heavy-duty/ceremony
7
0
Fork 2
Code Issues 1 Pull requests Projects Releases 8 Packages Wiki Activity Actions

actions/issueflow-reconcile + RELEASES.md + TRIAGE.md — membership record, carrier-gate fixes, CommonMark row parsing (upstream-0.6.3) #230

New issue
Closed
opened 2026-08-17 22:26:58 +00:00 by claude-bot-andresmgsl · 10 comments
claude-bot-andresmgsl commented 2026-08-17 22:26:58 +00:00
Member
Copy link

Context

Part of #228 (phase 2). Upstream-0.6.3 is the only code release in the
range: the issueflow reconciler learns a membership record. Both the
script and its test have diverged on this forge (Forgejo port), so this is
a logic port onto the forge files — upstream bytes will not apply.

Upstream source: tag upstream-0.6.3, commits ce67044 0bd30ec
f41c414 78199a5 65168fe 5fccb7e 056afba 449a9f1 0a92866
4291447 afe1ae7 6a6af8c 389979a (github.com/heavy-duty/ceremony,
read-only). Upstream's own changelog phrasing is the spec's source of
truth; their test additions (+306 lines in
test/issueflow-reconcile.test.sh) define the expected behavior.

Spec

  1. Membership record (upstream #343): a release issue records its
    window membership under a ## Members heading, read by heading and one
    bare #N per row; its Blocked by line answers the predecessor gate
    and nothing else. The standing-window decision and the non-member flag
    read that record with NO fallback to the gate: a release issue that
    enumerates no membership stands no window and draws no flag.
  2. Carrier-gate fixes (upstream #327): release-window carriers are
    excluded from their own gates; stale board-flag claims are suppressed
    after an issue pass changes queue state; the claimable-board-flag
    predicate is shared, not duplicated.
  3. CommonMark row parsing: a membership row is indented at most three
    spaces (not any run); an ordered marker is CommonMark's nine digits (not
    any run); a code block and a sub-row are different non-rows; every
    Markdown list marker opens a membership row. Pin the marker class and
    the record terminator in tests.
  4. Docs sections that belong to this change (upstream 056afba,
    449a9f1): membership gets its own record section in RELEASES.md and
    TRIAGE.md, plus the changelog fragment.
  5. Port the upstream test additions into the forge's test file, adapted to
    the forge harness.

Acceptance criteria

  • All ported membership/gate/parser behaviors covered by tests in
    test/issueflow-reconcile.test.sh, passing on this forge's runner.
  • jq programs version-portable (the runner's jq rejects $label-style
    reserved words that pass locally — the !23 lesson; check every new
    program).
  • shellcheck clean (use the chunked helper — full-tree shellcheck
    OOM-kills 7.7 GiB boxes).
  • RELEASES.md / TRIAGE.md membership sections present and consistent
    with the docs child's landed text.

Baseline note (triage, corrected 2026-08-23): main runs 44/44 on
test/labels.test.sh at f69224c — verified by running it. The 43/44
baseline recorded here on 2026-08-22 is spent: its sole failure,
an identity in the table but not the conf goes red, was the :249 fixture
#232 owned, and #232 landed 2026-08-23 as f69224c (!237). So the bar for this
issue is a full green suite, with no known failure to name.

Dependencies

Part of #228. Unblocked 2026-08-22 — the doctrine-docs child #229 landed
on main as 4f887a7 (!233 merged 22:16:23Z), so the RELEASES.md / TRIAGE.md
sequencing that held this issue is satisfied: build from current main, where
RELEASES.md already carries the post-merge gate-member rules and the holding
invariant this child extends, and add the membership sections beside them.

Blocks #231 (the release child) and #234 (collision edge — #234 changes the
same two reconciler files and waits on this issue closing). No collision edge
of its own: the other open non-epic issues touch disjoint surfaces — #235 is
actions/labels-reconcile, #236 and #240 are lib/forge-forgejo.sh and
test/forge-backends.test.sh, #238 is those two plus the other backend — so all
of them stay concurrently claimable. (#232, one string in test/labels.test.sh,
was on this list until it landed 2026-08-23.) No
release-window membership call applies: the 0.6.2 window is stood down (#231
carries no release label; the lead stood it down 2026-08-17T23:33Z).

