SymbolIndex / type-env-single-authority — the containment-tree naming authority (dissolves v1 global_bare census, peel guard, sigs walk, and the total-census memory wall). Owned under namespace-migration lane. by briansrls · Pull Request #6809 · gunb-ai/gunbc

briansrls · 2026-07-17T18:54:16Z

What

Folds build_global_bare_census into SymbolIndex.global_bare, removing the TypeEnv.global_bare dual-representation. The corpus-wide bare-name census now lives on the single SymbolIndex authority (built once by the topo prepass), not a second parallel map.

Why

§2/§3 single-authority: SymbolIndex is the qualified-name authority (Rule 1 forbids a second index). The standalone global_bare census was a dual representation of naming facts. This consolidation lands the settled substrate that the namespace-only-Y resolver flip sits on (design: SymbolIndex lands first, then namespace-only-Y is a policy layer over it — docs/plans/namespace-resolution-design.md §7.5/§8).

Behavior (byte-identical, fail-closed)

The global_bare population is byte-identical to the deleted census — deliberately behavior-preserving:

decl-only tracking (type/fn/data names via symbol_index_insert_decl / local_binding_for_item);
unique Disj variant aliases land in SymbolIndex.entries but do NOT enter global_bare (unchanged);
lookup_binding_by_name consults global_bare only after str_bindings / ancestry_str_bindings / intern+bindings all miss;
only GlobalBareUniqueBinding resolves; GlobalBareAmbiguousBinding stays Absent — fail-closed, never guesses (§5).

Load-bearing — needs operator sign-off

Touches 04_env.dag / 04_infer.dag / 04_resolve.dag (typecheck signatures thread symbol_index through build_type_env / typecheck_module). Per session policy these pipeline stages require operator sign-off before merge.

Test plan

Equivalence receipt (byte-identical census population on live floor corpus):

cargo test -p v1-compiler-tests global_bare --release — 6/6 PASS (~48s)
- global_bare_corpus_census_test::tier2_namespace_homonym_invariants — tier-2 triage invariants hold on SymbolIndex.global_bare built by build_symbol_index_census (dag/ + src/v2 floor roots)
- global_bare_corpus_census_test::tier2_census_disposition_snapshot — census leg refuses ABSENT names (shape-coverage gap unchanged)
- global_bare_corpus_census_test::corpus_census_state_of_failing_names — live census histogram over dag/ corpus
- global_bare_corpus_census_test::corpus_ambiguous_roster — AMBIGUOUS roster emitted from floor source-roots
- type_env_scope_chain_test::global_bare_fallback_resolves_when_corpus_wide_unique — lookup_binding_by_name resolves via symbol_index.global_bare when corpus-wide unique
- type_env_scope_chain_test::global_bare_fallback_stays_absent_when_corpus_wide_ambiguous — ambiguous stays Absent (fail-closed)

Resolver / type-env unit tests:

cargo test -p v1-compiler-tests type_env_scope_chain --release — 13/13 PASS (~1.2s) — import-chain scaling, str_bindings index, RC identity, local shadowing (all on updated build_type_env / symbol_index threading)

Self-host (no reconcile escape hatch):

regen_stage0 --verify — PASS (regen_divergence_count=0; fresh emit matches committed stage0 without post-process overwrite)
cargo test -p v1-compiler emitted_compiler_tests_has_no_stale_global_bare_arg --release — 1/1 PASS (~29s) — fresh compile_stage0 emit of compiler_tests.rs does not carry the removed global_bare argument

CI:

build — PASS (post-merge with origin/main @ Link the effect-namespace-grants design doc (fix main-red doc-orphan wall) #6821 orphan-doc fix)
ci — pending on 93fa7a92b

Review fixes landed (93fa7a92b):

Removed reconcile_compiler_tests_generated_source silent overwrite; emit is clean by execution
Dropped accidental doc_graph_orphan_inventory_for_ci_debug test; restored reachable_set_flags_orphan_node #[test]

Body filled by quiet-gull-833 (namespace-migration lane owner) to unblock review — #6809 is the critical-path substrate for finishing the migration. Test plan section completed by clever-ram-104 (owner).

…aming au

Remove the deleted global_bare parameter from compiler_tests profiling harness and its .dag source so fmt and compile stay green. Co-authored-by: Cursor <cursoragent@cursor.com>

…aming au

cursor · 2026-07-17T19:40:46Z

Bugbot is not enabled for your account, so this pull request was not reviewed.

Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs.

Split symbol_index_insert_decl for decl tracking; variant aliases use entry-only insert so global_bare stays byte-identical to the deleted build_global_bare_census population. Co-authored-by: Cursor <cursoragent@cursor.com>

gunbai-bot · 2026-07-17T19:49:52Z

Response to review 39316 (variant-alias census broadening)

Valid finding on a7cd4c9 — unintended, not a silent widen.

Fix (d2dcac + 6e08d55):

Split symbol_index_insert (entries only) from symbol_index_insert_decl (tracks global_bare via local_binding_for_item binding).
symbol_index_insert_unique_disj_variant_aliases uses entry-only insert — variant leaf names do not enter global_bare.
global_bare_fallback_invariant updated to document decl-only tracking and byte-identity with the deleted build_global_bare_census population.

This was also the root cause of the compile-clean failure on a7cd4c9 (variant names like Accepted/Found/Err polluting global-bare fallback and breaking inference across the corpus).

review 39318: Agreed — consolidation is coherent; the interim regression above is now repaired.

Re: symbol_index_qualified_leaf — kept for now but unused after the insert_decl split; can delete in a follow-on.

— sent from clever-ram-104

…aming au

gunbai-bot · 2026-07-17T20:02:24Z

Response to review 39319 (dead symbol_index_qualified_leaf)

Valid finding — fixed in 67f6e952.

Removed the unused helper from both authorities:

src/v1/04_env.dag (.dag source)
src/v1/stage0/src/v1_compiler_infer_env.rs (Rust mirror)

git grep symbol_index_qualified_leaf is now empty; no callers were ever wired.

…aming au

…ecursive scrutinees The per-module resolve memo (#6779) had two grains sharing one fp cache: a bounded memo (resolve_node_bounded — its main within-module speedup claim) and a scrutinee memo on resolve_scrutinee_type_node keyed by a FULL-subtree surface fingerprint. On a self-recursive scrutinee (e.g. MarkupNode reached via field_summary_for_type recursion) every expansion depth yields a DISTINCT fingerprint key, so the scrutinee grain NEVER hits yet pins every intermediate expanded subtree — measured >33GB / OOM on a 6-module closure and the whole-tree CI floor (the #6640 integration branch's 56m OOM-kill). Its dedup benefit cannot materialize on exactly the recursive case that makes it expensive, so it is pure liability. Removed. The bounded grain is unaffected and keeps #6779's speedup. Verified: whole-tree compile-clean now completes bounded (~1.1GB RSS, ~24s) where it previously OOM'd; per_module_resolve_memo_equivalence_test still green (3/3); workspace builds clean. Residue count unchanged (~1448 — the namespace-migration G1/strip residue, now finally visible because the tree completes). Unblocks #6640 (namespace integration) and #6809 (census consolidation). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

…ecursive scrutinees The per-module resolve memo (#6779) had two grains sharing one fp cache: a bounded memo (resolve_node_bounded — its main within-module speedup claim) and a scrutinee memo on resolve_scrutinee_type_node keyed by a FULL-subtree surface fingerprint. On a self-recursive scrutinee (e.g. MarkupNode reached via field_summary_for_type recursion) every expansion depth yields a DISTINCT fingerprint key, so the scrutinee grain NEVER hits yet pins every intermediate expanded subtree — measured >33GB / OOM on a 6-module closure and the whole-tree CI floor (the #6640 integration branch's 56m OOM-kill). Its dedup benefit cannot materialize on exactly the recursive case that makes it expensive, so it is pure liability. Removed. The bounded grain is unaffected and keeps #6779's speedup. Verified: whole-tree compile-clean now completes bounded (~1.1GB RSS, ~24s) where it previously OOM'd; per_module_resolve_memo_equivalence_test still green (3/3); workspace builds clean. Residue count unchanged (~1448 — the namespace-migration G1/strip residue, now finally visible because the tree completes). Unblocks #6640 (namespace integration) and #6809 (census consolidation). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Co-authored-by: Cursor <cursoragent@cursor.com>

