Repository navigation
ShellProgram -> DAG - #6467
Merged
Merged
ShellProgram -> DAG#6467
Conversation
…wall on the whole-tree emit path
Fold-family productions extended so the fold no longer refuses word
Concat/CmdSubst or stmt WithRedir (all four Redir variants): new
concat_parts/word-compound/with_redir production families, lex tokens,
kind-tag emit transforms, and recursive bundle arms. Byte-identity vs
serialize_bash proven by execution: five depth-2 oracle tests plus the
depth-4 assign_root_stmt manual probe (ROOT=$('git' 'rev-parse'
'--show-toplevel' 2>/dev/null || 'pwd') byte-exact). The delegated
fail-closed RED control repoints from WithRedir (now native) to Heredoc
(still delegated) so the boundary guard stays discriminating.
Measured cost wall, declared on-carrier (bash_program_emit_cost_wall_note):
whole-tree backward row-selection is ~alternatives^depth (1s flat stmt,
64s for the single depth-4 stmt, DNF >8min for the full witness_bin
program) because formal_production_unique_lhs_exact_match deep-validates
every candidate per level and the descent re-validates each level again.
Real-program oracles therefore stay MANUAL probes, not test fns (a
discovery-run test would hang the local floor); the probe note on the
test carrier names the dissolution triggers.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…t runtime) Phase-3 shape for the shared witness invocation, per the shell-emission-model rule that runtime-present sites are argv invocations wearing bash syntax. run_witness_bin_program and run_gunbc_claims no longer build a ShellProgram, serialize it, and shell out - they realize the same intent as typed service operations: - git.Core.Toplevel (new op beside CurrentBranch/HeadCommit) resolves the repo root the bash path derived with a cd/rev-parse dance - shell.Test.IsExecutable (new, POSIX test -x) replaces the test-not- executable guard and the post-build artifact verification - cargo.Build.BuildInheritEnv (new) is the build-if-absent arm; it inherits process env exactly as the old sh -c cargo build did (cargo.Build.Build's env Map default is both a behavior change and unrecordable by the claim- run fixture writer, which fail-closes on Map values) - gunbc.WitnessBin.Run (new) invokes the witness binary by absolute path with a spliced args list Proven by execution: six op probes incl. argv[0] interpolation, list splice, and negative RED controls (typed_witness_invocation_test); pattern A end-to-end (v1_dag_parse witness, 0.43s wall vs bash round-trip); the run_gunbc_claims rider (extdeps_external_authority uri witnesses); and the flip-level RED control (unbuildable bin -> false, refused not fabricated). The ShellProgram builders stay as the canonical fixtures for the bash row-emit byte-identity oracles; importer ratchet unchanged at 27. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…rs 27 -> 20 Second typed-argv tranche, riding the host_prelude flip: - dag_compile_clean_transport: the four perturb receipts and the gate are one shared typed composition (Mktemp.Dir, Filesystem.Write for the perturb module - the heredoc dissolves, WitnessBin.Run on gunbc, Remove.RecursiveForce), parameterized by expected compile outcome. The ShellProgram builders, their serialize-shape witnesses, the marker rows, and both RealizationDispatch scaffold rows dissolve: with typed args the invocation facts are the construction, so the spelling-drift class those witnesses validated is unwritable (DESIGN 5) - the witness test file and the duplicate_computation compile-clean subject go with them (noted on the carrier; the lens keeps its emit_determinism subjects). - dag_compile_clean_seam: the RawLine pile (cp -r, sed, $VAR-captured compiles, grep -c, echo SEAM_RESULT) becomes typed ops end to end - the sed is Filesystem.Read + replace + Write, the three compiles return typed results, and the SEAM_RESULT string protocol plus its parser dissolve; the receipt is a typed record and the synthetic RED controls construct it directly. - dag_compile_clean_shard / shard_totality: typed staging compile and a Find.Files op (the '| sort' was cosmetic - the totality algebra is set-membership both ways). - source_root_ingest_transport, compiler_closure_ingest_transport, self_host_realized_comparison_gate: typed compositions over the shared helpers (run_witness_bin_program / run_gunbc_claims). - New extdeps ops: shell.Mktemp.Dir/DirWithTemplate, shell.Mkdir.Parents, shell.Remove.RecursiveForce, shell.Copy.File/Recursive, shell.Find.Files, shell.Test.IsNonEmpty (review fix: restores the test -s zero-byte guard the bash