Say loudly what cannot be judged, read only a merge's changed directories, and move the taken key only by an applied merge (issues 1422, 1126; group fix/1422-trunk-unknown) #94

2026-10-28T10:20:47Z

mesh-tools commented

2026-10-28 10:20:47 +00:00

Delivery group fix/1422-trunk-unknown (mesh-delivery, mesh-catalog).

Issue 1422 (notes 4-8): a pull request of a repository the gate's source.main names no trunk head for was dropped silently by the checker (de-spiegel, novadock, hal).

Checker: such a candidate gets a 'cannot be judged: the trunk head of is not known' verdict, an error status and one log line per head, shown in checker_status; judged whole once the head is known.
Forge bridge: a pull.updated into a base other than main is written closed (ADR 0484 §11), so no candidate is made for it; a retarget to main is announced again by gitea (mesh-catalog member).

Issue 1126 (note 61): the bridge timed out ten times on mesh-catalog's merge 9fd9569 (140 modules, 420 file reads in one 30 s deadline) and dropped it; the indexer then moved its taken key without applying the merge, so show and view_reach said 'on the machines' falsely, and mesh-vault was never rebuilt.

Indexer: the taken key moves only by a merge it applied (ADR 0556 clause 1).
Bridge: a merge reads only module directories whose tree changed, kept by tree hash in memory and in a new state inputs; a failed try keeps what it read whole, so a large merge converges.
indexer_take_in re-reads a held module only when the forge says its record's commit is an ancestor of the source head (the recovery for mesh-vault after delivery).

Quick vs thorough: thorough for both (the timeout was met twice; the silent drop is the kind 1422 ends). Rule bent and said: the four 1126 commits had a first review reading in the review's third round, not counted against its limit.
Tests: rows and one renamed test only; mesh-delivery's count does not rise. Locally: gofmt, vet -tags integration, unit and the indexer's, bridge's, view's and checker's integration tests against an embedded bus pass.
Note kept: the bridge's inputs bucket has no TTL (issue 1126 note 62).

Reviewed-by: rev-1422@shanks at 2b801f5bd4

Delivery group fix/1422-trunk-unknown (mesh-delivery, mesh-catalog). Issue 1422 (notes 4-8): a pull request of a repository the gate's source.main names no trunk head for was dropped silently by the checker (de-spiegel, novadock, hal). - Checker: such a candidate gets a 'cannot be judged: the trunk head of <repo> is not known' verdict, an error status and one log line per head, shown in checker_status; judged whole once the head is known. - Forge bridge: a pull.updated into a base other than main is written closed (ADR 0484 §11), so no candidate is made for it; a retarget to main is announced again by gitea (mesh-catalog member). Issue 1126 (note 61): the bridge timed out ten times on mesh-catalog's merge 9fd9569 (140 modules, 420 file reads in one 30 s deadline) and dropped it; the indexer then moved its taken key without applying the merge, so show and view_reach said 'on the machines' falsely, and mesh-vault was never rebuilt. - Indexer: the taken key moves only by a merge it applied (ADR 0556 clause 1). - Bridge: a merge reads only module directories whose tree changed, kept by tree hash in memory and in a new state `inputs`; a failed try keeps what it read whole, so a large merge converges. - indexer_take_in re-reads a held module only when the forge says its record's commit is an ancestor of the source head (the recovery for mesh-vault after delivery). Quick vs thorough: thorough for both (the timeout was met twice; the silent drop is the kind 1422 ends). Rule bent and said: the four 1126 commits had a first review reading in the review's third round, not counted against its limit. Tests: rows and one renamed test only; mesh-delivery's count does not rise. Locally: gofmt, vet -tags integration, unit and the indexer's, bridge's, view's and checker's integration tests against an embedded bus pass. Note kept: the bridge's `inputs` bucket has no TTL (issue 1126 note 62). Reviewed-by: rev-1422@shanks at 2b801f5bd49a710203944af416a8bf641231f7f6

mesh-tools added 7 commits 2026-10-28 10:20:47 +00:00

Checker: say loudly that a pull request cannot be judged when its repository has no trunk head (issue 1422) 4e9b076408

A repository never merged through the gate has no entry in source.main, so its pull requests were dropped without a log
line, a verdict or a status, and checker_status counted them as new. Now the verdict says it cannot be judged and why,
the forge status is error with those words, checker_status names it under cannot_be_judged, the log says it once per
head, and the heads are judged whole once the trunk head is known. Quick fix: this. Thorough fix (the gate seeding
source.main from the pull message's base head) is still to do: it needs mesh-sdk, the forge bridge and the gate.