## Context Part of #228 (phase 2). Upstream-0.6.3 is the only code release in the range: the issueflow reconciler learns a **membership record**. Both the script and its test have diverged on this forge (Forgejo port), so this is a logic port onto the forge files — upstream bytes will not apply. Upstream source: tag `upstream-0.6.3`, commits `ce67044` `0bd30ec` `f41c414` `78199a5` `65168fe` `5fccb7e` `056afba` `449a9f1` `0a92866` `4291447` `afe1ae7` `6a6af8c` `389979a` (github.com/heavy-duty/ceremony, read-only). Upstream's own changelog phrasing is the spec's source of truth; their test additions (+306 lines in `test/issueflow-reconcile.test.sh`) define the expected behavior. ## Spec 1. **Membership record** (upstream #343): a release issue records its window membership under a `## Members` heading, read by heading and one bare `#N` per row; its `Blocked by` line answers the predecessor gate and nothing else. The standing-window decision and the non-member flag read that record with NO fallback to the gate: a release issue that enumerates no membership stands no window and draws no flag. 2. **Carrier-gate fixes** (upstream #327): release-window carriers are excluded from their own gates; stale board-flag claims are suppressed after an issue pass changes queue state; the claimable-board-flag predicate is shared, not duplicated. 3. **CommonMark row parsing**: a membership row is indented at most three spaces (not any run); an ordered marker is CommonMark's nine digits (not any run); a code block and a sub-row are different non-rows; every Markdown list marker opens a membership row. Pin the marker class and the record terminator in tests. 4. **Docs sections that belong to this change** (upstream `056afba`, `449a9f1`): membership gets its own record section in `RELEASES.md` and `TRIAGE.md`, plus the changelog fragment. 5. Port the upstream test additions into the forge's test file, adapted to the forge harness. ## Acceptance criteria - [x] All ported membership/gate/parser behaviors covered by tests in `test/issueflow-reconcile.test.sh`, passing on this forge's runner. - [x] jq programs version-portable (the runner's jq rejects `$label`-style reserved words that pass locally — the !23 lesson; check every new program). - [x] shellcheck clean (use the chunked helper — full-tree shellcheck OOM-kills 7.7 GiB boxes). - [x] RELEASES.md / TRIAGE.md membership sections present and consistent with the docs child's landed text. Baseline note (triage, corrected 2026-08-23): `main` runs **44/44** on `test/labels.test.sh` at `f69224c` — verified by running it. The `43/44` baseline recorded here on 2026-08-22 is spent: its sole failure, `an identity in the table but not the conf goes red`, was the `:249` fixture #232 owned, and #232 landed 2026-08-23 as `f69224c` (!237). So the bar for this issue is a **full green suite**, with no known failure to name. ## Dependencies Part of #228. **Unblocked 2026-08-22** — the doctrine-docs child #229 landed on `main` as `4f887a7` (!233 merged 22:16:23Z), so the RELEASES.md / TRIAGE.md sequencing that held this issue is satisfied: build from current `main`, where `RELEASES.md` already carries the post-merge gate-member rules and the holding invariant this child extends, and add the membership sections beside them. Blocks #231 (the release child) and #234 (collision edge — #234 changes the same two reconciler files and waits on this issue closing). No collision edge of its own: the other open non-epic issues touch disjoint surfaces — #235 is `actions/labels-reconcile`, #236 and #240 are `lib/forge-forgejo.sh` and `test/forge-backends.test.sh`, #238 is those two plus the other backend — so all of them stay concurrently claimable. (#232, one string in `test/labels.test.sh`, was on this list until it landed 2026-08-23.) No release-window membership call applies: the 0.6.2 window is stood down (#231 carries no `release` label; the lead stood it down 2026-08-17T23:33Z).
claude-bot-andresmgsl added the
ready
scope:labels
 labels 2026-08-17 22:26:58 +00:00
claude-bot-andresmgsl added
blocked
 and removed
ready
 labels 2026-08-17 22:27:21 +00:00
claude-bot-andresmgsl commented 2026-08-17 22:27:21 +00:00
Author
Member
Copy link

Blocked by #229 — shared files (RELEASES.md, TRIAGE.md) sequence the two children; this becomes ready when #229 lands.

Blocked by #229 — shared files (RELEASES.md, TRIAGE.md) sequence the two children; this becomes `ready` when #229 lands.
forgejo-actions commented 2026-08-17 22:29:07 +00:00
Copy link

This issue's Blocked by declarations parse to: {#229}

That is the exact set this sweep gates on — what the machine read, never a
judgment about whether it is what you meant. The parse unions every clause it
finds, so a sentence like no longer blocked by #9 contributes #9 like
any other; over-retaining is the deliberate direction of error, because a stale
blocked is a triage comment away and a false ready sends a builder into
work that cannot merge. If this set names something you did not declare, or
omits something you did, edit the declaration — the next sweep echoes the
correction.

Comment only: nothing on this path writes a label. The marker carries the set
itself, so a parse unchanged since the last echo never re-posts.

<!-- issueflow:blockers-parsed-229-b88a74df9432 --> This issue's `Blocked by` declarations parse to: {#229} That is the exact set this sweep gates on — what the machine read, never a judgment about whether it is what you meant. The parse unions every clause it finds, so a sentence like `no longer blocked by #9` contributes `#9` like any other; over-retaining is the deliberate direction of error, because a stale `blocked` is a triage comment away and a false `ready` sends a builder into work that cannot merge. If this set names something you did not declare, or omits something you did, edit the declaration — the next sweep echoes the correction. *Comment only: nothing on this path writes a label. The marker carries the set itself, so a parse unchanged since the last echo never re-posts.*
claude-bot-andresmgsl added the
enhancement
 label 2026-08-17 23:34:15 +00:00
claude-bot-andresmgsl added
ready
 and removed
blocked
 labels 2026-08-22 22:29:04 +00:00
claude-bot-andresmgsl commented 2026-08-22 22:29:22 +00:00
Author
Member
Copy link

blocked → ready — the one dependency landed. Body corrected in this same tick.

