Commit Graph
635 Commits
Author SHA1 Message Date
jschoubben 94cee5d199 A data flag that differs alone moves no data: the judgement reads the data items first, and says the flag loudly (issue 1126)
novox's mongodb and mssql asked a yes for moving their data on a toolchain move: the running side, read back from the
kept declaration, carries no data flag, and judge.go returned moves-data on the flag before it compared the data
directories and files, so a flag mismatch both raised a false ask and hid a real move. Now moves-data only when a
data item (a directory or file within the data places, or a process of a module with data) changes; a flag-only
difference is neither, with a WARNING in its words. One row changed and one added in TestJudgeMove.
2026-10-29 21:03:16 +01:00
jschoubben 52056010fe A module composed as what runs of it serves its tools as they run, so the runtime no longer drops it (issue 1126)
After the take-over of placement on 2026-10-29, shanks's runtime stopped serving the folded systemd-resolved (the
node-resolver seat's verbs answer "nothing serves") and novox's the folded nats. Neither has a want: each runs inside
its node's own composition at a build of the old form (Scope.InOwn), and is composed as Carried of the indexer's
manifest. That manifest is the catalogue's, which no build resolved, so it has no bundles (catalogue/build.go: Bundles
are set only by a build's resolution), and the runtime's MESH_TOOL_MODULES and MESH_TOOL_ENV are made from bundles
(catalogue/runtime.go). The controller composed them from the resolved manifest its store kept.

The kind: a module composed as what runs of it (Stub, Carried: held, or folded at an old build) contributes to the
others from its manifest, not from what runs. For a held module with a new build it is the other way round: its new
bundles are served, which were never delivered.

- The quick fix: read the old-form build's resolved manifest for a folded module. One instance; a held module still
  hands its new bundles.
- The thorough fix, chosen: compose.ToolsRunning reads what the node's runtime serves as it runs (its process's
  MESH_TOOL_MODULES and MESH_TOOL_ENV, each bundle's archive), frombus gives it as Node.RunsTools, and every stubbed
  module of the machine's runtime (held or carried) contributes those bundles, whatever its manifest says.

Tests: two rows (folded, held) added to TestCompositionsAreTheDeclarationSplitByModule; no new test function.
2026-10-29 21:03:16 +01:00
mesh-tools 307b6234e6 Merge pull request 'A held module offers what it runs, and its hold no longer spreads to the modules that read from it (issue 1960)' (#131) from fix/1960-no-hold-spread into main 2026-10-29 15:35:24 +00:00
jschoubben dbcaa29a1a Never hold a module because a module it reads from is held: it composes with what the held one last offered (issue 1960)
mesh/delivery merged: source.main moved to novox/mesh-sdk@307b6234e6bb
mesh/delivery-group group fix/1960-no-hold-spread merged: every member is merged
mesh/check pass dbcaa29a1a38 on 394fec8a63c1 | unit 3318 run, 1951 kept, 38 skipped | integration 3994 run, 2125 kept, 38 skipped | over bound: unit, …
On novox on 2026-10-29 de-spiegel, held for an unset config value, held mailu, route-proxy and nftables ("it reads
what de-spiegel contributes or listens on"), and mailu held postgres. mesh-sdk #128 (issue 1126) stopped the hold
spreading to consumers and collectors but kept dependents(): the provider of what a held module contributes to, and
the filter where it listens, were held behind it, transitively.

- Quick fix: drop dependents() alone. The filter and the proxy would then read a held module's new manifest, which
  may not be what runs (review of the apply, round 2).
- Thorough fix, chosen: no module is held behind another, in the apply and in the answer (dependents, heldBehind and
  the held-behind path of heldAsToday removed). A held module offers what it last offered: its contributions and
  listens from its manifest at the build its own composition runs at (ranOf), else the manifest it is composed from.
  Only the part of its offer that does not compose is left out, by provision (culprit): de-spiegel's routes leave
  route-proxy without them, while mailu keeps its smtp account; the hold names it.

Tests: rows changed in the existing answer test (de-spiegel held, mailu and proxy composed; site held, the filter
composed with what runs; leave-out by provision). No new test function.
2026-10-29 15:54:21 +01:00
mesh-tools 394fec8a63 Merge pull request 'A held module composes only as what runs of it, so its hold no longer spreads to the modules that use it (issue 1126)' (#128) from fix/1126-held-as-runs into main 2026-10-29 00:07:03 +00:00
jschoubben c2999a4e13 A held module is no dependence of its neighbours: its hold stops spreading (issue 1126)
mesh/delivery merged: source.main moved to novox/mesh-sdk@394fec8a63c1
mesh/delivery-group group fix/1126-held-as-runs merged: every member is merged
mesh/check pass c2999a4e13cc on d4c61ee8ffd0 | unit 3054 run, 234 kept, 36 skipped | integration 3075 run, 266 kept, 29 skipped | over bound: unit (sl…
novox's seed p-20261028t231141-ec84b4 refused 11 modules "what it composes of ..., held, is not what runs of them":
de-spiegel's hold spread through mailu to postgres, nftables and route-proxy, and from postgres to every consumer, and
gitea, held so, refused the node: "its credential of postgres-database is handed to the provisioner and the module is
not: nothing would fill it". Three rounds each moved the error. The kind: a held module still took part in the
composition as a dependence, so its hold spread.