Forge bridge: write a head pushed to a pull request into another base as closed, and say so (issue 1422) 93c0ae19dd

The checker judges pull requests into main only and source.main names main's head only (ADR 0484 §11), but a live update
into another base (de-spiegel#6 into nox-mesh) became a candidate, unlike at the start (OpenPulls) and at its merge. Once
the gate seeds a repository's trunk head, such a candidate would be judged against main's head: a verdict that lies.

The indexer's taken key moves only by a merge it applied, never by the gate's source head (issue 1126) 458e403a04

A drained forge stream does not say a merge was applied: the forge bridge was still
retrying mesh-catalog#535's merged message when the indexer started, and the source's
head moved taken.mesh-catalog to 9fd9569, so the view and show said the merge was on the
machines while mesh-vault's record stood at 78d7ef7 (ADR 0556 clause 1).

The forge bridge reads a merge's module directories only where their tree changed (issue 1126) 704fc8d467

mesh-catalog's 140 module directories are 420 forge reads per merge, past the 30 s bound
of a handled event while the forge is busy: the merged message of #535 timed out on every
try, and #3fb7951 needed seven (issue 1458 met the same kind at 131). A tree hash names its
bytes, so what was read at a tree is kept in the new state inputs and in memory, and a
merge reads only the directories it changed.

The forge bridge keeps the directories a failed try read whole, so a large merge converges over its tries (issue 1126) 89be76bb6a

Without it the first merge after the bridge starts, with nothing kept, still reads all 140
catalogue directories in one 30 s try and times out as before.

indexer_take_in reads a held module again at the source's head when its record stands behind it (issue 1126) 790653dfde

A merge whose merged message never reached the indexer had no recovery: take-in answered
held for mesh-vault, whose record stayed at 78d7ef7 after mesh-catalog#535.

indexer_take_in reads a held module again only when its record is behind the source's head, and the two new tests are rows (issue 1126, review)

mesh/delivery merged: source.main moved to novox/mesh-delivery@31bbf207ef3b

Details

mesh/delivery-group group fix/1422-trunk-unknown merged: every member is merged

Details

mesh/check being carried to 2b801f5bd49a on 31bbf207ef3b | from ae7ec702a5e9, mesh-catalog 9fd95693db54

2b801f5bd4

A head older than the record, or one whose order the forge cannot say, would have written
an older manifest back. The new tests become rows of existing ones: the repository is over
its 500 unit tests.

mesh-tools commented

2026-10-28 10:20:52 +00:00

Delivery novox/mesh-delivery@2b801f5bd49a — merged since 2026-10-28T10:34:54Z

Source — source.main moved to novox/mesh-delivery@31bbf207ef. Whether it reached the machines is mesh-delivery-view's, derived from source.main, the wants and the haves.

Group fix/1422-trunk-unknown: novox/mesh-catalog@f5d4b2f86e, novox/mesh-delivery@2b801f5bd4

composed together: pass — the checker's verdict at the members' heads (3bee9e0518c4)

Transitions

2026-10-28 10:20 (new) → proposed (announced): novox/mesh-delivery#94's head announced
2026-10-28 10:20 proposed → ready (accepted): a verdict names this commit and the trunk's head as its trunk commit, the gate passed or warned, and the repository's own check did not fail; or the check gates no merge (check-gates-merge false)
2026-10-28 10:34 ready → merged (merged): on the trunk its modules follow: merged there

The commit's note under refs/notes/mesh-plan keeps every transition of this change up to its merge: git log --notes=mesh-plan.

**Delivery** `novox/mesh-delivery@2b801f5bd49a` — **merged** since 2026-10-28T10:34:54Z **Source** — source.main moved to novox/mesh-delivery@31bbf207ef3b. Whether it reached the machines is `mesh-delivery-view`'s, derived from `source.main`, the wants and the haves. **Group** `fix/1422-trunk-unknown`: novox/mesh-catalog@f5d4b2f86ed4, novox/mesh-delivery@2b801f5bd49a - composed together: pass — the checker's verdict at the members' heads (3bee9e0518c4) **Transitions** - 2026-10-28 10:20 (new) → proposed (announced): novox/mesh-delivery#94's head announced - 2026-10-28 10:20 proposed → ready (accepted): a verdict names this commit and the trunk's head as its trunk commit, the gate passed or warned, and the repository's own check did not fail; or the check gates no merge (check-gates-merge false) - 2026-10-28 10:34 ready → merged (merged): on the trunk its modules follow: merged there The commit's note under `refs/notes/mesh-plan` keeps every transition of this change up to its merge: `git log --notes=mesh-plan`.

mesh-tools merged commit 31bbf207ef into main

2026-10-28 10:34:51 +00:00

mesh-tools deleted branch fix/1422-trunk-unknown

2026-10-28 10:34:51 +00:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: novox/mesh-delivery#94