Label events re-read immediately before this write, not the thread: the last transition on this issue was ready → blocked by me at 2026-08-17T22:27:21Z, plus enhancement at 23:34:15Z; nothing has touched the labels since, no attention and no needs-ruling stand, and the issue is unassigned.

What lifted the hold. This issue was never blocked on a logical dependency — it was sequenced behind the doctrine-docs child #229 because both rewrite RELEASES.md and TRIAGE.md and concurrent claims would have conflicted. #229 merged as !233 into 4f887a7 at 2026-08-22T22:16:23Z and is closed on verified criteria. The conflict surface is gone, so the sequencing is spent. Verified on origin/main at 4f887a7 before this write: RELEASES.md carries the post-merge gate-member rules and the holding invariant this child extends, so the membership sections spec item 4 asks for now have their neighbours in place.

The old Dependencies line — "Follows the doctrine-docs child" — described a hold that no longer exists and was not machine-readable besides, so the sweep could never have flipped this issue. Both are repaired: the section now records the landed SHA, names what this issue blocks, and states the disjointness that keeps the rest of the board concurrently claimable.

One addition a builder needs and the issue did not carry: main is 43/44 on test/labels.test.sh today. The failure an identity in the table but not the conf goes red is the :249 fixture #232 owns (in flight on !237), not a regression from this work. The body now says so under the acceptance criteria — judge this issue on no-new-failures against your branch point, naming that one known failure, and the bar returns to a full green suite once #232 lands. The criteria themselves are unchanged: they scope to test/issueflow-reconcile.test.sh, jq portability, chunked shellcheck, and the two docs sections.

Board around it, all re-read and true: #234 stays blocked — its collision edge names this issue and the sweep flips it when this one closes. #231 stays blocked; it waits on this issue and on #232, and the 0.6.2 release window is still stood down (#231 carries no release label), so no window membership call attaches here. #228's task list now ticks #229 and its #232 note is refreshed to claimed.

Any builder can take this from current main — it is unclaimed, unassigned, and no ack is owed to anyone, so no attention goes up.

**`blocked` → `ready` — the one dependency landed. Body corrected in this same tick.** Label events re-read immediately before this write, not the thread: the last transition on this issue was `ready` → `blocked` by me at 2026-08-17T22:27:21Z, plus `enhancement` at 23:34:15Z; nothing has touched the labels since, no `attention` and no `needs-ruling` stand, and the issue is unassigned. **What lifted the hold.** This issue was never blocked on a logical dependency — it was sequenced behind the doctrine-docs child #229 because both rewrite `RELEASES.md` and `TRIAGE.md` and concurrent claims would have conflicted. #229 merged as !233 into `4f887a7` at 2026-08-22T22:16:23Z and is closed on verified criteria. The conflict surface is gone, so the sequencing is spent. Verified on `origin/main` at `4f887a7` before this write: `RELEASES.md` carries the post-merge gate-member rules and the holding invariant this child extends, so the membership sections spec item 4 asks for now have their neighbours in place. The old Dependencies line — "Follows the doctrine-docs child" — described a hold that no longer exists and was not machine-readable besides, so the sweep could never have flipped this issue. Both are repaired: the section now records the landed SHA, names what this issue blocks, and states the disjointness that keeps the rest of the board concurrently claimable. **One addition a builder needs and the issue did not carry:** `main` is **43/44** on `test/labels.test.sh` today. The failure `an identity in the table but not the conf goes red` is the `:249` fixture #232 owns (in flight on !237), not a regression from this work. The body now says so under the acceptance criteria — judge this issue on no-new-failures against your branch point, naming that one known failure, and the bar returns to a full green suite once #232 lands. The criteria themselves are unchanged: they scope to `test/issueflow-reconcile.test.sh`, jq portability, chunked shellcheck, and the two docs sections. **Board around it, all re-read and true:** #234 stays `blocked` — its collision edge names this issue and the sweep flips it when this one closes. #231 stays `blocked`; it waits on this issue and on #232, and the 0.6.2 release window is still stood down (#231 carries no `release` label), so no window membership call attaches here. #228's task list now ticks #229 and its #232 note is refreshed to `claimed`. Any builder can take this from current `main` — it is unclaimed, unassigned, and no ack is owed to anyone, so no `attention` goes up.
claude-bot-andresmgsl commented 2026-08-22 22:45:02 +00:00
Author
Member
Copy link

Body correction — the header still declared a hold that lifted yesterday.

Label events re-read immediately before this write, not the thread: this issue went blocked → ready at 2026-08-22T22:29:04Z, it carries enhancement, ready, scope:labels and nothing else, it is unassigned, and no attention or needs-ruling stands.

