touches / function_edited into both sides and only exercises .dag floor_kernel_would_skip — disposition-kernel alignment on shared inputs, not impl-vs-impl witness (a).

Fix (8428818): Rescoped so the PR no longer presents Step-3 PROVE as closed:

  • Renamed Rust mod → floor_disposition_kernel_alignment; renamed floor witness → affected_set_disposition_both_axes_test.dag / affected_set_disposition_both_axes_witness_holds
  • Updated title/body to state explicitly: partial prep only; witness (a) (independent .dag affected_set_closure vs NodeFrontierSeeds, real diff, superset, RED control) remains follow-on per operator sequence

The .dag function_edited disposition change (affected_set_floor_runner.dag:79-88) and disposition witness remain; merge of this PR should be understood as landing that prerequisite, not discharging witness (a).

— sent from cool-fox-169

@gunbai-bot

gunbai-bot Bot commented Jun 30, 2026

Copy link
Copy Markdown
Contributor Author

cursor/composer-2.5 APPROVE — verified, no further code changes.

Checked against 08de05a0fa:

  • affected_set_floor_runner.dag:79-88 — both skip axes in disposition kernel ✓
  • affected_set_disposition_both_axes_test.dag — floor witness for isolated disposition logic ✓
  • cli_run.rs:4093-4097 — Rust harness explicitly scoped as disposition-kernel alignment (shared Rust inputs), not witness (a) impl-vs-impl ✓
  • PR title/body defer witness (a) with concrete open criteria ✓

Prior REQUEST_CHANGES (disposition tautology presented as witness (a)) addressed in 8428818 rescope + rename. This approval matches the corrected partial scope.

Merge readiness: 2/2 dashboard approvals (claude, cursor), no REQUEST_CHANGES, rust_tests green on HEAD; ci floor job still pending on this SHA.

— sent from cool-fox-169

briansrls and others added 7 commits June 30, 2026 19:21
…ate.

Documents that impl-vs-impl equivalence on a real diff remains open per #5994.

Co-authored-by: Cursor <cursoragent@cursor.com>
bootstrap::stage0_cargo_check and v1_compiler_lib_tests_compile_green
now use the same CARGO_BUILD_JOBS=1 / no-sccache retry ladder as ci.yml
when stderr shows temp-dir or sccache spawn errors under parallel nextest.

Co-authored-by: Cursor <cursoragent@cursor.com>
Rename tests/modules to stop claiming full-predicate equivalence; the
Rust harness is disposition-kernel alignment on shared Rust inputs only.
Witness (a) impl-vs-impl remains follow-on per #5994.

Co-authored-by: Cursor <cursoragent@cursor.com>
Co-authored-by: Cursor <cursoragent@cursor.com>
@briansrls
briansrls force-pushed the session/cool-fox-169 branch from 08de05a to 04fba00 Compare June 30, 2026 19:23
@gunbai-bot

gunbai-bot Bot commented Jun 30, 2026

Copy link
Copy Markdown
Contributor Author

claude/claude-opus-4-7 APPROVE — verified on rebased 04fba00c5c, no code changes.

Re-checked after rebase onto main (conflict in bootstrap.rs resolved — kept run_cargo_with_infra_retry, which subsumes main's nested-cargo sccache guard):

  • affected_set_floor_runner.dag:79-88 — function_edited axis in disposition kernel ✓
  • Existing call sites pass function_edited: false where appropriate; new tests cover the axis ✓
  • floor_disposition_kernel_alignment Rust mod is #[cfg(test)] disposition prep, not witness (a) PROVE gate ✓
  • helpers.rs cargo infra-retry helper retained post-rebase ✓

Approval matches PR scope. CI re-running on 04fba00c5c.

— sent from cool-fox-169

@gunbai-bot

gunbai-bot Bot commented Jun 30, 2026

Copy link
Copy Markdown
Contributor Author

claude/claude-opus-4-7 APPROVE (04fba00c5c) — verified, no code changes.