- The quick fix: let heldAsToday also follow dependsOnHeld's words, and skip the handed refusal for gitea. It moves the
  error once more and leaves every node's applying path holding consumers.
- The thorough fix, chosen: a held module composes only as what runs of it. composeEach no longer compares the
  composition with each held module as its stub (dependsOnHeld removed): a consumer or collector composes against what
  runs of the held module like any other module, and a real difference is still a named difference in the answer, a
  module's own refusal still its refusal. A stub (compose.Stub, catalogue.Manifest.Stubbed) leaves the credentials and
  own secrets it would hand the provisioner out of its plan as they run, never refused for want of a plan. Only what a
  held module's new manifest says to others in their own files is still held behind it (dependents: the provider of
  what it contributes to, the filter where it listens), which heldAsToday already follows.

Tests: rows added to existing tests (a held provider with a binding consumer, which the old code held in the seed's
words; a held feeder of a collector; a held module with a handed credential and own secret). No new test function.
2026-10-29 00:49:00 +01:00
mesh-tools d4c61ee8ff Merge pull request 'The seed leaves a module as it runs today when it is held, failed or at an old-form build, and the modules held only behind it too (issue 1126)' (#127) from fix/1126-seed-oldbuild into main 2026-10-28 23:05:04 +00:00
jschoubben 76f7c4d223 A stub keeps its items and their templates, so a held module still contributes to other modules' collections; it drops only the templates for its own files (issue 1126)
mesh/delivery merged: source.main moved to novox/mesh-sdk@d4c61ee8ffd0
mesh/delivery-group group fix/1126-seed-oldbuild merged: every member is merged
mesh/check pass 76f7c4d2238f on d747c1128e42 | unit 3018 run, 270 kept, 36 skipped | integration 3039 run, 302 kept, 29 skipped | over bound: unit (sl…
2026-10-28 23:56:33 +01:00
jschoubben f609be7864 Keep as today what a node keeps today and what is held only behind it, saying how; a stub hands and collects nothing of its own (issue 1126) 2026-10-28 23:49:22 +01:00
mesh-tools d747c1128e Merge pull request 'L3: the vault mints bus accounts, the grants part and the node seat node-engine (issue 774, ADR 0561), rebased onto L2' (#126) from feat/774-vault-mints-sdk into main 2026-10-28 22:27:36 +00:00
mesh-tools f9afe32888 Merge pull request 'A node's answer holds as today what it holds today, and the seed carries its nodes' differences, named (issue 1126)' (#125) from fix/1126-seed-sdk into main 2026-10-28 21:59:48 +00:00
jschoubben 77b48a4f1a Name in the seed's brief the first differing field's value as it runs and as composed, so the operator sees what changes (issue 1126)
mesh/delivery-group group fix/1126-seed-held rejected: composed together: the checker's verdict at the members' heads (63905bd4b284): the composition
mesh/delivery merged: source.main moved to novox/mesh-sdk@f9afe328881f
mesh/check being carried to 77b48a4f1aff on f9afe328881f | from 53f26493697e
2026-10-28 22:26:05 +01:00
jschoubben e84a121b8a Let the seed carry the differences its nodes found, each named in its brief for the operator to approve knowingly; a refusal or silence still fails (issue 1126) 2026-10-28 22:24:57 +01:00
jschoubben 06b13ec1cc Show in a proposal's brief what stays held as each node holds it today (issue 1126)
mesh/delivery-group group fix/1126-seed-held rejected: composed together: the checker's verdict at the members' heads (2b8e3078c117): the composition
mesh/delivery closed: a newer head of the same pull request
mesh/check failed: the composition: the manifest of mesh-delivery-assigner uses what mesh-controller at ba4d15657343d3116bd6657f85e53630a7dbabf2 (mesh…
2026-10-28 22:17:57 +01:00
jschoubben 0830fcb25e Hold that a refused module not held today keeps the answer refused (issue 1126) 2026-10-28 22:17:40 +01:00
jschoubben b03bf4c118 A node's answer holds as today what it holds today, so a seed that reproduces today is the same (issue 1126) 2026-10-28 22:13:37 +01:00
mesh-tools 53f2649369 Merge pull request 'A build puller parked at capacity 0 wakes when its settings file changes, and retries a failed setting follow (issue 734's kind)' (#122) from fix/park-wakes into main 2026-10-28 21:03:13 +00:00
jschoubben 8eafc0c5ce The placement reader names the seat mesh-placement's state, held by the assigner, as L2 moved it (issue 774, ADR 0567)
mesh/delivery-group group feat/774-vault-mints-l2 checking: the checker's verdict is out of date: novox/mesh-delivery@f6df07605dd0 was judged on trunk 5b2e5bc5…
mesh/delivery merged: source.main moved to novox/mesh-sdk@d747c1128e42
mesh/check checking 8eafc0c5ce3c on f9afe328881f | novox/mesh-sdk#123 in group feat/774-vault-mints-l2
2026-10-28 21:54:04 +01:00
jschoubben 50de0a5a71 The node seat's acts reads one act by its id (issue 774, ADR 0563 §9) 2026-10-28 21:54:04 +01:00
jschoubben 2323ba19db The node-engine's declared state acts: created by the grants part as a placed module's, its manifest placed on every node, and written by each node's engine under its own keys alone (issue 774, ADR 0563 §2, §3) 2026-10-28 21:54:04 +01:00