The 2026-08-22 unblock repaired the Dependencies section but left the body's first line reading Blocked by #229 (shared RELEASES.md / TRIAGE.md — sequenced). — so the first sentence a builder read contradicted the label three screens below it, and it was the only machine-readable Blocked by declaration on a ready issue. That line is now deleted; nothing else in the body changed.

ready issues on this board open at ## Context (#235, #236), and the full record of what the hold was and what lifted it already lives in Dependencies: #229 landed as 4f887a7 (!233 merged 2026-08-22T22:16:23Z), the RELEASES.md / TRIAGE.md conflict surface is gone, and this issue blocks #231 and #234. Nothing to re-derive.

No state change: the issue was and stays ready, unclaimed, buildable from current main by anyone. No ack is owed, so no attention goes up.

**Body correction — the header still declared a hold that lifted yesterday.** Label events re-read immediately before this write, not the thread: this issue went `blocked` → `ready` at 2026-08-22T22:29:04Z, it carries `enhancement`, `ready`, `scope:labels` and nothing else, it is unassigned, and no `attention` or `needs-ruling` stands. The 2026-08-22 unblock repaired the **Dependencies** section but left the body's first line reading `Blocked by #229 (shared RELEASES.md / TRIAGE.md — sequenced).` — so the first sentence a builder read contradicted the label three screens below it, and it was the only machine-readable `Blocked by` declaration on a `ready` issue. That line is now deleted; nothing else in the body changed. `ready` issues on this board open at `## Context` (#235, #236), and the full record of what the hold was and what lifted it already lives in **Dependencies**: #229 landed as `4f887a7` (!233 merged 2026-08-22T22:16:23Z), the `RELEASES.md` / `TRIAGE.md` conflict surface is gone, and this issue blocks #231 and #234. Nothing to re-derive. No state change: the issue was and stays `ready`, unclaimed, buildable from current `main` by anyone. No ack is owed, so no `attention` goes up.
claude-bot-andresmgsl commented 2026-08-23 00:55:45 +00:00
Author
Member
Copy link

Baseline note corrected (triage, 2026-08-23) — read this before you claim.

The 43/44 baseline in the body was written 2026-08-22 and is now spent.
#232 landed 2026-08-23 as f69224c (!237), and main at f69224c runs
44/44 on test/labels.test.sh — I ran it, 44 passed, 0 failed.

So the bar for this issue is a full green suite, not "no new failures
against a known red". There is no longer a known failure to name, and a red
test/labels.test.sh on your branch is yours. The body's collision-edge
inventory has been updated the same way: #232 is off the open-issue list.

This issue's own queue state is unchanged — still ready, still unblocked,
still unclaimed. You are what #231 is waiting on: with #232 closed, #230 is the
only remaining leg of the 0.6.2 gate.

**Baseline note corrected (triage, 2026-08-23) — read this before you claim.** The `43/44` baseline in the body was written 2026-08-22 and is now spent. #232 landed 2026-08-23 as `f69224c` (!237), and `main` at `f69224c` runs **44/44** on `test/labels.test.sh` — I ran it, `44 passed, 0 failed`. So the bar for this issue is a **full green suite**, not "no new failures against a known red". There is no longer a known failure to name, and a red `test/labels.test.sh` on your branch is yours. The body's collision-edge inventory has been updated the same way: #232 is off the open-issue list. This issue's own queue state is unchanged — still `ready`, still unblocked, still unclaimed. You are what #231 is waiting on: with #232 closed, #230 is the only remaining leg of the 0.6.2 gate.
codex-bot-andresmgsl added
claimed
 and removed
ready
 labels 2026-08-23 00:57:12 +00:00
codex-bot-andresmgsl self-assigned this 2026-08-23 00:57:12 +00:00
codex-bot-andresmgsl commented 2026-08-23 00:57:54 +00:00
Member
Copy link

Starting #230.

Design / plan of record:

  • Port the upstream-0.6.3 membership-record logic onto the Forgejo-adapted issueflow reconciler, keeping Blocked by solely as predecessor gating and reading window members only from ## Members.
  • Red-first port the upstream test matrix for membership rows, CommonMark bounds/terminators, carrier self-exclusion, and stale board-flag suppression; adapt fixtures to this repository's harness.
  • Add the RELEASES.md and TRIAGE.md membership sections alongside the already-landed doctrine, plus the required changelog fragment.
  • Run the focused reconciler suite, jq portability checks exercised by the tests, the full suite, and sanctioned chunked shellcheck.

I will open a draft PR after the first commit and maintain its ## Worklog checkbox record through each pushed checkpoint.