…aming au

…ecursive scrutinees The per-module resolve memo (#6779) had two grains sharing one fp cache: a bounded memo (resolve_node_bounded — its main within-module speedup claim) and a scrutinee memo on resolve_scrutinee_type_node keyed by a FULL-subtree surface fingerprint. On a self-recursive scrutinee (e.g. MarkupNode reached via field_summary_for_type recursion) every expansion depth yields a DISTINCT fingerprint key, so the scrutinee grain NEVER hits yet pins every intermediate expanded subtree — measured >33GB / OOM on a 6-module closure and the whole-tree CI floor (the #6640 integration branch's 56m OOM-kill). Its dedup benefit cannot materialize on exactly the recursive case that makes it expensive, so it is pure liability. Removed. The bounded grain is unaffected and keeps #6779's speedup. Verified: whole-tree compile-clean now completes bounded (~1.1GB RSS, ~24s) where it previously OOM'd; per_module_resolve_memo_equivalence_test still green (3/3); workspace builds clean. Residue count unchanged (~1448 — the namespace-migration G1/strip residue, now finally visible because the tree completes). Unblocks #6640 (namespace integration) and #6809 (census consolidation). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Co-authored-by: Cursor <cursoragent@cursor.com>

…aming au

…ecursive scrutinees The per-module resolve memo (#6779) had two grains sharing one fp cache: a bounded memo (resolve_node_bounded — its main within-module speedup claim) and a scrutinee memo on resolve_scrutinee_type_node keyed by a FULL-subtree surface fingerprint. On a self-recursive scrutinee (e.g. MarkupNode reached via field_summary_for_type recursion) every expansion depth yields a DISTINCT fingerprint key, so the scrutinee grain NEVER hits yet pins every intermediate expanded subtree — measured >33GB / OOM on a 6-module closure and the whole-tree CI floor (the #6640 integration branch's 56m OOM-kill). Its dedup benefit cannot materialize on exactly the recursive case that makes it expensive, so it is pure liability. Removed. The bounded grain is unaffected and keeps #6779's speedup. Verified: whole-tree compile-clean now completes bounded (~1.1GB RSS, ~24s) where it previously OOM'd; per_module_resolve_memo_equivalence_test still green (3/3); workspace builds clean. Residue count unchanged (~1448 — the namespace-migration G1/strip residue, now finally visible because the tree completes). Unblocks #6640 (namespace integration) and #6809 (census consolidation). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Co-authored-by: Cursor <cursoragent@cursor.com>

gunbai-bot · 2026-07-17T22:20:38Z

Addressed review 39382 in 93fa7a92b:

reachable_set_flags_orphan_node test — the accidental #[test] splice from the temporary debug helper is fixed: doc_graph_orphan_inventory_for_ci_debug is removed entirely; reachable_set_flags_orphan_node retains its #[test] attribute and runs under cargo test again.
reconcile_compiler_tests_generated_source escape hatch — deleted. Fresh compile_stage0 emit no longer carries the removed global_bare argument (verified by emitted_compiler_tests_has_no_stale_global_bare_arg and regen_stage0 --verify with regen_divergence_count=0). No silent seed overwrite; self-host is green-by-execution on this module.

PR body Test plan section now has the equivalence-receipt commands and results.

— sent from clever-ram-104

… live_deploy generated artifacts)

…n_wet. Brings in #6824 open-thread rows (wrap-decision, witness-realization, machine-shape) so the design_document emit matches committed DESIGN.md and generated_artifact_drift_gate passes on current main. Co-authored-by: Cursor <cursoragent@cursor.com>

gunbai-bot · 2026-07-17T22:48:57Z

Verified review 39419 against current HEAD (c078b822e): no findings to action.

TypeEnv no longer carries global_bare; census lives on SymbolIndex.global_bare only (04_env.dag)
build_global_bare_census dissolved into build_symbol_index_census; symbol_index_insert_decl tracks decl-only bare names, symbol_index_insert is entries-only for variant aliases — matches global_bare_fallback_invariant
Call sites, seed Rust, and tests (global_bare_corpus_census_test, type_env_scope_chain_test) follow the new shape

No code changes from this review. CI pending on drift-fix push.

— sent from clever-ram-104

gunbai-bot · 2026-07-17T22:49:17Z

Verified review 39420 against current HEAD (c078b822e): APPROVE stands, no code changes needed.

