216 Commits

This Branch

This Branch

All Branches

Author	SHA1	Message	Date
jschoubben	95f52afab2	Record the checksums of mesh-sdk feat/r4g 9205222a for the group's modules (R4g)	2026-10-23 16:26:11 +02:00
jschoubben	e4a40a1cb0	Pin mesh-sdk at feat/r4g's head 9205222a, the merge of lanes A, B, S8 and F1, so every module r4g touches builds against one SDK ... go.sum gains its lines from go mod tidy once mesh-sdk feat/r4g is pushed.	2026-10-23 16:22:02 +02:00
jschoubben	dd8e05142d	Merge branch 'feat/r4g-scripts' into feat/r4g	2026-10-23 16:20:17 +02:00
jschoubben	66be3dd0e4	Merge branch 'feat/r4g-view' into feat/r4g	2026-10-23 16:20:17 +02:00
jschoubben	527f87f94c	Merge branch 'feat/r4g-asks' into feat/r4g ... # Conflicts: # modules/mesh-delivery-checker/cmd/mesh-delivery-checker/judge.go # modules/mesh-delivery-checker/manifest_test.go # modules/mesh-delivery-checker/module.json	2026-10-23 16:20:17 +02:00
jschoubben	0463f21026	Merge branch 'feat/r4g-assigner' into feat/r4g	2026-10-23 16:19:45 +02:00
jschoubben	a4629bc649	Compute every module's fingerprint with no builder input and read no build-agent record for it, so a change of the builder rebuilds nothing (issue 1887 I1, ADR 0530 §1)	2026-10-23 16:14:19 +02:00
jschoubben	50836eb4d6	Format the watchman's integration test (issue 1887 A8b)	2026-10-23 16:14:03 +02:00
jschoubben	79ef768ea4	Show the backlog of build asks per language and kind: waiting, held by a node, and the oldest waiting ask's age (issue 1887 A8a)	2026-10-23 16:14:00 +02:00
jschoubben	a8ba5c2fd5	Mirror a waiting build ask as a pulse until a node writes its step, and drop the read of the controller's checks-running (issue 1887 A8b)	2026-10-23 16:12:43 +02:00
jschoubben	d0354574c6	Look the assigner's compositions bucket up at each check until it is there, so an assigner delivered after the checker started is read without a restart (issue 1297, review round 1 item 3)	2026-10-23 16:12:18 +02:00
jschoubben	69577f739d	Judge hq as any repository, so its pull requests get a verdict and mesh/check from its Go tests and no script judges it (issue 1887 A10, ADR 0527 §2)	2026-10-23 16:11:43 +02:00
jschoubben	d4977d9c9d	Start the checker whole with the assigner's verdict on the bus and read its line, the level 1297/AC3 asks (issue 1297, review round 1 item 2)	2026-10-23 16:11:31 +02:00
jschoubben	179664898a	Merge feat/r4g-asks into the view lane, keeping the checker's assigner compositions	2026-10-23 16:11:15 +02:00
jschoubben	77af82cfa2	Keep the asks state and write the ask about a held fingerprint, so the view's and the teller's read of mesh-delivery-assigner.asks is granted and has something to answer (issue 1297, review round 1 item 1; ADR 0528 clause 6)	2026-10-23 16:11:07 +02:00
jschoubben	f37e5d5fa1	Delete merge-check.sh; the checker's levels and the build module's fetch beside run what it ran (issue 1295, S8; ADR 0529)	2026-10-23 16:10:46 +02:00
jschoubben	d9d1e57e3a	Delete the checker's hold, its own capacity and places, its node moves and its read of the controller's plans-building, so a check waits only in the build seats' streams (issue 1887 A5b, closes 1685; ADR 0525 §4, §7)	2026-10-23 16:09:34 +02:00
jschoubben	e4b639c6de	Judge each pull request's check as one ask on the build seats' check stream, a re-judge on recheck, so the checker chooses no node and a lost node's check is redelivered, not moved (issue 1887 A5a, ADR 0525 §1, §3, ADR 0536 §4)	2026-10-23 16:05:55 +02:00
jschoubben	271d1ca759	Pin mesh-sdk at feat/r4g's head c3915ffb, the merge of lane B's layouts, so every r4g module builds against one SDK ... go.sum gains its lines from go mod tidy once mesh-sdk feat/r4g is pushed.	2026-10-23 16:01:15 +02:00
jschoubben	2227e7a81f	Merge branch 'feat/r4g-view' into feat/r4g	2026-10-23 15:59:55 +02:00
jschoubben	173178df86	Merge branch 'feat/r4g-assigner' into feat/r4g	2026-10-23 15:59:55 +02:00
jschoubben	9f27e2b1f8	Build each build module's part of a fingerprint as one ask on its seat's stream, named in the key before it is put, so the builder chooses no node, holds no capacity and never asks twice (issue 1887 A4a, A4b, A4c, ADR 0525 §1-§6)	2026-10-23 15:57:03 +02:00