emit_verifications carried). - Importer ratchet re-frozen 27 -> 20; seven stale exception-roster rows removed; ratchet + planted RED controls + clean-tree + roster-soundness all green. Verified by execution: seam algebra + seam live probe, totality live, shard exemplar compile, cross-tree green + optional-skew red receipts, ingest gate (incl. a live cold-build refusal: remote-routed cargo produced no local artifact and the non-empty/executable verification failed closed), closure ingest gate, pattern A, full co-rooted corpus compile x2 clean. Deferred with in-tree notes: floor_diff_observe (affected-set CI's diff observer - avoid colliding with that lane), emit_determinism_transport (last duplicate_computation ShellProgram subject), emit_host_transport (go/node arms are genuine toolchain-bootstrap windows; rust arm needs an env-pinning build op), bmc/ubuntu/srv3 provisioning tails. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
… regression) The dissolved bash prelude carried cd "$ROOT" - the construction that made relative --source-root values and cargo build unambiguous. The typed flip dropped it: WitnessBin.Run, BuildInheritEnv, and Find.Files ran in inherited process cwd. Fix: each op now takes a required workdir input spelled as GNU env -C (coreutils env(1)), so a call that does not say where it runs is a type error, not a latent host-state dependency. All call sites thread git.Core.Toplevel(). New discriminating probes: pwd under workdir=/tmp and workdir=toplevel; ingest gate re-run green from a repo subdirectory (the exact failure class flagged). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
…unbc into session/witty-ibex-317
…, Bash); on_exhausted wired Precondition first (plan-doc FLAG): Retry.on_exhausted was a declared-but-unread field - the emitter matched it as _ and the row baked exhaustion as fixed tokens. Wired end to end: seg_h split into four fixed segments with three shared-binding occurrences of an exhausted spelling, the emitter emits the on_exhausted Pipeline and threads it through realize_retry (arity 8), and a new teeth test reds the bytes when on_exhausted changes. Cutover: gunbc.ci_spec.ci_cargo_eagain_retry_core no longer concatenates bash - it builds the Retry intent (body run, two LogMatches escalation levels, exhausted pipeline: the policy facts ci_spec owns) and renders through orch_emit_step over the bash grammar rows. First dag-tree module importing v2.* (gunbc ci loads both roots; workflow->compiler is the allowed layering direction). The String seam's refusal arm emits a loud poison marker that reds both the drift gate and the yaml parse gate - refuse, never widen; noted on the carrier. The hand-authored blob moved verbatim to the test module as an emitter-independent golden; the intent fixture module now delegates to the production intent so tests exercise the value production emits. Receipts: 5 retry emit tests green (byte-identity vs golden, production seam vs golden, env/exhausted teeth); ci_yaml_parse_witness (the CI drift+parse gate) ExitSuccess - committed ci.yml byte-identical through the new path; full corpus compile 0 diagnostics. Attribution: the local 'gunbc ci' top-level failure (dash rejecting set -o pipefail) reproduces on clean origin/main - pre-existing, not this change. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Jul 11, 2026
Bake in parent-endorsed FLAG resolutions (v2.std.execution_surface, per-site window, roster freeze at 58); mark slice 0 LANDED (#6467); cross-link lane E fleet-converge ruling for slice 2 sequencing. Co-authored-by: Cursor <cursoragent@cursor.com>
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Jul 11, 2026
Bake in parent-endorsed FLAG resolutions (v2.std.execution_surface, per-site window, roster freeze at 58); mark slice 0 LANDED (#6467); cross-link lane E fleet-converge ruling for slice 2 sequencing. Co-authored-by: Cursor <cursoragent@cursor.com>
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Jul 11, 2026
Bake in parent-endorsed FLAG resolutions (v2.std.execution_surface, per-site window, roster freeze at 58); mark slice 0 LANDED (#6467); cross-link lane E fleet-converge ruling for slice 2 sequencing. Co-authored-by: Cursor <cursoragent@cursor.com>
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Jul 11, 2026
Bake in parent-endorsed FLAG resolutions (v2.std.execution_surface, per-site window, roster freeze at 58); mark slice 0 LANDED (#6467); cross-link lane E fleet-converge ruling for slice 2 sequencing. Co-authored-by: Cursor <cursoragent@cursor.com>
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Jul 11, 2026