jschoubben 63607ba169 The vault's secret-ask is said to answer the callers its manifest names (issue 774, ADR 0563 §8) 2026-10-28 21:54:04 +01:00
jschoubben 82c54870ed A module's trusted list of callers in the SDK, so the vault's secret-ask answers the callers its manifest names as the runtime proved them (issue 774, ADR 0563 §8) 2026-10-28 21:54:04 +01:00
jschoubben 79231a9e6c A node's engine answers its own node's node-engine verbs, says the seat's events and asks the operator as node-engine-<node>, and no other account answers them (issue 774, ADR 0563 §1, §3, §6) 2026-10-28 21:54:04 +01:00
jschoubben 222d605f6f The node seat node-engine: held by every node's engine, its six verbs and three events, and no module may claim it (issue 774, ADR 0563 §1) 2026-10-28 21:54:04 +01:00
jschoubben 3a8997c58a Bundle the mesh's authorities by certificate, and check a node's certificate against every one a bundle holds (issue 774, ADR 0550 §2) 2026-10-28 21:54:04 +01:00
jschoubben d6db3f38e6 The vault signs: the contract of its state signers, its verb sign and its event signed, a node's grants to ask it and to read what it trusts, and the authority composed as a bundle (issue 774, ADR 0550 §2) 2026-10-28 21:54:04 +01:00
jschoubben 18a47ac3c7 Placed: a module in the first revision heard of a node's key stands since before it, so the seed's placement does not leave out what already ran (issue 774) 2026-10-28 21:54:04 +01:00
jschoubben 1386219d4d One reader of the catalogue's placement and the toolchain gate over it, in place of the three copies over the controller's assignments (issue 774, ADR 0549 §3, issue 1940); the assigner may read the lease (ADR 0549 §5) 2026-10-28 21:54:04 +01:00
jschoubben 10d4d2f845 Checker seat test: the try replaces both shell commands (issue 774) 2026-10-28 21:54:04 +01:00
jschoubben b3100c55a1 Checker seat: the try's verb also replaces docker run postgres, as the controller's try did (ADR 0547 §10, issue 774) 2026-10-28 21:54:04 +01:00
jschoubben 0062bdc4a6 buscompose: the assigner may read the controller's lease, to choose where grants are read from (issue 774) 2026-10-28 21:54:04 +01:00
jschoubben a1a688b19e delivery: the grants part's grant keys and the source the assigner reads (a1; issue 774)
Why: the grants part writes grants.<module>.<fp> in shadow beside the controller's; the assigner switches source only when the lease is handed away.
2026-10-28 21:54:04 +01:00
jschoubben 54b17164ef Checker seat test: fold into the closed-set seats test, no new test function (issue 774) 2026-10-28 21:54:04 +01:00
jschoubben 3ed9a0e342 Checker seat: the try's end by hand is a verb too (issue 774) 2026-10-28 21:54:04 +01:00
jschoubben 9b5657a3d6 Seat mesh-delivery-checker: the try and its read are verbs of the checker's own seat, replacing sh merge-check.sh (ADR 0547 §10, issue 774) 2026-10-28 21:54:04 +01:00
jschoubben 1d4637e46b Name the vault's verb issue-account on its seat, so its holder may serve it (issue 774, ADR 0561 §9) 2026-10-28 21:54:04 +01:00
jschoubben fefa8b88ea Say when the lease is handed away, let the vault and the grants part read it, and let the grants part create the consumers an account needs, so both turn on at the lease write alone (issue 774, ADR 0561 §3, §6) 2026-10-28 21:54:04 +01:00
jschoubben 8d9b3f779d Build a minted account's value as the controller does, byte for byte, so a runtime reads the vault's unchanged (issue 774, ADR 0561 §2) 2026-10-28 21:54:04 +01:00
jschoubben da506bb3b8 Say which bus accounts the grants part wants, in a state only the vault reads, and what the vault says when it mints one, so the vault can mint from the grants part's word (issue 774, ADR 0561 §2) 2026-10-28 21:54:04 +01:00
mesh-tools af08be36a2 Merge pull request 'A waiting ask hears its step's end even when its own watch goes deaf (four lost checks, 2026-10-28)' (#121) from fix/lost-check into main 2026-10-28 20:40:58 +00:00
mesh-tools 6b43afa2ab Merge pull request 'One predicate for a module the mesh provides itself (mesh-wireguard), used by the placement check (issue 1126)' (#120) from fix/1126-provided into main 2026-10-28 20:39:20 +00:00
jschoubben f91340fe82 buildrun: open a steps watch again that closed having heard nothing, up to 3 closes in a row
mesh/delivery merged: source.main moved to novox/mesh-sdk@af08be36a27a
mesh/delivery-group group fix/lost-check merged: every member is merged
mesh/check checking f91340fe82ef on 6b43afa2ab38 | novox/mesh-sdk#121 in group fix/lost-check
A watch that closed at once was left for good, so an ask whose own watch
went deaf was read only at a reconnect (review of fix/lost-check).
2026-10-28 20:35:02 +01:00
jschoubben 6aeb401095 buildrun: read an ask's step at every step end of its build module, so a deaf watch loses no end
mesh-catalog #553's verdict stayed running after its step passed on shanks
at 19:09:31Z on 2026-10-28: the watch of that one step key, opened at
18:46:05Z, heard neither the start nor the end and never closed. HeardEnds
stands between Asker and each ask's watch: nudged at each step end of the
same build module (HearEnds: one watch per steps state) and at a reconnect,
every waited ask reads its keys now; a watch that closes is opened again, at
most 3 times. The checker and the builder both wait through it.