jschoubben	d5868c178f	Take the words of a module the change removes from mesh-sdk in the indexer and the assigner, so the assigner's reading of a removal cannot drift from what the indexer writes (issue 1297, B3b)	2026-10-23 15:44:00 +02:00
jschoubben	7cbb1c4207	Wait for the assigner's composition only while the assigner is online, say at once that it is not judged when it is not, and take it in when the assigner writes it later, so no check waits its bound for a composition nobody makes (issue 1297, B3d)	2026-10-23 15:44:00 +02:00
jschoubben	4c98841d2a	Judge a repository the mesh knows no module of with both build modules, so mesh-lab's nested Go modules and its npm test both run (issue 1887 A12)	2026-10-23 15:41:57 +02:00
jschoubben	c9ee92a4ab	Pin the assigner and the checker to the SDK commit that holds the assigner's ask value, so both build on the layouts the view reads (issue 1297, subtask B1c)	2026-10-23 15:39:49 +02:00
jschoubben	0f7091e4ac	Place the change's own repository beside it at its trunk, so a shrink-only list test compares with the trunk's copy at $MESH_CHECK_BESIDE/<name> (issue 1887 A12, ADR 0529 §2)	2026-10-23 15:35:37 +02:00
jschoubben	ed98a7519e	Write every difference between a want and what the controller sends, and delete one that ended, so the view and acceptance test 17 read them on the bus (issue 1297, subtask B5b; ADR 0528 clause 4)	2026-10-23 15:31:11 +02:00
jschoubben	d7ba5aa91e	Compare each node's want with the fingerprint of the build the controller sends it, explaining a difference by the rule that holds the want back or saying it unexplained (issue 1297, subtask B5a; ADR 0528 clause 4)	2026-10-23 15:30:31 +02:00
jschoubben	016385751d	Write each node's want by the six rules, the canary the first online node by name, and hold a fingerprint whose health failed with its nodes set back, so R5's wants can be compared with what the controller sends (issue 1297, subtasks B4a-B4b; ADR 0429 §6, ADR 0528 clauses 1 and 3, ADR 0435 §3)	2026-10-23 15:29:57 +02:00
jschoubben	5091cab652	Read the assigner's composition into the checker's verdict in place of 'not yet run', failing a check whose composition failed, naming its machine (issue 1297, subtask B3d; ADR 0528 clause 5)	2026-10-23 15:27:59 +02:00
jschoubben	d8599855e1	Compose the machines a change touches, from the trunk's and the change's manifests and the controller's assignments on the bus, and write the verdict under the checker's key; what needs the facts snapshot is said as not judged (issue 1297, subtasks B3a-B3c; ADR 0446 clauses 1-3, ADR 0528 clause 5, ADR 0482 clause 3 as R4g delivers it)	2026-10-23 15:25:22 +02:00
jschoubben	57258b6fa4	Take the assigner's ask in the SDK's layout and name the asking in the view's answer, so a stale answer never fills a later asking (issue 1297 B9 and B10)	2026-10-23 15:22:32 +02:00
jschoubben	c2b7b8867e	Start the assigner whole on an embedded bus in an integration test: it says its three states and refuses to start without a bucket it reads (issue 1297, subtask B2b)	2026-10-23 15:19:49 +02:00
jschoubben	a48279b31a	Fill an assigner ask's answer from the view's answer under its view key, once and never for an ask that needs proof (issue 1297 B10)	2026-10-23 15:19:37 +02:00
jschoubben	034a6de9a7	Make the module mesh-delivery-assigner: its manifest's reads and states, its own credential and the states it announces, so the controller grants it what R5 reads (issue 1297, subtask B2a; ADR 0528, ADR 0429 §1)	2026-10-23 15:19:06 +02:00
jschoubben	e7ed48b81a	Give the command an answer verb, so the operator answers a hold or restore ask from a terminal as in the page (issue 1297 B9b)	2026-10-23 15:18:04 +02:00
jschoubben	f7dd4ab586	Let the view answer the assigner's asks that need no proof, under the ask's view key, and refuse the merge's Approve on the phone's behalf (issue 1297 B9a)	2026-10-23 15:17:40 +02:00
jschoubben	57b0030258	Show every difference between the assigner's wants and the controller's sends with its rule, so the operator sees what holds a node back (issue 1297 B8)	2026-10-23 15:15:32 +02:00
jschoubben	4ffe5f4d52	Start a never-judged head's check before the re-judges a trunk move asks, oldest first (issue 1897 S1) ... A new pull request waited 25 to 45 minutes behind every open pull request's re-judge after a merge. The semaphore becomes a ranked admission (ADR 0536 §4); checker_status names the waiting checks by kind.	2026-10-23 14:09:42 +02:00
