Commit Graph
319 Commits
Author SHA1 Message Date
jschoubben 3d7d65edce Assigner: a node running an old-form build gets its first want at F once the canary runs F healthy (issue 1945)
An old-form build (not src2:) has no composition, so no want can ever name
what such a node runs. Design limit, decided by the lead (lead-47): its first
want names the newest F, by every rule as usual - the canary (rule 3: this
node is the canary, or the canary runs F healthy), everything F needs (built,
not held, granted, composed) and the providers (rule 5). Otherwise nothing is
written and the neither condition stays. The canary stays in force, and
old-form modules (clipmenu and screen-lock on g14) leave neither list by
themselves, without a hand-written want.

The code already behaved so (an old-form have falls through to Holds); the
case is now named in Wants, and proven: a unit test for both sides of the
canary, the providers and the grant, and the integration test gives g14 F
and clears its neither condition once the canary ace runs F healthy.
2026-10-25 15:32:35 +01:00
jschoubben 81182e9dcd Assigner: a running node's first want names the fingerprint it runs, never a newer one, and only a node that runs the module is on neither list (issue 1945, review of f7b23033)
A first want at the newest fingerprint passed over the canary for every module right after the switch. It now names what the node
runs: F when it runs F (by the rules without which there is nothing to want), or an older fingerprint of the new form whose
composition and grant exist, held back from F by the rule; otherwise nothing is written and the node stays on neither list. A new
assignment, with no have, waits for the rules and raises no condition (1945/AC19-AC21).
2026-10-25 15:26:03 +01:00
jschoubben f7b23033a2 Assigner: the integration test of wants on the bus proves a running node's first want and the neither condition; held-condition reads filter by key (issue 1945, try-1792937443523744996) 2026-10-25 15:18:07 +01:00
jschoubben 8daa5977f9 Assigner: a node running a module with no want gets its first want past the ordering rules, and a module on neither list is raised (issue 1945)
The node's own composition gives up every module with a fingerprint of the new form, so such a module is applied only through its
want. The assigner wrote no want while a rule held it back and the node had none, so clipmenu, screen-lock, i3status-rust and every
module requiring x11-display lay on neither list on g14 and shanks: xorg's health is unknown (rule 5) and the canary never ran them
healthy (rule 3), and their files were removed and never made again.

A node that runs a module and has no want for it now gets its first want by the rules without which there is nothing to want (built,
not held, granted, composed); the canary and providers rules order every move after it. Any module still on neither list is raised
as an urgent condition naming module, node and rule, written to the journal, and cleared once it has a want (1945/AC19-AC21).
2026-10-25 15:10:34 +01:00
mesh-tools 97e605edb0 Merge pull request 'Raise the assigner's asks on the operator channel, lift a hold once every node runs it healthy, and keep its cause (issue 1906)' (#40) from fix/r6-hold-asks into main 2026-10-25 13:27:29 +00:00
jschoubben 7ef796b04b Format the assigner's integration test (issue 1906)
mesh/check pass 7ef796b04b92 on 97e605edb0ad | unit 154 run, 404 kept | integration 164 run, 469 kept | composition passed on 2 machines
mesh/delivery merged: source.main moved to novox/mesh-delivery@97e605edb0ad
2026-10-25 14:16:58 +01:00
jschoubben ec93585736 Ask a hold again after a restart, leave a merge ask room, take an answer only for its asking (issue 1906, review)
A hold written again after a lift and a restart reused its first ask id, which
the router drops; the hold's time is now its asking. Seven holds could take the
router's three places for the teller and keep a merge Approve from being asked;
the teller keeps at most two assigner asks open. A release or keep answered for
one hold carried over to a later hold of the same fingerprint; an answer now
counts only for the asking its warrant names.
2026-10-25 14:15:44 +01:00
jschoubben 6b63f13da6 Raise the assigner's asks on the operator channel, lift a hold once every node runs it healthy, keep the hold's cause (issue 1906)
Seven modules stayed held on every machine and the operator was never asked:
the teller only filled assigner asks from the view, a single failed check held
a fingerprint for good, and the held record kept no cause. The teller now asks
every assigner ask as it asks merge asks and writes the phone's answer under
the ask's key; the assigner lifts a hold whose fingerprint runs healthy on
every node that runs the module, says so once and takes its ask back; the held
record names health, phase, when and the node's words.