Confirmed:

build_global_bare_census deleted; census built in build_symbol_index_census → SymbolIndex.global_bare
symbol_index_insert_decl / symbol_index_insert split preserves decl-only global_bare population (variant aliases excluded) per global_bare_fallback_invariant
emitted_compiler_tests_has_no_stale_global_bare_arg guards fresh emit (no removed global_bare arg on build_type_env/typecheck_module)
Tests updated: global_bare_corpus_census_test, type_env_scope_chain_test

— sent from clever-ram-104

…ecursive scrutinees The per-module resolve memo (#6779) had two grains sharing one fp cache: a bounded memo (resolve_node_bounded — its main within-module speedup claim) and a scrutinee memo on resolve_scrutinee_type_node keyed by a FULL-subtree surface fingerprint. On a self-recursive scrutinee (e.g. MarkupNode reached via field_summary_for_type recursion) every expansion depth yields a DISTINCT fingerprint key, so the scrutinee grain NEVER hits yet pins every intermediate expanded subtree — measured >33GB / OOM on a 6-module closure and the whole-tree CI floor (the #6640 integration branch's 56m OOM-kill). Its dedup benefit cannot materialize on exactly the recursive case that makes it expensive, so it is pure liability. Removed. The bounded grain is unaffected and keeps #6779's speedup. Verified: whole-tree compile-clean now completes bounded (~1.1GB RSS, ~24s) where it previously OOM'd; per_module_resolve_memo_equivalence_test still green (3/3); workspace builds clean. Residue count unchanged (~1448 — the namespace-migration G1/strip residue, now finally visible because the tree completes). Unblocks #6640 (namespace integration) and #6809 (census consolidation). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Co-authored-by: Cursor <cursoragent@cursor.com>

Restore main's infer_lookup/env/resolve/sigs and emit after merge conflict resolution; pass source_roots to enable_floor_compile_clean_lazy_install. Co-authored-by: Cursor <cursoragent@cursor.com>

Rebase onto main @ c0cacc4 (#6809 SymbolIndex). Atomic Wave-1 import-strip from collapsed PRs #6603/#6600/#6607/#6610; stage0 aligned to #6809 infer_lookup/env/resolve/sigs authority. Residue ~984 until namespace-only-Y flip (quiet-gull). Co-authored-by: Cursor <cursoragent@cursor.com>

…Var fallback - Register unique Disj variants into global_bare (symbol_index_insert_decl, corpus-ambiguity fail-closed) — was entries-only/qualified per #6809 - ExprVar path consults global_bare census when locals + func_sig miss - Mirrored into stage0 seed - Whole-tree compile-clean residue 4847 -> 1583 (baseline 9310, cumulative -83%) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

…e dispatch in as the fourth consumer Two updates to the namespace-only resolution design. 7.5 sequencing: SymbolIndex LANDED (#6809, dag/std/symbol_index.dag). The loyal-heron scaling gate this lane was waiting on is discharged; the doc still described it as pending. namespace-only is now a policy layer over a settled, in-tree index. New 10: the fork is three-way, not two. Proven by instrumented probe (reading lookup_resolved_sig's return value directly, because the compiler rejects neither label nor return-type mismatches for this shape, so absent diagnostics carry no information): a module importing twin_sig from twin_p binds twin_q's signature. func_env.parents records WHICH MODULES are imported, never the selective name list, and its order is incidental. So neither typecheck nor runtime honors the import list -- 2.1's dual-representation argument is empirically confirmed and stronger than stated: imports are a DEAD authority, not a redundant working one. Folding dispatch in converts component 1 of the bare-name dispatch fork from 'build a fourth resolver' into 'delete a fork': fn_nodes dissolves rather than being re-keyed, GlobalBareLookupAmbiguous is already the 2.4 typed refusal, and dispatch stays one map hit (on QualifiedName, not a per-call ancestor walk) -- the claim most worth falsifying before implementation. Records why no standalone fix is defensible, and the counted-ledger interim if the flip is not immediate.

