Repository navigation
Resolver cost: demand-scoped typed census, import-facts cost shape, bare-admission scoping - #12970
gunbai-bot[bot] wants to merge 45 commits into
Conversation
The via-index parse and the parse-cache miss arm re-lexed every closure file the pool census had already lexed under the same spelling. Both now ask pool_acquire. Adds the per-term attribution probe and a live identity differential (pooled vs fresh artifact over the whole pool). Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…-parsing pool_acquire holds one file-local heads reading (empty intern table, occurrence ordinals from zero). module_path_index reads it as-is; the pool census maps it into its threaded intern/occurrence space by a total projection (occurrence ids offset by the entry base, idents relabeled through the file's string list) instead of parsing the file a second time. Anything the parser does not produce refuses. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Every Node field and every ExprData / MatchPattern / InferredNode arm is named with no '..' or wildcard, so a field or variant added later fails to compile at the walker instead of passing through with file-local ids (review 72735). Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
build_symbol_index_census_nodes is census_with_resolved_fn_sigs over the raw census; closure_name_census(index, Some(root)) already builds and memoizes exactly that raw census over the same nodes and source indices. tree_bare_census_for_root now upgrades the memoized value. Adds a live whole-SymbolIndex differential per root. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…heads # Conflicts: # src/v1/stage0/src/cli_run/entry_resolve.rs
…ites it, so Some refuses Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…s' into session/bold-bat-516-bare-fill
…typed-census # Conflicts: # src/v1/stage0/src/cli_run/entry_resolve.rs
…t-516-bare-fill # Conflicts: # src/v1/stage0/src/v1_compiler_infer.rs
…refusal, builtin arm Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…into session/bold-bat-516-bare-scope
…re admission instrument Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…ads name index once per index; transitive-pick row rung and trigger restated Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…bare-fill # Conflicts: # src/v1/stage0/src/cli_run.rs # src/v1/stage0/src/cli_run/entry_resolve.rs # src/v1/stage0/src/v1_compiler_infer.rs
…into session/bold-bat-516-bare-scope
The builtin arm no longer pulls another tree's fn get for a builtin get call, so these pairs no longer occur (floor: RosterStale ... retire it as ImportsFixed). Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…essWet rows, with its declared rung drop A changed witness whose file is a BinWitnessWet WitnessExclusionRow is declined, counted and located per row, instead of planned into a certain route gap: no CI lane executes that class. The loss is declared as gunbc.rung_drop.edited_bin_witness_wet_rows_not_executed_by_ci (appended to the roster; docs/design-rung-drops.md regenerated). Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…ane' into session/bold-bat-516-bare-scope
…; retire the 12 get pairs as NotAReference (review 73273) A test-row provider was never pulled, so it does not refuse. The get pairs stopped occurring because the reader now resolves get as the builtin, not because any import was fixed. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…3267) The rung drop's population is one data list, edited_bin_witness_wet_rows_not_executed_by_ci_population, which the RungDrop row and the floor gate both read. A BinWitnessWet row it does not name is not declined, so the drop is bounded and widening it is an edit to that list. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…ane' into session/bold-bat-516-bare-scope
…issing on this head Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…n entries Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…ane' into session/bold-bat-516-bare-scope
…bare-scope # Conflicts: # src/v1/stage0/src/cli_run/entry_resolve.rs
…disposition gunbc#12794's decline pushed a disposition row, but the sublane's exactness join counted only PlannedAsChangedWitness, so every declined selection refused as selected_without_disposition (observed on gunbc#12741's floor, run 36768985832). The decline was unreachable by its route. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…-join' into session/bold-bat-516-bare-scope
…austively, and file the class changed_witness_sublane_join is extracted and driven with a DeclinedNoCiWetLane selection (red on the pre-fix allow-list predicate with ChangedWitnessSublaneJoinInexact, green here); decides_a_changed_selection is an exhaustive match, so a new disposition arm cannot compile without stating its membership. Class filed as gunbc.recurring_failure_mode.a_new_decision_arm_the_downstream_join_does_not_admit. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…-join' into session/bold-bat-516-bare-scope
…73414) Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…12897 decider mirror test Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
… typed-graph exclusive-bytes reader Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…as DeclinedNoCiWetLane suppress_withheld keeps every changed witness in the expected-red, route-gap and non-verdict rosters on the premise that the changed sublane executes it. A DeclinedNoCiWetLane decline falsifies that premise, so the route-gap reverse join refused three declined witnesses as stale on this PR's floor. They are now suppressed with their own ground (ExpectedRedSuppressionGround::DeclinedNoCiWetLane), decided by an exhaustive match. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…h; census filed as the RFM row's second specimen Five hand-written seed sites decided by catch-all or matches!. partition_cost_debt_roster's Some(_) contradicted its .dag authority for the two changed-selection declines (now OutsideThisRunsUniverse, unit red on the restored catch-all). The others are converted without behavior change; the site-projection line now counts the two declines it omitted. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 315912398e
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| }; | ||
| let mut route_gap_roster = route_gap_roster; | ||
| let _ = suppress_withheld(&mut route_gap_roster, "floor_route_gap"); | ||
| let _ = suppress_declined_no_ci_wet_lane(&mut route_gap_roster, "floor_route_gap"); |
There was a problem hiding this comment.
When a changed BinWitnessWet identity also has a typed route-gap expectation—several declared members do, such as the namespace-structural and self-host artifact witnesses in src/v2/workflow/floor_route_gap.dag—this line removes the identity from route_gap_roster, but the expectation decoder at lines 9559-9562 still iterates its expectation and refuses because the identity is now absent from that roster. Consequently, edits to those explicitly non-blocking wet rows still fail the required floor with located identity is absent from derived roster; suppress the corresponding expectation under the same DeclinedNoCiWetLane condition.
Useful? React with 👍 / 👎.
| @@ -0,0 +1,8 @@ | |||
| CG=/sys/fs/cgroup/gunbc-probe; S=""; command -v sudo >/dev/null && S=sudo | |||
| echo "+memory" | $S tee /sys/fs/cgroup/cgroup.subtree_control >/dev/null; $S mkdir -p $CG; echo 23622320128 | $S tee $CG/memory.max >/dev/null; echo $$ | $S tee $CG/cgroup.procs >/dev/null | |||
| git stash push -q -- src/v1/stage0/src/cli_run/required_floor_runner.rs | |||
There was a problem hiding this comment.
When this helper is run with local changes to required_floor_runner.rs, git stash push -q -- <pathspec> removes those changes from the worktree, but the script never executes git stash pop or git stash apply. Because the stash is also quiet, the helper appears to finish while leaving the user's edits hidden in the stash; preserve and restore the file without mutating the caller's stash state.
Useful? React with 👍 / 👎.
|
Closing rather than rebasing. This PR was auto-opened when bold-bat-516's session was closed out. Its branch is #12741's history, which already landed on main by squash (hence the conflicts), plus one close-out WIP commit (3159123) that contains only untracked scratch scripts (target_*.sh) and isn't meant to land. Nothing here needs merging. The branch is kept. — sent from jolly-boar-500 |
Auto-opened by session-dashboard for session
bold-bat-516.Pushing to
session/bold-bat-516-bare-scopeadvances this PR.Worker attestation
Before flipping this PR to ready for review, confirm each item:
npm test,cargo test) and the result.Closes #Ndirective.Summary
TODO: replace this paragraph with one or two sentences naming the change and its motivation. Reviewers read this first.
Test plan