jschoubben	2bdb8d0e00	Ask the Go build module to run the candidate when a merge changes it, and say in the line which program judged (issue 1898 K3) ... A change to mesh-build-go was judged by the installed module it replaces, so such pull requests were merged by hand. ADR 0535 §1, §2, §4, §5.	2026-10-23 14:06:57 +02:00
jschoubben	75a4821eff	Merge pull request 'Make checker_status count only steps that can still run, so a key a build module left running is not read as a full node (issue 1890)' (#24) from fix/1890-checker-stale-steps into main ... Reviewed-on: #24	2026-10-23 11:01:10 +00:00
jschoubben	e762028760	Make checker_status count only steps that can still run, so a key a build module left running is not read as a full node (issue 1890) ... The checker's own places are released when a step ends or its wait passes the step bound; checker_status read every 'running' key in the build modules' steps buckets as a step in flight, so keys left by stopped, restarted or unassigned build modules looked like full nodes. It now lists them under left_running with why (past the bound, module not on the node, module not running), and leaves hq-only groups out of the unjudged candidates.	2026-10-23 09:03:45 +02:00
jschoubben	36bf3b7143	Merge pull request 'The builder says it writes references only once its build modules proved it (issue 1886, ADR 0522)' (#23) from fix/1886-builder-says-references into main ... Reviewed-on: #23	2026-10-23 00:26:18 +00:00
jschoubben	9fad90ee89	Say that the builder writes references only once each build module answered one (issue 1886, ADR 0522 review) ... The references come from the build modules' own pushes, so a new builder beside an old build module writes none; saying it at start would send the controller to keys that fail not stored.	2026-10-23 00:52:50 +02:00
jschoubben	4b8b1c047c	Say in the builder's own bucket of instances that it writes references (issue 1886) ... The controller delivers from the builder's keys only while a running builder says its done keys name references; a builder from before R4e says nothing, so R4e's own delivery is asked of build-agent instead of waiting for ever. The say is put once started and taken away when stopped.	2026-10-23 00:48:28 +02:00
mesh-tools	b714eab995	Merge pull request 'R4e (mesh-delivery): the builder pushes and keys artifacts by reference, the checker's verdict and try, the view, and the checker's owed state read by the forge' (#22) from feat/r4e into main	2026-10-22 22:13:05 +00:00
jschoubben	efca168c55	Stop the forge bridge performing mesh/check: the forge module writes it (R2-1) ... R4e round 2 finding R2-1, ADR 0521: nothing in the delivery plan orders the forge module before the bridge, and the forge module's durable consumer keeps its deliver policy when its filters change, so a check.status emitted before it consumed the event was lost. The forge module now watches the checker's owed state and writes mesh/check itself, once per revision. The bridge emits no check.status, declares no emits, and neither reads the checker's owed nor hears its stream; OwedAct still refuses mesh/check. The integration test shows the bridge calls no tool and publishes nothing of its own when the checker owes mesh/check.	2026-10-22 21:40:09 +02:00
jschoubben	19d06163de	Announce the checker's states, owed readable, so the forge module may read it (R2-1) ... R4e round 2 finding R2-1, ADR 0521: the forge module writes mesh/check from the checker's owed state, read by a watch, because no event order between the forge bridge and the forge module is kept by the delivery plan. A reader is granted from the owner's announcement (ADR 0495 §2-§3), and the checker announced nothing. It now says every state it keeps, with owed, previews and verdicts readable: the forge module and the bridge read owed, the indexer previews, the gate verdicts.	2026-10-22 21:38:53 +02:00
jschoubben	0997b071a3	Pin every R4e member module to mesh-sdk's final R4e head 104533e8, which holds review fixes S1 and S2, so the group pins one SDK	2026-10-22 21:29:20 +02:00