…nce census spec (docs only) (#6933) * WIP: collapse the fork * collapse-the-fork: remove exploratory homonym probe fixtures (design review deliverable; repro recipe in report, not landed ahead of design sign-off) * WIP: collapse the fork * WIP: collapse the fork * WIP: collapse the fork * revert stray dag/gunbc contamination: a mid-turn auto-commit captured a conflicted git-stash pop The worktrees share one git stash stack; a concurrent session (deep-wren-620) pushed a stash during a long corpus run, so a pathspec-limited 'stash pop' here popped THEIR entry and left conflict markers in six dag/gunbc/*.dag files, which an auto-commit then captured (ec0afe1). Those files are restored to their origin/main content; the other session's stash entry is intact and untouched. No dag/gunbc change was ever intended by this branch. * WIP: collapse the fork * WIP: collapse the fork * WIP: collapse the fork * add discriminating witness for the application-site contract wall; drop the qualified-dispatch scaffold * restore stray #[cfg(test)] above cross_representation_equality_test (review 39978) * WIP: collapse the fork * WIP: collapse the fork * remove Q1 sigprobe fixtures (inconclusive investigation, not a landable artifact) * WIP: collapse the fork * WIP: collapse the fork * namespace lane: record that the import list is inert, and fold runtime dispatch in as the fourth consumer Two updates to the namespace-only resolution design. 7.5 sequencing: SymbolIndex LANDED (#6809, dag/std/symbol_index.dag). The loyal-heron scaling gate this lane was waiting on is discharged; the doc still described it as pending. namespace-only is now a policy layer over a settled, in-tree index. New 10: the fork is three-way, not two. Proven by instrumented probe (reading lookup_resolved_sig's return value directly, because the compiler rejects neither label nor return-type mismatches for this shape, so absent diagnostics carry no information): a module importing twin_sig from twin_p binds twin_q's signature. func_env.parents records WHICH MODULES are imported, never the selective name list, and its order is incidental. So neither typecheck nor runtime honors the import list -- 2.1's dual-representation argument is empirically confirmed and stronger than stated: imports are a DEAD authority, not a redundant working one. Folding dispatch in converts component 1 of the bare-name dispatch fork from 'build a fourth resolver' into 'delete a fork': fn_nodes dissolves rather than being re-keyed, GlobalBareLookupAmbiguous is already the 2.4 typed refusal, and dispatch stays one map hit (on QualifiedName, not a per-call ancestor walk) -- the claim most worth falsifying before implementation. Records why no standalone fix is defensible, and the counted-ledger interim if the flip is not immediate. * remove Q1 instrumented-probe scaffolding (throwaway investigation; finding recorded in namespace-resolution-design.md 10.1) * namespace lane 10.4: spec the divergence census as the next step (operator-directed) * fix review 40267: wrong SymbolIndex carrier path, and drop the racing consumer ordinal --------- Co-authored-by: Brian Searls <briansearls1@gmail.com>

…uling (#6973) Folds the operator ruling of 2026-07-21 into namespace-resolution-design.md as §13. Semantics are ratified (operator, in-session: "fail loudly" / "full path required at all times + users can alias" / "familiar frontend, very strict/ precise underneath" / "this looks good"); wording is the lane's draft, operator reviews here. The rule: resolve a reference's first segment to the UNIQUE binder on its ancestor chain — zero = loud Unresolved, 2+ = located typed Ambiguous with the full candidate list + fix menu, exactly one = project downward. No nearest-wins; shadowing is a refusal (Elm precedent); refusal at the reference site. Amends §3/§6 ("exactly one nearest" -> "exactly one on the chain; multiple = refusal"). Invariant: no edit elsewhere can silently CHANGE what an existing reference means — only loudly break it. Fallback chains (nearest-wins / global-unique-fallback / first-hit / silent-pick) rejected as a class; uniqueness replaces priority order (Rule 2 on the ancestor chain). Also captures: aliases as ordinary binding nodes (import->alias transmutation, source of truth = the walk's resolved target, #6936 buckets); global_bare dies as a resolution mechanism (migration-oracle only, then delete); builtins bind at root (dissolves the prelude-shaped neither_bound class); the (Y) expected-type filter FLAGGED not ruled (adjacent to the rejected fallback class, needs its own ruling); and sequencing (§10 step-1 in flight #6968; transmutation lands with-or-before the strict flip, never imports-first). Relayed from sunny-wolf-225 (fix governor); gates Phase 1 dispatch. Sibling to the DESIGN.md #6809 gate-discharge doc fix (#6969), kept separate to preserve that PR's approvals. Co-authored-by: Brian Searls <briansrls@gunb.ai> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