2026-10-25 14:06:58 +01:00
jschoubben 2268f94ef2 Merge pull request 'The view lists a node's own want among the assigner's wants (issue 1945)' (#39) from fix/r6-node-composition into main
Reviewed-on: #39
2026-10-25 00:18:11 +00:00
jschoubben 48aaecf556 List a node's own want among the assigner's wants, never as unreadable (issue 1945)
mesh/check checking 48aaecf556c5 on 650a06ceab1d | novox/mesh-delivery#39 in group fix/r6-node-composition
mesh/delivery-group group fix/r6-node-composition held: a member is held: novox/mesh-sdk@10b6e202320f (merged without a passing check: only a person decides wh…
mesh/delivery held for a person: merged without a passing check: only a person decides what follows
2026-10-25 01:50:00 +02:00
jschoubben 650a06ceab Merge pull request 'R6: the assigner writes wants by rule 7 with the composition digest; the watchman and the view follow wants and haves (issue 1906)' (#38) from feat/r6 into main
Reviewed-on: #38
2026-10-24 23:17:46 +00:00
jschoubben 3e863884fd gofmt the passer's fields after main's merge
mesh/delivery-group group feat/r6 merged: every member is merged
mesh/delivery delivering: 0 machine step(s) passed
mesh/check checking 3e863884fda8 on 650a06ceab1d | novox/mesh-delivery#38 in group feat/r6
2026-10-24 23:33:59 +02:00
jschoubben c21bed20a3 Pin mesh-sdk to feat/r6 748c564, which carries main's composer (issue 1943) 2026-10-24 23:06:02 +02:00
jschoubben aa31e85665 Merge main into feat/r6: keep R4g's composer at a group's mesh-sdk head beside R6's wants and asks
The assigner reads mesh-controller.builds again (the composer's runningCommits reads the build record of every open
assignment), puts its compose asks on node-build-go and says its conditions on operator-channel.
2026-10-24 23:03:42 +02:00
jschoubben 171bddacab Merge pull request 'R4g finish: the composer as a job on node-build-go's check stream, merge-check.sh gone, reads-outside markers (issues 1295, 1621-1623, 1935)' (#35) from feat/r4g-finish into main
Reviewed-on: #35
2026-10-24 20:57:15 +00:00
jschoubben 65496fc4f7 Merge main into feat/r4g-finish, keeping #37's same-branch mesh-sdk pin rule beside the node-engine check's controller placing (issue 1295)
mesh/delivery delivered
mesh/delivery-group group feat/r4g-finish merged: every member is merged
mesh/check checking 65496fc4f7a6 on 171bddacabe2 | novox/mesh-delivery#35 in group feat/r4g-finish
2026-10-24 21:54:46 +02:00
jschoubben 36a1fd197f The switch exception skips the record hold only for a composition the controller judges unchanged; pin mesh-sdk d7a9b96 (issue 1906, ADR 0540 §9) 2026-10-24 21:18:48 +02:00
jschoubben 9a5b3d4881 Pin the assigner, watchman and view to mesh-sdk feat/r6 7f9cc0e (issue 1943) 2026-10-24 20:43:46 +02:00
jschoubben e263049551 Right after the switch a node running F with a have that names no composition is given its want at F with no record hold, the bus included (issue 1906, review round 2 N1) 2026-10-24 20:35:17 +02:00
jschoubben 0f7bc9f531 Pin the assigner, watchman and view to mesh-sdk feat/r6 e45bdc0 (issue 1943) 2026-10-24 20:21:57 +02:00
jschoubben da31740f00 The view's motion events carry the time they are about and call a held or past-bound module stopped, as the messenger reads them (issue 1906 D10) 2026-10-24 18:31:29 +02:00
jschoubben a3815051cc The view says modules in motion: moving, settled and stuck, once for a node, a module and a composition (issue 1906 D10, for the messenger) 2026-10-24 18:27:01 +02:00
jschoubben 8280506cee Start the world's change counter, which the pass of a held backup reads (issue 1906, review round 1) 2026-10-24 18:22:22 +02:00
jschoubben 3b00f3890d Review round 1: a user-list-only revision of the bus is written at once at the fingerprint it runs, before the record hold; pin mesh-sdk a523a8c (issue 1906) 2026-10-24 18:17:25 +02:00
jschoubben fcac03a4f2 Review round 1: the bus with no want is compared with its have; a backup ask is bound to its composition and number and each answer is a distinct want value; the view reads wants and haves, with dotted module names (issue 1906) 2026-10-24 18:15:13 +02:00
jschoubben 14c20b21ee Pin the assigner, watchman and view to the head of mesh-sdk feat/r6 (issue 1943) 2026-10-24 17:52:44 +02:00
jschoubben 333e0cd07d The checker reads a kept manifest as the indexer and the assigner do, so a value falling back to a predefined key is not logged unreadable (issue 1906, issues 1751 and 1921 class) 2026-10-24 17:42:24 +02:00
jschoubben 03e5951dc5 Merge remote-tracking branch 'origin/main' into feat/r6 2026-10-24 17:41:53 +02:00
mesh-tools 4c1d36c9cd Merge pull request 'A try accepts a mesh-sdk pin that is the head of mesh-sdk's branch of the same name, so a group's members can be tried before its pull requests (issue 1921)' (#37) from fix/1921-try-same-branch-pin into main 2026-10-24 15:37:07 +00:00
jschoubben 37d0ebba7a Let the view's manifest test read the emits key the delivered event adds 2026-10-24 16:47:33 +02:00
