Repository navigation
Link the effect-namespace-grants design doc (fix main-red doc-orphan wall) - #6821
Merged
Merged
Conversation
…wall) docs/plans/effect-namespace-grants.md (#6811 effects-decomposition design anchor) landed with zero inbound references, so doc_graph_has_no_orphan_docs reds on main and every open PR inherits it (branch-merged-with-main CI). Fix = one inbound reference via the design_document.dag open-threads authority (the established pattern — every docs/plans design doc is linked from an open-thread bullet), faithful to the doc's framing (dissolve Hermetic | Wet into frame × verb × subtree grants over namespaces). DESIGN.md regenerated byte-exact (expected_design_md() drift gate green). Green by execution: doc_graph_has_no_orphan_docs + doc_graph_has_no_dangling_links PASS. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
gunbai-bot
Bot
changed the title
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Jul 17, 2026
Resolve design_document.dag conflict keeping both open-thread bullets; regenerate DESIGN.md from expected_design_md(). Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
added a commit
that referenced
this pull request
Jul 17, 2026
…IGN anchor (#6821/#6822) The doc-orphan was fixed canonically on main: #6821 adds a dedicated DESIGN.md bullet framing the effect-namespace-grants model (in design_document.dag), and #6822 binds the doc's provenance in host_effect_orchestration.dag. My earlier ROADMAP link (via roadmap_authority.dag) was only a workaround for the same orphan and is now a redundant second reachability path (DESIGN 2), dragging generated-artifact regen churn (ROADMAP.md + live-deploy-srv1-apply.sh) into what should be a clean P-A model PR. Reverted those three files to main. Net PR is now just the P-A model: dag/std/effect_grant.dag + its witness. The grants doc stays reachable via the DESIGN anchor. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
Jul 17, 2026
…n_document.dag open-threads conflict #6821 added the effect-namespace-grants open-thread bullet in the same open_threads_blocks() list where #6815 added the shell → intent bullet. Both are distinct threads — keep both li() rows (effect-grants then shell → intent) and render both in DESIGN.md in matching order. Resolved in the design_document.dag source first; DESIGN.md reconstructed to match (drift gate is the oracle). Doc-reachability verified orphans=0.
briansrls
added a commit
that referenced
this pull request
Jul 17, 2026
briansrls
added a commit
that referenced
this pull request
Jul 17, 2026
Merge origin/main (#6821 doc-orphan fix). Delete byte-identical h2/p/li/ul/cell/row copies from wiring_liveness_preflight and design_document (keep unique h1/rule); global-bare resolves bare refs to gunbc.plans.md_helpers sole authority (-465 residue bucket). Co-authored-by: Cursor <cursoragent@cursor.com>
gunbai-bot Bot
pushed a commit
that referenced
this pull request
Jul 17, 2026
Co-authored-by: Cursor <cursoragent@cursor.com>
This was referenced Jul 17, 2026
briansrls
added a commit
that referenced
this pull request
Jul 17, 2026
…admissibility (#6817) * WIP: effects rework * WIP: effects rework * WIP: effects rework * WIP: effects rework * Fix ROADMAP orphan the right way: link grants doc in roadmap_authority (the generator), regenerate The batch-4 generated_artifact_drift_gate was red because the prior orphan fix hand-edited ROADMAP.md — a GENERATED artifact (RoadmapArtifact => expected_roadmap_md() in gunbc.roadmap_authority). A hand-edit to a derived projection is exactly what the drift gate refuses (DESIGN 3: single authority, no dual representation). This masked on main because main dies at batch-2 (doc orphan) before reaching batch-4; fixing the orphan exposed the drift. Correct fix: add the effect-namespace-grants link to the `5-defork` entry in dag/gunbc/roadmap_authority.dag (the single authority), then regenerate the derived artifacts. ROADMAP.md and .github/live-deploy-srv1-apply.sh (which embeds ROADMAP content) are regenerated from that authority. Verified green by execution: - generated_artifact_gate main -> ExitSuccess (all 14 registered artifacts) - doc_graph_has_no_orphan_docs -> true (grants doc reachable from ROADMAP.md) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Drop redundant ROADMAP orphan-link workaround; rely on operator's DESIGN anchor (#6821/#6822) The doc-orphan was fixed canonically on main: #6821 adds a dedicated DESIGN.md bullet framing the effect-namespace-grants model (in design_document.dag), and #6822 binds the doc's provenance in host_effect_orchestration.dag. My earlier ROADMAP link (via roadmap_authority.dag) was only a workaround for the same orphan and is now a redundant second reachability path (DESIGN 2), dragging generated-artifact regen churn (ROADMAP.md + live-deploy-srv1-apply.sh) into what should be a clean P-A model PR. Reverted those three files to main. Net PR is now just the P-A model: dag/std/effect_grant.dag + its witness. The grants doc stays reachable via the DESIGN anchor. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Brian Searls <briansrls@gunb.ai> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
Jul 17, 2026
Merge origin/main (#6821 doc-orphan fix). Delete byte-identical h2/p/li/ul/cell/row copies from wiring_liveness_preflight and design_document (keep unique h1/rule); global-bare resolves bare refs to gunbc.plans.md_helpers sole authority (-465 residue bucket). Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