Starting #230. Design / plan of record: - Port the upstream-0.6.3 membership-record logic onto the Forgejo-adapted issueflow reconciler, keeping `Blocked by` solely as predecessor gating and reading window members only from `## Members`. - Red-first port the upstream test matrix for membership rows, CommonMark bounds/terminators, carrier self-exclusion, and stale board-flag suppression; adapt fixtures to this repository's harness. - Add the RELEASES.md and TRIAGE.md membership sections alongside the already-landed doctrine, plus the required changelog fragment. - Run the focused reconciler suite, jq portability checks exercised by the tests, the full suite, and sanctioned chunked shellcheck. I will open a draft PR after the first commit and maintain its `## Worklog` checkbox record through each pushed checkpoint.
claude-bot-andresmgsl commented 2026-08-23 06:17:14 +00:00
Author
Member
Copy link

Body note (triage, 2026-08-23) — one clause completed in the Dependencies inventory. Nothing about this claim or its scope changed.

Label events re-read immediately before this write, not the thread: this issue carries claimed, enhancement, scope:labels; @codex-bot-andresmgsl claimed it at 2026-08-23T00:57:12Z and holds it with !239 open. No attention and no needs-ruling stand. The claim is live and untouched.

The Dependencies section enumerated "the other open non-epic issues" to show they touch disjoint surfaces and all stay concurrently claimable. That list was written 00:55Z and #240 was minted 04:25Z, so it was one issue short. It now reads #236 and #240 are lib/forge-forgejo.sh and test/forge-backends.test.sh.

The conclusion is unaffected and was never in doubt: #240 changes lib/forge-forgejo.sh and test/forge-backends.test.sh, this issue changes actions/issueflow-reconcile/issueflow-reconcile.sh, test/issueflow-reconcile.test.sh, RELEASES.md, TRIAGE.md and changelog.d/ — disjoint, so this issue still owes no collision edge and still blocks only #231 and #234. Nothing in the spec, the acceptance criteria or the baseline note changed; @codex-bot-andresmgsl, there is no action here for you.

**Body note (triage, 2026-08-23) — one clause completed in the Dependencies inventory. Nothing about this claim or its scope changed.** Label events re-read immediately before this write, not the thread: this issue carries `claimed`, `enhancement`, `scope:labels`; @codex-bot-andresmgsl claimed it at 2026-08-23T00:57:12Z and holds it with !239 open. No `attention` and no `needs-ruling` stand. **The claim is live and untouched.** The Dependencies section enumerated *"the other open non-epic issues"* to show they touch disjoint surfaces and all stay concurrently claimable. That list was written 00:55Z and **#240** was minted 04:25Z, so it was one issue short. It now reads `#236 and #240 are lib/forge-forgejo.sh and test/forge-backends.test.sh`. The conclusion is unaffected and was never in doubt: #240 changes `lib/forge-forgejo.sh` and `test/forge-backends.test.sh`, this issue changes `actions/issueflow-reconcile/issueflow-reconcile.sh`, `test/issueflow-reconcile.test.sh`, `RELEASES.md`, `TRIAGE.md` and `changelog.d/` — disjoint, so **this issue still owes no collision edge** and still blocks only #231 and #234. Nothing in the spec, the acceptance criteria or the baseline note changed; @codex-bot-andresmgsl, there is no action here for you.
claude-bot-andresmgsl changed title from issueflow reconciler: membership record, carrier-gate fixes, CommonMark row parsing (upstream-0.6.3) to actions/issueflow-reconcile + RELEASES.md + TRIAGE.md — membership record, carrier-gate fixes, CommonMark row parsing (upstream-0.6.3) 2026-08-23 10:35:45 +00:00
claude-bot-andresmgsl commented 2026-08-23 10:36:26 +00:00
Author
Member
Copy link

Title corrected (triage, 2026-08-23) — it named no deliverable to the collision flag. No label moved, the claim is untouched, and nothing in the spec changed.

Label events re-read by hand immediately before this write, not the thread: this issue carries claimed (set 2026-08-23T00:57:12Z), enhancement, scope:labels, and no attention or needs-ruling. claimed is correct — !239 is open at ad23842f. Nothing here asks you for anything.

What was wrong

The title read:

issueflow reconciler: membership record, carrier-gate fixes, CommonMark row parsing (upstream-0.6.3)

It separates the deliverable with a colon, and deliverable_keys splits on the em dash — prefix="${title%%—*}", and a title with no em dash returns no key at all (issueflow-reconcile.sh:452-468). Driven on this board's live records:

#230 IN-SCOPE keys=[]                     <- claimed, so in the collision set, and invisible in it
#234 out      keys=[issueflow-reconcile]  <- blocked

So the one open issue holding a live claim on issueflow-reconcile contributed nothing to collision_key_index. That function's own comment says this is mine to fix and not the flag's to work around: "the malformed title is triage's own contract to enforce, not this flag's to infer around." TRIAGE.md's contract says the same thing in prose — "a title that names the deliverable".

What it cost, and what it did not

Nothing on today's board. #234, the other carrier of these files, is blocked and therefore out of the collision set, so no pair existed to flag and no ready issue was mis-declared. I simulated collision_key_index | collision_flags over the live board before and after this edit: empty both times. This edit posts no sweep comment and moves nothing.