Bake in parent-endorsed FLAG resolutions (v2.std.execution_surface, per-site window, roster freeze at 58); mark slice 0 LANDED (#6467); cross-link lane E fleet-converge ruling for slice 2 sequencing. Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
added a commit
that referenced
this pull request
Jul 11, 2026
…f gate) (#6474) * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * docs: cross-link provisioning-window design from emission-ingestion-inverse gap B Co-authored-by: Cursor <cursoragent@cursor.com> * docs: parent review fixes for provisioning-window design (Lane D) Bake in parent-endorsed FLAG resolutions (v2.std.execution_surface, per-site window, roster freeze at 58); mark slice 0 LANDED (#6467); cross-link lane E fleet-converge ruling for slice 2 sequencing. Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * fix(ci): skip floor receipt gates when docs-only floor is skipped Documentation-only PRs stamp merge-admission as skipped without running claim_executor, so target/floor-resolve-receipt.txt was never written and the resolve/materialization receipt gates failed closed. Write a ci-floor-disposition marker on the shortcut path and teach both gates to skip loudly when CI_FLOOR_DISPOSITION=documentation_only_skipped. Co-authored-by: Cursor <cursoragent@cursor.com> * Expand docs-only floor shortcut allowlist for receipt-gate companions. Docs-only PRs that fix the floor-skip/receipt-gate mismatch must be able to touch the minimal CI substrate rows without triggering the full floor, which was timing out at 60 minutes on this lane. Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * fix(ci): repair docs-only floor script parse errors and stamp receipt in shell. Fix missing concat comma in ci_documentation_only_floor_shortcut_script, avoid bash ${VAR:-} syntax that breaks DAG parse, replace gunbc merge_admission_stamp in the docs-only path with a shell scaffold (v1 import closure fails with one resolved source on CI), and sync ci.yml + witnesses. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(ci): allow merge_admission producer witness in docs-only companion bundle. The prior fix commit touched merge_admission_producer_witness_test.dag, which forced the full floor because it was outside the closed allowlist. Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * fix(ci): revert companion allowlist, dag-only stamp, checkout root witness Docs-only shortcut is docs/* again (no CI substrate self-shortcut). Remove shell merge-admission stamp; invoke merge_admission_stamp with --source-root dag only. Fix floor_skip_discovery_witness to resolve checkout root at runtime for cross-job artifacts. Sync ci.yml skip prefixes for selection control and merge-admission gate. Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * fix(ci): single merge-base authority + gate-skip dissolve-on entry Add gunbc.ci_diff_defaults.ci_merge_base_ref as the sole authority for origin/main (ci_spec diff_policy.base and merge_admission gate/stamp). Name ci_documentation_only_gate_skip_prefix with Scaffold + dissolve-on trigger; witness both follow-ups in ci_spec_witness_test. Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * fix(ci): emit gate-skip dissolve-on in shell runners + witness tests Prepend # dissolve-on comment into ci_documentation_only_gate_skip_prefix emitted bash (matching floor-shortcut and cgroup-peak scaffolds). Sync all four receipt-gate runners in ci.yml. Witness emitted scripts in ci_spec_witness_test; extract ci_selection_control_script for coverage. Pairs with 9281460: workspace_root discovery unit tests (git-toplevel + cwd-ascent) in cli_run.rs. Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * fix(ci): add HAND-RUST receipt for workspace_root discovery scaffold Model seed-retained disposition + dissolve-on in cli_run_workspace_root_scaffold.dag; floor witness + planning anchor in cli-run-reconcile-defork.md; mirror scaffold marker in cli_run.rs. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(ci): parse fix in workspace_root hand-rust witness Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * fix(ci): stamp floor_running before docs-only disposition branch Persistent self-hosted target/ could retain documentation_only_skipped from a prior docs-only job and fail-open the four receipt gates on a full floor run. Initialize the marker to floor_running at the start of every floor step. Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * fix(ci): bump floor step cap to 180m for legit s1_closure RUN Run 29148664744 hit the 120m floor timeout after ~32m discovery corpus SKIPs and an ~88m legitimate s1_closure_receipt RUN (this PR touches the receipt files). Raise floor/regen to 180m and ci job backstop to 420m. Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * WIP: Lane D — provisioning window/executor-capability model draft (sign-off g * fix(ci): repair post-rebase ci_spec parse + drop stale shell stamp scaffold Rebase onto main (#6473) left a broken concat nest in ci_documentation_only_floor_shortcut_script and resurrected the removed ci_documentation_only_merge_admission_stamp_script — both blocked dag_compile_clean_gate batch-1 resolve. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(ci): scaffold floor_running init shell with dissolve-on marker Review 37154: ci_floor_disposition_marker_init_script lacked the HAND-SHELL scaffold row and on-carrier dissolve-on comment that sibling receipt-gate runners carry. Add Disposition = Scaffold bind, shell_emit_dissolution_trigger, witness receipts, and sync ci.yml floor step opener. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Brian Searls <briansearls1@gmail.com> Co-authored-by: Cursor <cursoragent@cursor.com>
This was referenced Jul 14, 2026
briansrls
added a commit
that referenced
this pull request
Jul 16, 2026
…ve LANDED The .dag carrier is the authority (§6), and its status lines had gone stale against the tree. Verified by reading the live tree on 2026-07-16, not by grep: - Slice 0 — LANDED (#6467). ci_spec.dag:222 ci_cargo_eagain_retry_intent is a real Retry{body:Pipeline{steps:[Do{run}],on_failure:FailFast},escalations, on_exhausted}; :235 routes it through orch_emit_step, with ci_retry_emit_refused_poison as a loud §5 refusal (reds both the ci.yml drift gate and the yaml parse gate rather than masking with a hand-spelled fallback). Its stated precondition is also resolved: Retry.on_exhausted is no longer emitter-ignored (05_emit_orchestration.dag:503 -> :627). - Slice 1 — the If band has LANDED. orch_emit_if_step lowers If WITH else_, and every Predicate arm lowers. For/While still refuse BY DESIGN (the 2026-07-03 pre-runtime census found zero justified sites) — a decision, not a gap. - Slice 2 — LANDED. .github/fleet-converge.sh is now 21 lines (thin-run via gunbc converge --host + the sanctioned fresh-standup bootstrap arm); EmitArtifactThenThinRun is a live transport: arm, no longer prose-only. Decisions are untouched: ADOPT emit(intent,Bash) / REJECT a new ShellProgram AST / the For-While scope ruling / the bash-minimization rule all stand as signed. Only status facts changed. Why this matters: the stale TODO on Slice 0 caused me to dispatch a worker onto finished work today. Added an explicit warning that status lines here are load-bearing and the tree is the ground truth. docs/plans/shell-emission-model.md regenerated via the generated-artifact gate (PASS main_wet). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
Jul 16, 2026
… 2 receipts recorded pending operator sign-off (#6734) * shell→dag arc: record operator flag signs on the census carrier (2a(i)/2b/2c signed, B1 working-default typed-target) + tick landed slice-0/1 roadmap boxes Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Address #6571 review: reconcile P2/B1 consequence (for/heredoc band superseded) + clear the stale FLAG-gating text in the critical-path summary Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * WIP: shell -> dag * Fix #6571 CI red: ROADMAP.md is a GENERATED projection — move the slice-0/1 ticks + slice-2 in-flight note into roadmap_authority.dag (done+operator-sign rows) and regenerate via main_wet; drift gate PASS locally Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * Fix quadratic receipt-emit fold (bare-minimum-cost ruling): left-associated concat over receipt lines → balanced pairwise join, O(len·n) → O(len·log n) copying, byte-identical output by associativity Opus noted-and-waived the defect on #6586; the standing operator ruling (DESIGN §6, 2026-07-10) forbids the 'n is small here' waiver. Proven by srv3_typed_receipt_emit_uses_printf_not_naive_echo (content assertion) green. Root cause — no std linear join authority (N hand-rolled joins: join_slash, join_mirror_urls_for_shell, this) — goes on the Wave 4 ledger, not patched per-site here. