Commit Graph
1804 Commits
Author SHA1 Message Date
jschoubben 24f50c7c86 mesh-build-go: name go.mod among the narrowing test's paths, since mesh-sdk 104533e8 keeps only the manifest and the lock files beside them (review finding S1) 2026-10-22 21:39:55 +02:00
jschoubben e3db481491 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 21:29:19 +02:00
jschoubben 21b4925892 Merge remote-tracking branch 'origin/fix/r4e-k1' into r4e-int2 2026-10-22 21:28:24 +02:00
jschoubben 0fdfecfa29 K3: pin gitea to mesh-sdk 104533e8, the group's final SDK head after S1 and S2 (ADR 0500 §2) 2026-10-22 21:19:11 +02:00
jschoubben 97f0f4813c K2: keep the captured claude-code manifest in a test file (issue 1751)
The catalogue's paths check (ADR 0260) reads every file of a module but its documentation and tests, and
the verbatim manifest names claude-code's own paths; as a constant in a _test.go file it stays a real
capture without being taken for one module reaching into another's state.
2026-10-22 21:18:18 +02:00
jschoubben 8d3f81c2c0 Write mesh/check from the forge bridge's own event, never from a tool (K1)
Review finding K1 of R4e, issue 1301 S4 (ADR 0503 §3, ADR 0513 §4: the forge
requires mesh/check alone). gitea_check_status assumed only the bridge could
call it, but a tool is told no caller and node-tools invokes every tool, so any
session could set mesh/check to success and merge past the checker: the fault
issue 293 closed for the two statuses before it.

gitea now consumes mesh-delivery-forge-bridge.check.status, a subject the bus
grants to the bridge's own account alone, and check.ts writes mesh/check from
it: the source is checked, the head's whole sha and the state, and a revision
older than one already written on the head is dropped so a redelivered stale
pending never hides a verdict; a forge refusal throws, so the event is
delivered again. gitea_check_status is removed, and gitea_commit_status keeps
refusing mesh/check.
2026-10-22 21:14:01 +02:00
jschoubben 8f10d3192c K3: pin gitea to mesh-sdk 875c680d, the group's one SDK head (ADR 0500 §2)
gitea's Go code (cmd/repositories) changes in the R4e group, so it pins the SDK head every other member
pins; it still held 66e8fdaa.
2026-10-22 21:08:33 +02:00
jschoubben 1e3439429f K2: read a manifest for its build inputs in mesh-build-go and mesh-build-image (issue 1751, ADR 0514 §2)
The build step parsed the manifest as a judge of config values, so claude-code, dunst and xterm, whose
values fall back to a predefined key, got a fingerprint but their tools bundle was refused at the build.
The builder holds no published list of predefined keys; ParseManifestForInputs is its read. mesh-build-node
reads module.json with its own narrow struct and needs no change.
2026-10-22 21:08:12 +02:00
jschoubben e9e8c7c901 Pin every R4e member module to mesh-sdk's final R4e head 875c680d, which holds main's #68 and #69, so the group pins one SDK 2026-10-22 20:45:17 +02:00
jschoubben db35f2b0cb Merge origin/main into feat/r4e (R4e integration, before the final SDK pin) 2026-10-22 20:43:32 +02:00
jschoubben f90d3bfcc0 Merge feat/r4e-leaf into feat/r4e again: the gate's read states announced readable, pulls too (R4e integration) 2026-10-22 20:36:42 +02:00
jschoubben 37c47f8f95 Mark the delivery gate's pulls readable, as build-agent reads it (issue 1790) 2026-10-22 20:36:21 +02:00
jschoubben a45e467ddd Merge feat/r4e-leaf into feat/r4e again: the gate's read states announced readable (R4e integration) 2026-10-22 20:35:52 +02:00
jschoubben d3f9a1855d Mark the delivery gate's asks, merges, owed and source readable, as mesh-delivery's parts read them (issue 1790) 2026-10-22 20:35:29 +02:00
jschoubben c97cea21cf Merge feat/r4e-leaf into feat/r4e (R4e integration)
gitea's index.ts conflicted on imports: kept both, the leaf's statesAnnouncement and the gate's checkStatusTool.
mesh-delivery-gate is pinned to the sdk lane's newest head d1af87a (the leaf had a90f523) until the final SDK head.
