Repository navigation
feat(frontier): --only / --since walk only the affected specs - #10244
Merged
Merged
Conversation
Contributor
This was referenced Oct 18, 2026
…10236) `t27c frontier` decided every spec under specs/. Add `--only <spec>...` and `--since <rev>`: the walk is the named specs (or the specs `git diff --name-status <rev>` names) plus every spec that imports them, directly or transitively. The selection is specs/ci/affected.t27 select_affected, the rule CI already uses; the Rust gathers the diff, the `use` edges and the corpus and hands them to affected's own generated code, as CI's C stub does. A change affected.t27 cannot judge (a path outside specs/) walks the whole corpus and says so. With neither flag the output is unchanged. specs/verified/frontier_scope.t27 holds the two decisions around the selection (scope_wanted, scope_narrows). It does not `use ci::affected`: a spec that imports it does not compile, because affected declares in_list, ext_dot, KINDS and PROSE_EXTS and imports policy::own_language, which declares the same four names differently, so the resolver leaves them unspliced. The stale four-line doc comment above the Frontier variant (the run_record text, attached by a merge) goes, and `total` in run_frontier is inlined, so the hand-written Rust loses more lines than it gains. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
… the tri hot doc (Refs #10236) Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
…0236) The t27b ratchet names a spec the reference passes and the ledger does not list as UNLISTED. frontier_scope.t27 passes under t27b, like its neighbours. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
gHashTag
force-pushed
the
claude/10236-frontier-only-since
branch
from
October 18, 2026 05:43
2e3851d to
1857352
Compare
gHashTag
enabled auto-merge (squash)
October 18, 2026 05:43
Contributor
PR DashboardGenerated at: 2026-10-18 05:49:54 UTC
Summary
Seal Status
|
This was referenced Oct 18, 2026
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.
Closes #10236
Epic #8095 (hot reload), ledger #8737. First step of the change-only inner loop.
What changes
t27c frontier --only <spec>...and--since <rev>walk the named specs (or the specsgit diff --name-status <rev>names) plus every spec that imports them, directly or transitively. The selection isspecs/ci/affected.t27select_affected, the rule CI already uses; each spec in scope still getsseal_identity.t27node_decision, and its closure is read in full.specs/) walks the whole corpus and the output says so.--onlyand--sinceare refused together by clap.frontier: scope N of M specsis printed only when a flag is given. With neither flag the output is unchanged.specs/verified/frontier_scope.t27holds the two decisions around the selection (scope_wanted,scope_narrows; 4 tests, 1 invariant, negative control for the sentinel). It does notuse ci::affected: a spec that imports it does not compile (t27c test-report: BLOCKED,use of undeclared identifier 'own_language'), because affected declaresin_list,ext_dot,KINDSandPROSE_EXTSand importspolicy::own_language, which declares the same four names differently, so the resolver leaves them unspliced. The Rust therefore calls affected's own generated code (bootstrap/gen/rust/ci/affected.rs, standalone), as CI's C stub does. Renaming affected's four helpers would let importers compile; that is a separate change to a CI-critical spec and is not made here.bootstrap/src/main.rsandbootstrap/src/service.rs(the stale four-line run_record doc comment aboveFrontiergoes;totalinrun_frontieris inlined).bootstrap/src/test_report.rsis untouched.Measured on the Railway lab (load average about 96, shared)
frontier --listdiffers from master only by the new spec itself (6 specs/verified/frontier_scope.t27, total 2227 -> 2228, missing 464 -> 465); every other line is identical.frontier --only specs/verified/watch.t27: scope 1 of 2228, 0.18 s, against 0.75 s for the full warm walk.frontier --only specs/verified/reuse.t27: scope 45 of 2228 (reuse plus its transitive importers); all 7 lines it listed are also in the full list.frontier --since HEADwith a one-line edit ofreuse.t27in the working tree: the same 45 specs, 0.43 s.--since HEAD~3(touchesbootstrap/src): whole corpus.--since nosuchrev: error, not an empty scope.--onlywith--since: clap refusal.t27c test-report specs/verified/frontier_scope.t27: 4 of 4 pass, 0 vacuous.🤖 Generated with Claude Code