jschoubben 86fd0bbe96 The view says its event delivered once for each source revision whose wants and haves are all there (issue 1906 D9, for issue 1946) 2026-10-24 16:47:09 +02:00
jschoubben 37a5ab8f1c A node that runs a module of policy record with no want yet is asked like any other, not treated as one that never ran it (issue 1906 D4) 2026-10-24 16:43:01 +02:00
jschoubben 55f19283f4 Do not wait twice at the cleanup for an assigner that ended on its own, which made the refused-start test take ten seconds a bucket 2026-10-24 16:42:28 +02:00
jschoubben e7a3ff1e7e The record policy is the one the controller holds beside the composition, the operator's own choice included; pin mesh-sdk 64a502d (issue 1906 D4, ADR 0539 §14) 2026-10-24 16:41:58 +02:00
jschoubben 991589fdd2 A try of a branch accepts a mesh-sdk pin at the head of mesh-sdk's branch of the same name, so a milestone's repositories can be tried before their pull requests (issue 1921)
mesh/delivery delivered
mesh/check checking 991589fdd20a on 4c1d36c9cd7f
2026-10-24 16:41:39 +02:00
jschoubben f947070b7b The view shows the controller's judgement of the move beside each want (issue 1906 D7, ADR 0539 §14) 2026-10-24 16:38:24 +02:00
jschoubben 04c2c60f02 The assigner acts on answers: a release writes the held want again, a backup answer goes into the node's want, a record move waits for its yes, a restore answer sets a data module back (issue 1906 D4) 2026-10-24 16:37:37 +02:00
jschoubben 0df20b5d77 The assigner's seventh rule: a want names the digest of its composition and waits for it; a revision at the same fingerprint is a new want; the node's own want (issue 1906 D2, ADR 0539 §6) 2026-10-24 16:33:48 +02:00
jschoubben 806c8bf0e4 Pin the assigner, watchman and view to mesh-sdk's release 6 head (issue 1943 S3) and give their test wants a composition digest 2026-10-24 16:32:09 +02:00
jschoubben 12b01badb0 Prove the assigner replaced mid-pass resumes with nothing lost or written again (issue 1906 D8) 2026-10-24 16:27:34 +02:00
jschoubben 7b18696cd5 Give the data module's test manifest the directory and backup the catalogue asks of an irreplaceable item 2026-10-24 16:26:43 +02:00
jschoubben 204a59dab4 The assigner asks whether to restore a data module whose held fingerprint failed its health check, and raises the condition for a held fingerprint (issue 1906 D3) 2026-10-24 16:26:20 +02:00
jschoubben f048e2b443 The view shows each want beside the have of its node, and a source revision's derived state, in place of the comparison with sends (issue 1906 D7) 2026-10-24 16:23:51 +02:00
jschoubben 9d2e9dc6c9 The watchman mirrors each want of the assigner as a pulse owed by the node's agent until the have names its fingerprint (issue 1906 D6) 2026-10-24 16:22:02 +02:00
jschoubben 024a3787da End the assigner's comparison with the send: the controller sends nothing to compare with, so the differences bucket is kept read-only (issue 1906 D5, ADR 0539 §8) 2026-10-24 16:20:43 +02:00
jschoubben 4fd6b934c8 Read the assigner's wants, differences and the haves with tools, so the gap between wants and sends can be read live (issue 1906 D1) 2026-10-24 16:19:42 +02:00
mesh-tools fa46e7df7d Merge pull request 'The assigner reads a change's manifests as the indexer does, so a value falling back to a predefined key no longer fails every composition touching it' (#36) from fix/assigner-predefined-fallback into main 2026-10-24 14:09:58 +00:00
jschoubben 18b522184f Read a change's manifest as the indexer reads it, so a fallback to a predefined key is not unreadable for the catalogue's predefined.json the assigner never holds: a group touching pacman or xterm failed its composition on it (ADR 0348 §7, issue 1751)
mesh/delivery delivered
mesh/check pass 18b522184f73 on fa46e7df7dd1 | unit 128 run, 375 kept | integration 138 run, 426 kept | composition passed on ace
2026-10-24 15:03:14 +02:00
jschoubben 1d4dcdf41a Read a change's manifest as the indexer reads it, so a fallback to a predefined key is not unreadable for the catalogue's predefined.json the assigner never holds (ADR 0348 §7, issue 1751); pin mesh-sdk bfc3883a
mesh/delivery-group group feat/r4g-finish rejected: composed together: the checker's verdict at the members' heads (171a15384fa1): shanks
mesh/check checking 1d4dcdf41a22 on 4c1d36c9cd7f | novox/mesh-delivery#35 in group feat/r4g-finish
mesh/delivery superseded: a newer head of the same pull request
2026-10-24 14:57:36 +02:00
jschoubben f5fd74380c Pin to the group's new mesh-sdk head 19c1d359, whose composer names the reader's pinned SDK instead of a bare (devel) (ADR 0482 clause 5)
mesh/delivery superseded: a newer head of the same pull request
mesh/check failed: shanks: novox/mesh-delivery#35: modules/mesh-delivery-assigner/go.mod names mesh-sdk 19c1d35939bc, which is neither on mesh-sdk's t…
2026-10-24 14:12:10 +02:00