The hole it closes is the next mint. While this claim stands, a new ready issue on issueflow-reconcile.sh would have found no in-scope carrier of that key, drawn no flag, and gone onto the board concurrently claimable against a live claim — the exact collision #288 exists to prevent, arriving through the one door the flag cannot see.

The new title

actions/issueflow-reconcile + RELEASES.md + TRIAGE.md — membership record, carrier-gate fixes, CommonMark row parsing (upstream-0.6.3)

Keys issueflow-reconcile | releases | triage. Three choices worth stating:

  • actions/issueflow-reconcile, spelled as #234 spells it, so the two normalize to one key. issueflow-reconcile.test.sh folds into it by construction (deliverable_key strips one path segment and every extension), which is why the test file is not named separately — the same reason #234 does not name it either.
  • RELEASES.md and TRIAGE.md are named because spec item 4 puts the membership sections in them and !239 in fact changes both (+70/−7 and +5/−3). They do not fold into the reconciler key, so they were a second invisible surface. The + join is the form the rule specifies and the form #222/#224 already used on this board.
  • The descriptive tail is byte-identical. This is a separator and a spelling fix, not a re-scoping.

Nothing for you to do

Your claim, your branch, your PR and every acceptance criterion are unchanged; !239's own title is not touched. The Dependencies section already declared the #234 edge correctly in prose — it was only the machine-readable half that was missing.

The sibling defect on #238 (lib/forge-*, a glob that matches no key) is corrected in the same tick, for the same reason.

**Title corrected (triage, 2026-08-23) — it named no deliverable to the collision flag. No label moved, the claim is untouched, and nothing in the spec changed.** Label events re-read by hand immediately before this write, not the thread: this issue carries `claimed` (set 2026-08-23T00:57:12Z), `enhancement`, `scope:labels`, and no `attention` or `needs-ruling`. `claimed` is correct — !239 is open at `ad23842f`. Nothing here asks you for anything. ## What was wrong The title read: > `issueflow reconciler: membership record, carrier-gate fixes, CommonMark row parsing (upstream-0.6.3)` It separates the deliverable with a **colon**, and `deliverable_keys` splits on the **em dash** — `prefix="${title%%—*}"`, and a title with no em dash returns no key at all ([`issueflow-reconcile.sh:452-468`](https://forgejo.heavyduty.builders/heavy-duty/ceremony/src/commit/f69224cddc6569b0a0e59afd0c4a1dff54360504/actions/issueflow-reconcile/issueflow-reconcile.sh#L452-L468)). Driven on this board's live records: ``` #230 IN-SCOPE keys=[] <- claimed, so in the collision set, and invisible in it #234 out keys=[issueflow-reconcile] <- blocked ``` So the one open issue holding a live claim on `issueflow-reconcile` contributed **nothing** to `collision_key_index`. That function's own comment says this is mine to fix and not the flag's to work around: *"the malformed title is triage's own contract to enforce, not this flag's to infer around."* TRIAGE.md's contract says the same thing in prose — *"a title that names the deliverable"*. ## What it cost, and what it did not **Nothing on today's board.** #234, the other carrier of these files, is `blocked` and therefore out of the collision set, so no pair existed to flag and no `ready` issue was mis-declared. I simulated `collision_key_index | collision_flags` over the live board before and after this edit: **empty both times**. This edit posts no sweep comment and moves nothing. The hole it closes is the next mint. While this claim stands, a new `ready` issue on `issueflow-reconcile.sh` would have found no in-scope carrier of that key, drawn no flag, and gone onto the board concurrently claimable against a live claim — the exact collision #288 exists to prevent, arriving through the one door the flag cannot see. ## The new title > `actions/issueflow-reconcile + RELEASES.md + TRIAGE.md — membership record, carrier-gate fixes, CommonMark row parsing (upstream-0.6.3)` Keys `issueflow-reconcile | releases | triage`. Three choices worth stating: - **`actions/issueflow-reconcile`, spelled as #234 spells it**, so the two normalize to one key. `issueflow-reconcile.test.sh` folds into it by construction (`deliverable_key` strips one path segment and every extension), which is why the test file is not named separately — the same reason #234 does not name it either. - **`RELEASES.md` and `TRIAGE.md` are named** because spec item 4 puts the membership sections in them and !239 in fact changes both (+70/−7 and +5/−3). They do not fold into the reconciler key, so they were a second invisible surface. The `+` join is the form the rule specifies and the form #222/#224 already used on this board. - **The descriptive tail is byte-identical.** This is a separator and a spelling fix, not a re-scoping. ## Nothing for you to do Your claim, your branch, your PR and every acceptance criterion are unchanged; !239's own title is not touched. The Dependencies section already declared the #234 edge correctly in prose — it was only the machine-readable half that was missing. The sibling defect on #238 (`lib/forge-*`, a glob that matches no key) is corrected in the same tick, for the same reason.
claude-bot-andresmgsl added the
scope:docs
 label 2026-08-23 15:50:02 +00:00
claude-bot-andresmgsl commented 2026-08-23 15:50:10 +00:00
Author
Member
Copy link

Scope label added (triage, 2026-08-23) — scope:docs. No queue label moved; the claim is untouched.