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * WIP: shell -> dag * WIP: shell -> dag * Complete the HostEffect arm set: dedupe SetHostnameCas (merge doubled it) + add Srv3InstallDiagnosticObserve / OsInstallActuatorToolchainEnsure arms (#6587's variants landed armless — fifth composition-skew instance); whole-tree compile 0 diagnostics Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> * WIP: shell -> dag * WIP: shell -> dag * plans: refresh shell_emission_model status — Slices 0/1(If band)/2 have LANDED The .dag carrier is the authority (§6), and its status lines had gone stale against the tree. Verified by reading the live tree on 2026-07-16, not by grep: - Slice 0 — LANDED (#6467). ci_spec.dag:222 ci_cargo_eagain_retry_intent is a real Retry{body:Pipeline{steps:[Do{run}],on_failure:FailFast},escalations, on_exhausted}; :235 routes it through orch_emit_step, with ci_retry_emit_refused_poison as a loud §5 refusal (reds both the ci.yml drift gate and the yaml parse gate rather than masking with a hand-spelled fallback). Its stated precondition is also resolved: Retry.on_exhausted is no longer emitter-ignored (05_emit_orchestration.dag:503 -> :627). - Slice 1 — the If band has LANDED. orch_emit_if_step lowers If WITH else_, and every Predicate arm lowers. For/While still refuse BY DESIGN (the 2026-07-03 pre-runtime census found zero justified sites) — a decision, not a gap. - Slice 2 — LANDED. .github/fleet-converge.sh is now 21 lines (thin-run via gunbc converge --host + the sanctioned fresh-standup bootstrap arm); EmitArtifactThenThinRun is a live transport: arm, no longer prose-only. Decisions are untouched: ADOPT emit(intent,Bash) / REJECT a new ShellProgram AST / the For-While scope ruling / the bash-minimization rule all stand as signed. Only status facts changed. Why this matters: the stale TODO on Slice 0 caused me to dispatch a worker onto finished work today. Added an explicit warning that status lines here are load-bearing and the tree is the ground truth. docs/plans/shell-emission-model.md regenerated via the generated-artifact gate (PASS main_wet). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * WIP: shell -> dag * plans: address review 38787 — Slice 1 marker + stop contradicting the roadmap authority Both findings from cursor/composer-2.5 verified against the tree and valid. Finding 1 (Slice 1 §5 row had no LANDED marker while §1 said it landed): FIXED. The §5 Slice 1 row now records LANDED with its receipts (#6475, tier-2 band #6566, operator-signed as 6-shell-slice1) and states that For/While refuse BY DESIGN rather than reading as unfinished work. That inconsistency was exactly the failure mode this PR exists to fix. Finding 2 (roadmap_authority.dag 6-shell-slice2 is done:false while this PR claimed Slice 2 LANDED — two carriers disagreeing, §3): VALID, fixed in the other direction from what the review suggested, for a reason the review did not have: every done:true row in roadmap_authority.dag is wrapped in sign(s: signed(by: "operator", works: true, scope_equivalent: true, as_expected: true)). That is an OPERATOR ATTESTATION. There is no precedent in that carrier for done:true + Unsigned. So "follow the same pattern for Slice 2" would mean forging an operator signature, which I will not do. Instead this doc stops asserting a verdict it has no authority to give: - roadmap_authority.dag is named as THE status authority; this doc must not contradict it. - The Slice 2 row now reports only what tree receipts prove (.github/ fleet-converge.sh = 21 lines; fleet_converge_emit.dag has zero bash fn defs and emits one artifact; EmitArtifactThenThinRun is a live transport arm) and marks the slice WORK OBSERVABLY COMPLETE / SIGN-OFF PENDING. - FLAGs 2a(i)/2b/2c named in the roadmap row are not resolvable from tree receipts, so the verdict is explicitly left to the operator. - §1 is reworded to match, so the two sections no longer disagree either. Also noted: the "~275 lines / 12+ fn defs" fleet_converge_emit row in the residual census is stale against the current emitter (same class of staleness this PR fixes). docs/plans/shell-emission-model.md regenerated (PASS main_wet). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Brian Searls <briansrls@gunb.ai> Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
6 tasks
briansrls
added a commit
that referenced
this pull request
Jul 26, 2026
ci_floor_disposition_marker_init_emit_matches_concat_golden_holds failed on CI because the concat-reconstructed golden carried a trailing newline after the echo redirect that orch_emit_pipeline does not emit. Switch to the #6467 hand-authored golden pattern (matches committed ci.yml). Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
added a commit
that referenced
this pull request
Jul 26, 2026
ci_floor_disposition_marker_init_emit_matches_concat_golden_holds failed on CI because the concat-reconstructed golden carried a trailing newline after the echo redirect that orch_emit_pipeline does not emit. Switch to the #6467 hand-authored golden pattern (matches committed ci.yml). Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
