Term 1's foundation. lib/forge.sh gains forge_select, which sources exactly
one of lib/forge-github.sh or lib/forge-forgejo.sh; both define the same
verbs, so no branching reaches the 61 call sites. The github backend is the
current gh invocation extracted 1:1 — term 5 is kept by making that path
boring.
The page size moves OUT of the call sites and into the backend, because it
is not portable and fails silently. Measured 2026-08-02:
?per_page=100 GitHub 100 items Forgejo 30 items (ignored)
?limit=100 GitHub 30 items Forgejo 50 items (capped)
Both answer HTTP 200 with valid JSON. Every call site here is GitHub-shaped,
so a verbatim port would have swept 30 of rig's 137 issues and printed
"reconciled." — criterion 2 failing green, the same failure class as the
blind sweep. Both page_url helpers strip a stray page-size parameter in
either dialect, so a call site cannot reintroduce it by accident.
Forgejo caps a page at 50 whatever is asked, so pagination is mandatory, not
an optimisation. The gather is then PROVEN complete against x-total-count
rather than assumed complete because a loop ended.
@kimi-reviewer-andresmgsl's hardening (#4699): a missing x-total-count is
itself a loud refusal. Header exposure is a server setting, and an assert
that cannot run must not silently pass — that is the failure class
re-entering through the guard built to stop it.
Call sites are not ported yet; that is the next commit.
Refs #188
Term 3 of #188. Forgejo has no GraphQL API, so these two gathers could not
be translated — there is no endpoint to translate them to. A real
forgejo-runner job says so from the other side: GITHUB_GRAPHQL_URL arrives
set to the empty string (probe task 278).
MERGED_REF_PR_RECORDS was already a body parse; GraphQL was buying
pagination, nothing semantic. OPEN_PR_ISSUES used GitHub's own parse of the
closing keywords, so it becomes lib/closes_references.sh — a sibling of
refs_references, sharing its LOCAL/CROSS classifier so rig#112 can still
never be read as local #112 (#61).
Both gathers now read /pulls, which /api/v3 and /api/v1 return in the same
shape (measured on both). merged_at replaces GraphQL's states: MERGED.
Bodies travel base64: jq's @tsv escapes a newline to a literal backslash-n,
which a line parser reads as one line and loses every declaration after the
first.
The accepted delta, written down rather than rediscovered: GitHub also
records closing links attached through the PR development sidebar, which
live in no body. This family declares links in the body, so the delta is
zero here.
Refs #188
The preflight half of #188, landed first so it stands alone: the forge is
decided once, before any sweep, and a client that cannot speak it exits
non-zero with a named reason.
Measured against forgejo.heavyduty.builders at 84bb1a4 — two of the three
actions reported SUCCESS having read nothing:
labels-scope exit 0 "no .github/labeler.yml" (the file is HTTP 200)
labels-reconcile exit 0 "reconciled." (zero PRs enumerated)
issueflow-reconcile exit 1 "unexpected end of JSON input"
labels-reconcile's blind-sweep warning (#96) could not fire: it counts
unreadable PRs against a list `gh pr list` never produced, and a process
substitution's failure does not trip set -e, so total stayed 0. Installing
gh makes it worse, silencing the one loud failure.
Detection is measured, not inferred from docs: a real forgejo-runner v6.3.1
job (probe task 278) shows Forgejo populating the whole GITHUB_* namespace,
so GITHUB_ACTIONS proves nothing. GITHUB_API_URL's shape, GITEA_ACTIONS and
GITHUB_SERVER_URL do. The same probe shows the runner image carries neither
gh nor stoke, which is what makes the forgejo backend REST.
Tests declare CEREMONY_FORGE at the forge boundary rather than stubbing gh
and staying silent about the forge — the boundary move term 5 asks for.
Refs #188
Command substitution strips every trailing newline, so 'grouped\n\n'
reached the case as a clean word and passed — the round's shared blocker
(codex, grok). A line count taken from the file itself now refuses any
physically multi-line sentinel before the word check runs, with the
existing diagnosis naming the file. Red rows: grouped/flat with a
trailing blank line in the unit suite, grouped in the armed suite.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The sentinel pins the set's shape, outranking the newest-published-section
inference (#182 D2); malformed content is a red diagnosis naming the file.
bin/changelog-assemble skips 'shape' in the stray-file loop so the sentinel
survives the consumption (#182 D5).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
One definition in the fragment predicate; changelog-armed reds the PR
that writes the fragment and the assembler refuses at release, both by
inheritance. Doctrine names the number in BUILDER.md and CHANGELOG.md.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
A repository's first push to main is a branch-create push whose head is a
root commit: event.before is all-zeros and MERGE_SHA^1 does not exist, so
the fallback died at exit 128 before establishing a fact. The parent count
is now read via rev-list --parents (a fact, not an inferred failure), the
no-base path skips the belt-and-braces fetch and the base git show, and an
unresolvable MERGE_SHA still fails loudly.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
lib/changelog.sh gains changelog_fragments (publication order: trailing
issue number descending, filename tie-break), changelog_fragment_problem
(the release-time rules moved onto the PR that writes the fragment, #112
D9), and changelog_assemble (canonical group order per D5, one shape per
repo per D4). bin/changelog-assemble folds changelog.d/ into one release
section, deletes exactly what it consumed, and --check proves the body
without touching the tree.
Part of #112. Closes#114 groundwork; tests follow.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
lib/ruling.sh gains the mechanical half of #50 D12-D15: the comment body
rides the fetch as base64 (rows stay line-oriented), ruling_shape_decision
checks the four line-anchored bold-tolerant field labels,
ruling_deadline_decision derives the rung from the labeled epoch,
ruling_default_decision parses the Default: line for wording only, and
reconcile_ruling wires them with bare-stops-here exclusion — every write a
comment, one per marker per episode, both surfaces via the shared lib.
Part of #73.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
One implementation for both surfaces: the stale exemption, the bare-flag
mechanical proxy (15-minute back-window against the newest labeled event),
the marker-scoped idempotency, the 7-day nudge with no marker (the comment
is the activity that resets its own window), and the one impure
orchestrator both reconcilers will source.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Closes#9. .github/workflows/release.yml (workflow_call) replaces box's,
rig's and cast's release.yml: the merge door (facts -> decide -> notes ->
tag+publish+bump, every decision in a tested lib script) and the tag door
(the manual fallback and backfill, no bump). Plus lib/facts.sh (the merge
door's impure half, contract-tested against fixture repos with a stubbed
gh), the self-ref pin guard (.github/scripts/self-ref-check.sh + CI step +
tests), the release-exercise scratch caller (dry wiring), the end-to-end
script-chain rehearsal, and the caller + artifact-hook contracts in
docs/CONSUMERS.md.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The six-row decision table from issue #8 as a pure script: the workflow
gathers facts (VER, BASE_VER, RELEASED, LABELED), this script decides.
Notices/refusals ported near-verbatim from box's decide step, de-repo-ified;
design lineage box#96 / rig#47 / cast#111.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The seam that lets box/rig/incubator (a VERSION file) and cast
(package.json + lockfile sync) share every other ceremony component
unchanged. Sourced, pure (no git), fail-loud on every unreadable state.
Carries cast's pkg_version discipline (node's parser, never regex) and
its lockfile-only bump incantation; refuses pre-release arithmetic
(-dev/-rc1) per box's prefix-confusion lore. The npm-backed write test
skips locally without npm but CI sets CEREMONY_REQUIRE_NPM so the skip
is a failure there — the case can never quietly stop running.
Closes#3
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