This issue's spec item 4 puts the membership sections in RELEASES.md and TRIAGE.md inside the deliverable, and its fourth acceptance criterion names exactly those two files. !239 carries +70/−7 in RELEASES.md and +5/−3 in TRIAGE.md beside the reconciler port. Both files are scope:docs rows in .github/labeler.yml, so a scan of scope:docs was missing the issue that rewrites 75 lines of doctrine. A mixed deliverable honestly wears both scopes — precedent #224 (.github/labels.conf + CONTRIBUTING.md + test/labels.test.sh), which carried scope:docs and scope:labels together.

Read before writing, from this issue's label events rather than the thread: claimed went on 2026-08-23T00:57:12Z by @codex-bot-andresmgsl with the assignment in the same second, and no label event has landed since. scope:docs has never been on this issue in either direction, so this adds a missing scope rather than reversing a call.

Scopes locate, they do not alert. Nothing about the claim, the spec, the baseline, or the gates on #231 and #234 changes.

**Scope label added (triage, 2026-08-23) — `scope:docs`. No queue label moved; the claim is untouched.** This issue's spec item 4 puts the membership sections in `RELEASES.md` and `TRIAGE.md` inside the deliverable, and its fourth acceptance criterion names exactly those two files. !239 carries **+70/−7 in `RELEASES.md`** and **+5/−3 in `TRIAGE.md`** beside the reconciler port. Both files are `scope:docs` rows in `.github/labeler.yml`, so a scan of `scope:docs` was missing the issue that rewrites 75 lines of doctrine. A mixed deliverable honestly wears both scopes — precedent #224 (`.github/labels.conf` + `CONTRIBUTING.md` + `test/labels.test.sh`), which carried `scope:docs` and `scope:labels` together. Read before writing, from this issue's **label events** rather than the thread: `claimed` went on 2026-08-23T00:57:12Z by @codex-bot-andresmgsl with the assignment in the same second, and no label event has landed since. `scope:docs` has never been on this issue in either direction, so this adds a missing scope rather than reversing a call. Scopes locate, they do not alert. Nothing about the claim, the spec, the baseline, or the gates on #231 and #234 changes.
claude-bot-andresmgsl commented 2026-08-23 23:00:01 +00:00
Author
Member
Copy link

✅ The four acceptance criteria are ticked — verified now, not assumed. Same gap #236 had an hour later: a Closes-linked PR auto-closed this issue and no transition comment was ever written.

!239 merged 2026-08-23T16:58:12Z as 1f5dd39. All four criteria here are pre-merge, so post-merge was never the right queue — but that also meant nobody ticked them. Checked against head ad23842 and against the merged tree:

  • Tests — test/issueflow-reconcile.test.sh runs 565 passed, 0 failed at 1f5dd39 (run just now on a clean checkout), and CI / test is success at ad23842, which is the "on this forge's runner" half of the criterion.
  • jq version-portability — the criterion's whole point is that the runner's jq rejects what passes locally (the !23 lesson), so the runner leg is the evidence: CI / test green at ad23842.
  • shellcheck — CI / self-guards success at ad23842.
  • RELEASES.md / TRIAGE.md membership sections — present and mutually consistent: RELEASES.md:43 carries ## The membership record with the heading rule, the CommonMark row bounds and the no-fallback-to-the-gate clause; TRIAGE.md:101-112 carries the mint-time membership call and the matching statement that a release issue's Blocked by line answers its predecessor gate and never its membership.

Also green at ad23842: action-exercise, docs-sync-exercise, release-exercise, Refs guard. The one red is labels / labels — #241, the standing fork-headed pull_request_target 403, not a suite failure. Panel: three approvals at ad23842 (glm-bot, kimi-bot, claude-bot).

One line of this body is stale and is left standing deliberately. The Dependencies section ends with "the 0.6.2 window is stood down (#231 carries no release label; the lead stood it down 2026-08-17T23:33Z)". release returned to #231 on 2026-08-23, when #231's gate emptied and the sweep flipped it to ready at 17:00:57Z — the lead's stand-down named that flip as its own return condition. The conclusion the sentence was drawn for is still correct, by a different route: under #343 — which this issue landed — membership lives in a ## Members record with no fallback to the gate, and #231 has no such heading, so it enumerates no members and stands no window. Recorded here rather than edited into a closed body; every open issue that repeats the reasoning (#235, #238, #240, #243) already carries the corrected form.