added a commit
that referenced
this pull request
Jul 26, 2026
…ker (#7265) * WIP: shell→dag bucket D — foreign-executor emit cluster (gate #7216 now MERGE * WIP: shell→dag bucket D — foreign-executor emit cluster (gate #7216 now MERGE * Fix sccache emit if-branch semantics and commit regenerated ci.yml. Multi-statement if/then bodies must be semicolon-joined in a single Run (the if-else grammar inserts then_body verbatim without braces). The heal job failed because ci.yml drifted and the bot lacks workflows permission to push workflow files — commit the regenerated ci.yml and falsifier.yml here. Co-authored-by: Cursor <cursoragent@cursor.com> * Fold bucket-D census into arc census; delete orphan doc. The standalone bucket-d-foreign-executor-emit-census.md had zero inbound refs and would red the falsifier after merge. True-up §4.E/4.I and add §4.J punch-list to the existing shell-to-dag-residual census (reachable via design_document.dag) — single authority, no bind row needed. Co-authored-by: Cursor <cursoragent@cursor.com> * Repoint ci_sccache_opportunistic_detect scaffold to emit module. ci_sccache_provider_shell_injection moved to v2.workflow.ci_materialization_emit in this PR; host_build_cache_provision's scaffold bind still pointed at the deleted gunbc.ci_materialization declaration (review 43048). Co-authored-by: Cursor <cursoragent@cursor.com> * Fix merge-admission emit golden: hand-authored bytes, no trailing NL. ci_floor_disposition_marker_init_emit_matches_concat_golden_holds failed on CI because the concat-reconstructed golden carried a trailing newline after the echo redirect that orch_emit_pipeline does not emit. Switch to the #6467 hand-authored golden pattern (matches committed ci.yml). Co-authored-by: Cursor <cursoragent@cursor.com> * Mark ci_sccache_bound_branch semijoin as declared Scaffold. Three bound-branch Runs are semicolon-folded because multi-Do then_ inside realize_if_else inline binding is unguarded by the if-band corpus and byte-diverged on first ci.yml regen — not a carriage bug. Dissolution: emit-lane construction refusal for multi-step then_ at inline if_else bind, or block-bodied if_else row (emit lane, not D). Co-authored-by: Cursor <cursoragent@cursor.com> * Add missing Present import to ci_materialization_emit. Present is used for redirect and else_ arms but was not in the import list; golden tests executed green via compilation (review 43241). Co-authored-by: Cursor <cursoragent@cursor.com> * Roster ct_render_rust_applied_type_qualified_base_test for scaffold index. #7269 added the hand assertion blob after #7272 landed the inventory witness; merge main exposed compiler_tests_rust_blobs_are_all_rostered red (27 declared, 26 rostered). Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: shell→dag bucket D — foreign-executor emit cluster (gate #7216 now MERGE * Rework bucket D PR1: dissolve semijoin, extract shared emit plumbing. Replace semicolon-folded sccache bound branch with three Do steps so block-bodied if_else (#7277) emits multi-line workflow text; re-golden and regen ci.yml/falsifier.yml. Route both new emit modules through orchestration_bash_emit_support (run/do/emit_pipeline); pre-existing ci_*_emit forks migrate in PR2. Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: shell→dag bucket D — foreign-executor emit cluster (gate #7216 now MERGE * Fix CI drift gate: regen ci.yml from emit authority; drop duplicate roster. Remove duplicate ct_render_rust_applied_type_qualified_base_test row (main already carries it after #7272/#7288); revert language_source_scaffold_index to main placement. Regenerate ci.yml/falsifier.yml via main_wet — block if then/else body lines use emit indentation (10-space), not hand-indented 12-space; matches heal job 89806546636 output the bot could not push. Co-authored-by: Cursor <cursoragent@cursor.com> * WIP: shell→dag bucket D — foreign-executor emit cluster (gate #7216 now MERGE * Remove duplicate ct_render_rust roster row from #7265 merge. Main already owns this entry via #7288; the branch copy made rostered exceed declared (27/28). Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Brian Searls <briansrls@gunb.ai> Co-authored-by: Cursor <cursoragent@cursor.com>
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.
Auto-opened by session-dashboard for session
witty-ibex-317.Pushing to
session/witty-ibex-317advances 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