…te is discharged (#6969) The namespace-only-resolution open-thread bullet still read "gated on loyal-heron's scaling receipt before any resolver surgery", but namespace-resolution-design.md §7.5 records that gate as DISCHARGED at #6809 ("the lane is no longer blocked on it"). This is a summary-to-authority alignment, not a design change: the open-thread summary was stale relative to its own authority doc, and was mis-gating readers into believing resolver surgery is still blocked on the scaling receipt when the design doc says it is not. Both surfaces updated in lockstep to avoid a §3 dual-representation drift: the DESIGN.md bullet and its mirror in dag/gunbc/design_document.dag (li text) carry the identical correction. Rationale flagged for operator review per sunny-wolf-225's authorization; operator can veto in review if the gate is considered live for other reasons. Co-authored-by: Brian Searls <briansrls@gunb.ai> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Brian Searls and others added 5 commits July 17, 2026 18:54

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

e29f3ef

…aming au

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

f459633

…aming au

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

000c79d

…aming au

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

4d27851

…aming au

Fix build_type_env call sites after global_bare folded into SymbolIndex.

a7cd4c9

Remove the deleted global_bare parameter from compiler_tests profiling harness and its .dag source so fmt and compile stay green. Co-authored-by: Cursor <cursoragent@cursor.com>

gunbai-bot Bot marked this pull request as ready for review July 17, 2026 19:30

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

d2dcac1

…aming au

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

67f6e95

…aming au

Brian Searls added 2 commits July 17, 2026 20:19

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

09759b5

…aming au

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

479f77e

…aming au

gunbai-bot Bot mentioned this pull request Jul 17, 2026

fix(infer): remove #6779 scrutinee resolve-memo grain — whole-tree compile-clean OOM #6813

Closed

Brian Searls added 4 commits July 17, 2026 20:49

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

13fd412

…aming au

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

9078dd4

…aming au

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

32fa589

…aming au

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

fe6ba1f

…aming au

Brian Searls added 4 commits July 17, 2026 21:44

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

cf7b558

…aming au

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

006eb69

…aming au

Merge remote-tracking branch 'origin/main' into session/clever-ram-104

57d7578

WIP: SymbolIndex / type-env-single-authority — the containment-tree n…

93fa7a9

…aming au

Brian Searls and others added 2 commits July 17, 2026 22:29

Merge origin/main into session/clever-ram-104 (pick up effect_grant +…

5646913

… live_deploy generated artifacts)

briansrls merged commit c0cacc4 into main Jul 18, 2026
3 checks passed

briansrls deleted the session/clever-ram-104 branch July 18, 2026 01:46

gunbai-bot Bot mentioned this pull request Jul 20, 2026

Namespace lane: runtime dispatch is the fourth consumer + the divergence census spec (docs only) #6933

Merged

This was referenced Jul 21, 2026

docs: align DESIGN.md open-thread with #6809 (loyal-heron scaling gate discharged) #6969

Merged

docs(namespace): §13 — operator-ratified unique-on-chain resolution ruling #6973

Merged

Navigation Menu

Repository navigation

SymbolIndex / type-env-single-authority — the containment-tree naming authority (dissolves v1 global_bare census, peel guard, sigs walk, and the total-census memory wall). Owned under namespace-migration lane. - #6809

SymbolIndex / type-env-single-authority — the containment-tree naming authority (dissolves v1 global_bare census, peel guard, sigs walk, and the total-census memory wall). Owned under namespace-migration lane.#6809
briansrls merged 20 commits into
mainfrom
session/clever-ram-104

briansrls commented Jul 17, 2026 •

edited by gunbai-bot Bot

Loading

Uh oh!

cursor Bot commented Jul 17, 2026

Uh oh!

gunbai-bot Bot commented Jul 17, 2026

Uh oh!

gunbai-bot Bot commented Jul 17, 2026

Uh oh!

gunbai-bot Bot commented Jul 17, 2026

Uh oh!

gunbai-bot Bot commented Jul 17, 2026

Uh oh!

gunbai-bot Bot commented Jul 17, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

briansrls commented Jul 17, 2026 • edited by gunbai-bot Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What

Why

Behavior (byte-identical, fail-closed)

Load-bearing — needs operator sign-off

Test plan

Uh oh!

cursor Bot commented Jul 17, 2026

Uh oh!

gunbai-bot Bot commented Jul 17, 2026

Uh oh!

gunbai-bot Bot commented Jul 17, 2026

Uh oh!

gunbai-bot Bot commented Jul 17, 2026

Uh oh!

gunbai-bot Bot commented Jul 17, 2026

Uh oh!

gunbai-bot Bot commented Jul 17, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

briansrls commented Jul 17, 2026 •

edited by gunbai-bot Bot

Loading