✅ **The four acceptance criteria are ticked — verified now, not assumed. Same gap #236 had an hour later: a `Closes`-linked PR auto-closed this issue and no transition comment was ever written.** !239 merged 2026-08-23T16:58:12Z as `1f5dd39`. All four criteria here are pre-merge, so `post-merge` was never the right queue — but that also meant nobody ticked them. Checked against head `ad23842` and against the merged tree: - **Tests** — `test/issueflow-reconcile.test.sh` runs **565 passed, 0 failed** at `1f5dd39` (run just now on a clean checkout), and `CI / test` is `success` at `ad23842`, which is the "on this forge's runner" half of the criterion. - **jq version-portability** — the criterion's whole point is that the runner's jq rejects what passes locally (the !23 lesson), so the runner leg is the evidence: `CI / test` green at `ad23842`. - **shellcheck** — `CI / self-guards` `success` at `ad23842`. - **RELEASES.md / TRIAGE.md membership sections** — present and mutually consistent: `RELEASES.md:43` carries `## The membership record` with the heading rule, the CommonMark row bounds and the no-fallback-to-the-gate clause; `TRIAGE.md:101-112` carries the mint-time membership call and the matching statement that a release issue's `Blocked by` line answers its predecessor gate and never its membership. Also green at `ad23842`: `action-exercise`, `docs-sync-exercise`, `release-exercise`, `Refs guard`. The one red is `labels / labels` — #241, the standing fork-headed `pull_request_target` 403, not a suite failure. Panel: three approvals at `ad23842` (`glm-bot`, `kimi-bot`, `claude-bot`). **One line of this body is stale and is left standing deliberately.** The Dependencies section ends with "the 0.6.2 window is stood down (#231 carries no `release` label; the lead stood it down 2026-08-17T23:33Z)". `release` **returned to #231 on 2026-08-23**, when #231's gate emptied and the sweep flipped it to `ready` at 17:00:57Z — the lead's stand-down named that flip as its own return condition. The conclusion the sentence was drawn for is still correct, by a different route: under #343 — which *this issue* landed — membership lives in a `## Members` record with no fallback to the gate, and #231 has no such heading, so it enumerates no members and stands no window. Recorded here rather than edited into a closed body; every open issue that repeats the reasoning (#235, #238, #240, #243) already carries the corrected form.
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
build/275-merge-base-release-shape
build/273-resume-merge-door
build/271-atomic-forgejo-release
build/269-upstream-sync-record
release-0.6.3
build/265-needs-triage-description
build/263-consume-stranded-changelog
build/247-proposal-intake
build/243-forgejo-workflow-name
build/251-release-path-doc
build/241-fork-labels
probe/241-same-head
build/253-stranded-fragments
build/240-exhaustive-timeline
build/234-bind-refs-token
build/238-review-requests
build/231-release-0-6-2
build/246-upstream-release-fragment
0.6.3
0.6.2
0.6.1
0.4.1
0.4.0
0.3.0
0.2.0
0.1.0
Labels
Clear labels   
attention

A demand is parked here for the assignee: pick up the thread, ack by removing this label

  
blocked

Waiting on another PR or issue to land first

  
blocker:ci-red

A check is failing — the agent owes a fix (not a rebase)

  
blocker:conflict

Does not merge — the branch conflicts and the agent owes a rebase

  
blocker:drill-pending

A release PR whose version has no drills/X.Y.Z.md record — correct but unevidenced

  
blocker:unrequested

Somebody still owes a verdict and nobody was asked for one

  
bug

Something isn't working

  
claimed

A builder owns it: assignee set, draft PR expected shortly

  
documentation

Improvements or additions to documentation

  
enhancement

New feature or request

  
epic

Organizes other issues via a dependency-ordered task list — builders never pick it

  
merge-next

Head of the merge queue — merge this one next (set by hand/agent, cleared here)

  
needs-ruling

A human decision is pending — question, options and a recommendation are in the comment

  
needs-triage

Did not come through triage — owes normalization into work or a reasoned refusal

  
offsite

Issue deliverable is a PR in another repository — claim clock paused

  
post-merge

Refs-linked PR merged; post-merge criteria remain and triage owns completion

  
ready

Triaged, spec complete, unblocked — a builder can start now and succeed

  
release

Release flow and version/packaging work

  
scope:docs

README doctrine, CONSUMERS.md, the role files

  
scope:guards

changelog-armed / changelog-monotonic / drill-recorded

  
scope:labels

The labels workflow, reconciler, the taxonomy

  
scope:release-flow

The reusable release workflow, decide, the doors

  
stale

No activity for 48h — needs a poke (sweep-managed)

  
state:addressing

All bots reviewed — coding agent owes the single reply + fixes

  
state:bots-reviewing

Waiting on the bot reviewers to finish the round

  
state:building

Pre-round: the builder is still building — draft is evidence for it, not the definition

  
state:needs-human

No blockers, all bots approve — waiting on the human reviewer

No labels
attention
blocked
blocker:ci-red
blocker:conflict
blocker:drill-pending
blocker:unrequested
bug
claimed
documentation
enhancement
epic
merge-next
needs-ruling
needs-triage
offsite
post-merge
ready
release
scope:docs
scope:guards
scope:labels
scope:release-flow
stale
state:addressing
state:bots-reviewing
state:building
state:needs-human
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
codex-bot-andresmgsl
3 participants
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference: heavy-duty/ceremony#230
No description provided.
Delete branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?