added a commit
that referenced
this pull request
Jul 17, 2026
Merge origin/main (#6821 doc-orphan fix). Delete byte-identical h2/p/li/ul/cell/row copies from wiring_liveness_preflight and design_document (keep unique h1/rule); global-bare resolves bare refs to gunbc.plans.md_helpers sole authority (-465 residue bucket). Co-authored-by: Cursor <cursoragent@cursor.com>
briansrls
pushed a commit
that referenced
this pull request
Jul 18, 2026
…han wall + drift gate) (#6828) #6824 landed the design_document.dag source edit (5 orphan-doc links) but the resolve-heavy DESIGN.md regen had not finished, so main went drift-red AND still orphan-red. This is the missing regen: DESIGN.md now links rc-ownership-wrap-decision-design, witness-realization-plan, v2-memory-control-audit, machine-shape-orthogonal-scheduling (effect-namespace-grants already linked via #6821). Generated via generated_artifact_gate main_wet; design_document.dag unchanged since #6824 so this is pure regen, no source re-litigation. Co-authored-by: Brian Searls <briansearls1@gmail.com> Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
briansrls
added a commit
that referenced
this pull request
Jul 18, 2026
* docs: anchor the shell → intent rescope (§3/§4) in DESIGN.md + plan doc Restates §3/§4 for the shell case: the .dag intent may not name a target language. A workflow is an ordinary dependency graph over modeled operations; rendering to bash is a separate, target-parametrized concern (emit), distinct from realization (pure-spec → host-effect). Bash appears only in its grammar spec + emit rows, never in the intent — exactly where Rust appears. Enforceable invariant: the intent imports no language-construction vocab (bash_build, the ShellStmt coproduct, serialize_bash) — realization_vocabulary_ containment generalized with the intent layer in scope, green = conforms. Names the residual the operator flagged 2026-07-17: (1) raw concat shell (gunbc_ci_deploy_invoke, untouched by the bash_build migration) and (2) structured-but-still-bash (floor_diff_observe). Phases the rescope with the sidecar delete demoted to Phase 0; git.diff/floor_diff_observe as the flagship. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs: fill the shell-in-intent residual census (2026-07-17) ~110 raw-concat shell builders + ~9 structured-bash sites across ~50 files, concentrated in dag/gunbc (std/lens/compiler are clean). Collapses to ~10 operation families to model once each — not 110 independent problems. Adds the per-layer/per-operation breakdown, per-file hotspots, and the §3 policy note on the extdeps extern-tool wrappers (legitimate as one-of-N transport handlers). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * WIP: shell -> dag * WIP: shell -> dag * docs: restore generated DESIGN.md anchor + fix census class label DESIGN.md is generated from dag/gunbc/design_document.dag (DesignArtifact); the anchor lives in the source's open_threads_blocks() li and DESIGN.md is its rendered output. An auto-commit had reverted the generated DESIGN.md, leaving source-with-anchor vs generated-without = drift. Restores the rendered bullet so committed DESIGN.md == expected_design_md() again (review 39352 finding 1). Plan doc: relabel "Class 1 + 3" to "Class 1" — the census fragment-helper sub-kind is a raw-string builder (Class 1), not a separate undefined Class 3 (review 39352 finding 2). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * Merge origin/main into shell-intent-design: resolve DESIGN.md + design_document.dag open-threads conflict #6821 added the effect-namespace-grants open-thread bullet in the same open_threads_blocks() list where #6815 added the shell → intent bullet. Both are distinct threads — keep both li() rows (effect-grants then shell → intent) and render both in DESIGN.md in matching order. Resolved in the design_document.dag source first; DESIGN.md reconstructed to match (drift gate is the oracle). Doc-reachability verified orphans=0. * WIP: shell -> dag * WIP: shell -> dag --------- Co-authored-by: Brian Searls <briansrls@gunb.ai> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.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.
What
Fixes the main-red
doc_graph_has_no_orphan_docsgate.docs/plans/effect-namespace-grants.md(the #6811 effects-decomposition design anchor) landed with zero inbound references, so the orphan wall reds onmainitself and every open PR inherits it (CI tests branch-merged-with-main). This blocks the fleet — including the merge-ready #6814.Fix
One inbound reference via the
design_document.dagopen-threads authority — the established pattern (everydocs/plans/*.mddesign doc is linked from an open-thread bullet). The bullet is faithful to the doc's framing: dissolveHermetic | Wetinto (frame × verb × subtree) grants over namespaces.DESIGN.mdregenerated byte-exact (expected_design_md()drift gate green;README.md/CLAUDE.mdsymlinks intact).Green by execution
claim_batch --entry dag/test/claim/doc_reachability_witness_test.dag:doc_graph_has_no_orphan_docsPASS (the orphan is now referenced)doc_graph_has_no_dangling_linksPASS (the link resolves to a real file)doc_graph_universe_is_nonemptyPASSScope:
design_document.dag+ regeneratedDESIGN.mdonly. No code paths touched.🤖 Generated with Claude Code