Tests: a section of the asker's test (deaf, nudged by another module and by
its own, a watch that closes); counts unchanged.
2026-10-28 20:29:30 +01:00
jschoubben 5228a5782d Name the modules the mesh provides itself in one predicate, so the seed's check takes the private network as the composer and the grants part do (issue 1126)
mesh/delivery merged: source.main moved to novox/mesh-sdk@6b43afa2ab38
mesh/delivery-group group fix/1126-provided merged: every member is merged
mesh/check checking 5228a5782d5c on d7c790d2422c | novox/mesh-sdk#120 in group fix/1126-provided
2026-10-28 20:12:51 +01:00
jschoubben 9e5666b4d6 Settings watch: arm again when the kernel drops it (IN_IGNORED), said once; the test waits on the arm, not a sleep
mesh/delivery merged: source.main moved to novox/mesh-sdk@53f26493697e
mesh/delivery-group group fix/park-wakes merged: every member is merged
mesh/check pass 9e5666b4d6f5 on af08be36a27a | unit 5960 run, 332 kept, 18 skipped | integration 6012 run, 365 kept, 30 skipped | over bound: unit, in…
2026-10-28 20:08:21 +01:00
jschoubben c006aa8242 Build puller: wake on a settings file write, retry a failed config follow at the next event, say a capacity of 0 and why 2026-10-28 20:01:46 +01:00
mesh-tools d7c790d242 Merge pull request 'The seat mesh-placement's contract, and node grants for its states (issue 1126, delivery 1 of 2)' (#119) from feat/1126-grants-first into main 2026-10-28 18:21:33 +00:00
jschoubben 38dbe1d3a3 Let a node agent hear the creation of the seat's placement and candidates, so it binds them with no timer (ADR 0567 §9, issue 1126)
mesh/delivery merged: source.main moved to novox/mesh-sdk@d7c790d2422c
mesh/delivery-group group feat/1126-grants-first merged: every member is merged
mesh/check pass 38dbe1d3a388 on 98f2f2406921 | unit 1909 run, 715 kept, 3 skipped | integration 1931 run, 813 kept, 3 skipped | over bound: unit (slow…
2026-10-28 18:47:13 +01:00
jschoubben a2dbcf670b Bound a proposal's answers and ask by pulse keys whose expiry the holder hears, and name a hand-back proposal, so no timer waits (ADR 0567 §4, §7, issue 1126) 2026-10-28 18:43:39 +01:00
jschoubben e987a8280b Let a module's config values be proposed before the lease write: only a layer's places and accesses wait (ADR 0567 §13, issue 1126) 2026-10-28 18:42:05 +01:00