Findings match current tree: function_edited threaded uniformly through disposition kernel/scaffold; coproduct dispatch preserved; .dag witnesses cover both axes; Rust alignment test calls shared floor_kernel_would_skip (disposition prep, not witness (a)); cargo infra-retry helper is test scaffolding consistent with CI discipline.

— sent from cool-fox-169

@gunbai-bot

gunbai-bot Bot commented Jun 30, 2026

Copy link
Copy Markdown
Contributor Author

cursor/composer-2.5 APPROVE (04fba00c5c) — verified, no code changes.

Confirmed: disposition kernel + both-axis witnesses land correctly; Rust module is honestly scoped disposition prep; witness (a) deferral explicit in PR body. Matches prior rescope that addressed the earlier REQUEST_CHANGES.

— sent from cool-fox-169

@gunbai-bot

gunbai-bot Bot commented Jun 30, 2026

Copy link
Copy Markdown
Contributor Author

cursor/composer-2.5 APPROVE (872595503) — verified after main merge, no code changes.

Post-merge check: function_edited disposition kernel, floor witness (affected_set_disposition_both_axes_test.dag), and floor_disposition_kernel_alignment Rust prep remain intact on this SHA. Partial scope / witness (a) deferral unchanged.

— sent from cool-fox-169

@gunbai-bot

gunbai-bot Bot commented Jun 30, 2026

Copy link
Copy Markdown
Contributor Author

claude/claude-opus-4-7 APPROVE (872595503) — verified, no code changes.

Confirmed on post-main-merge HEAD: function_edited axis in disposition kernel, six test call sites updated, floor_test_function_edited_runs_holds wired into policy aggregate, affected_set_disposition_both_axes_test.dag witness present, fail-closed paths preserved, Rust additions scoped as disposition prep + cargo retry scaffolding. Matches PR intent.

rust_tests green on this SHA; ci floor job still pending.

— sent from cool-fox-169

briansrls and others added 2 commits July 1, 2026 00:23
Release build verified green locally; CI failed with sccache compile
exit status 2 under fleet pressure, not a code regression.

Co-authored-by: Cursor <cursoragent@cursor.com>
@gunbai-bot

gunbai-bot Bot commented Jul 1, 2026

Copy link
Copy Markdown
Contributor Author

claude/claude-opus-4-7 APPROVE — verified on 1693478bb5, no code changes.

Re-checked current HEAD:

  • affected_set_floor_runner.dag:62-88 — function_edited threaded through floor_witness_run_disposition; skip only when both touches_frontier and function_edited are false ✓
  • floor_kernel_would_skip / floor_host_scaffold_would_skip mirror the same predicate ✓
  • affected_set_disposition_both_axes_test.dag — floor witness covers both axes in isolation ✓
  • cli_run.rs floor_disposition_kernel_alignment — #[cfg(test)] disposition prep with explicit PROVE-gate-open comment; not presented as impl-vs-impl ✓
  • helpers.rs::run_cargo_with_infra_retry — shared cargo infra-retry ladder used by bootstrap.rs and v1_compiler_lib_test.rs ✓

Approval matches the narrowly scoped partial prep. Witness (a) impl-vs-impl on a real diff remains follow-on per #5994.

— sent from cool-fox-169

@briansrls briansrls closed this Jul 1, 2026
briansrls added a commit that referenced this pull request Jul 1, 2026
Fold cool-fox-169 floor-enrolled claim witness into this migration;
adapted for entry_file_touched third axis already on this branch.

Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls added a commit that referenced this pull request Jul 1, 2026
Fold cool-fox-169 floor-enrolled claim witness into this migration;
adapted for entry_file_touched third axis already on this branch.

Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls added a commit that referenced this pull request Jul 1, 2026
Fold cool-fox-169 floor-enrolled claim witness into this migration;
adapted for entry_file_touched third axis already on this branch.

Co-authored-by: Cursor <cursoragent@cursor.com>
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