2026-10-22 20:34:48 +02:00
mesh-tools 2f05e70734 Merge pull request 'Upgrade the nodes one at a time on the bus, with a package record, a kernel hold, an after-upgrade check, a rollback and a fallback kernel (upgrade safety, issues 1840 and 1841)' (#450) from feat/upgrade-safety into main 2026-10-22 18:34:15 +00:00
jschoubben 5ab9d6c296 Merge feat/r4e-buildmods into feat/r4e (R4e integration) 2026-10-22 20:30:01 +02:00
jschoubben e61c5ee5bf Merge feat/r4e-gate into feat/r4e (R4e integration) 2026-10-22 20:30:00 +02:00
jschoubben 0786408109 Say loudly, not fail, where a check has no bubblewrap to run a build module's shut-off steps (issue 1309)
build-agent's check container has no bwrap; the build nodes, where the checker runs from
R4e, do. The sibling test and the real Node build skip naming it, and the pull-head pin
test still runs its fetch half and skips only the shut-off build.
2026-10-22 19:04:24 +02:00
jschoubben a18eeeef5a Say gitea's states with the SDK's emit, as the toolchain's SDK has no keeps yet (issue 1790, issue 1852) 2026-10-22 19:04:21 +02:00
jschoubben 5b6dfc70f4 Wait until the older clipmenud holds the lock before the launcher starts (issue 1814, S1) 2026-10-22 19:03:03 +02:00
jschoubben 59c28971cb Narrow both build modules' fetch with mesh-sdk's buildstep.Narrow, the one copy (issue 1758, ADR 0499 §4)
The local narrowing of mesh-build-go and mesh-build-node moves into mesh-sdk (feat/r4e-sdk
d1af87a), with its tests; each module keeps its fetch-level test on real git. The three
build modules are pinned to d1af87a.
2026-10-22 18:42:17 +02:00
jschoubben 035c950ff2 mesh-build-node: prove with the real npm that a trunk package is published once and a candidate's never (issue 1587, S20, ADR 0498 §5)
A registry of the test's own speaks npm's publish protocol and asks for the credential: a
candidate's build publishes nothing, a trunk build publishes the packed file the build
digested, and a second trunk build of the same package finds it and publishes nothing.
2026-10-22 18:40:05 +02:00
jschoubben 5a79776625 mesh-build-node: name each bundle's launchers by mesh-sdk's one rule, and move the recipe for the SDK a member's lock names (R4e, lead decision L10)
writeLaunchers takes its entrypoint-to-launcher names from buildstep.Launchers, the rule the
controller reads a bundle by. RecipeSteps goes to 2: a member is now built against the SDK
its own lock names, which can change what a build produces (ADR 0405 §7).
2026-10-22 18:39:06 +02:00
jschoubben b9cdca78b0 mesh-build-node: narrow the fetch to the named paths, build members against the SDK their lock names, and build, pack and publish a package (issues 1758, 1587 S15-S19; ADR 0499 §4, ADR 0498 §5, ADR 0500 §4-§6; lead decision L10)
build_node_fetch takes buildstep.FetchRequest (dir and paths narrow the workspace) and an
optional package {repository, commit, ref}: the group's package member, built and packed
shut off and published nowhere. Each member's lock decides its SDK: the group's packed
file when the version and integrity are the lock's, else the registry's (integrity
checked); a version the registry lacks fails waits for: @novox/mesh-sdk@<version>. A
package artifact is built by its own script and packed with npm pack, shut off; only a
trunk build publishes it, from this process with the write credential (never in a
workspace): absent is published, the same file is done, another is version taken:. The
registry configuration file now lies outside every workspace.
2026-10-22 18:38:22 +02:00
jschoubben fde561075d mesh-build-go: take the fetch's request in mesh-sdk's one form and narrow the workspace to the module's named paths (issue 1758, ADR 0499 §4, lead decision L8)
build_go_fetch parses its arguments with buildstep.ParseFetchRequest; with dir and paths
(from the builder, never from a manifest) the tree is narrowed after the fetch and the
merge to those paths, the manifest and the lock files, so a build reads nothing else. The
narrowing waits to move into mesh-sdk as buildstep.Narrow.
2026-10-22 18:32:36 +02:00
jschoubben 3cce15fa81 restic: let the thinning wait up to ten minutes for a held repository lock (issue 620)
mesh/check failed: unit: modules/gitea.test: [gitea] where the poll of merges last looked from could not be read, so none is kept — no runtime | 3cce1…
mesh/merge-gate head 3cce15fa811e merged into trunk 835bf22b7d40 — pass: builds restic → ace, g14, novox, shanks; no bus step; every machine composes with …
mesh/repo-check head 3cce15fa811e merged into trunk 835bf22b7d40 — pass: its merge-check.sh passed, each part in its toolchain: merge-check.sh (go), merge-…
mesh/delivery delivered
A restic outside the night's lock, such as a listing while a delivery plan backs up
data modules before their moves, held the lock 246 ms before the thinning asked for it,
and the thinning failed at once, raising an urgent condition for a timing collision.
restic's --retry-lock now waits up to ten minutes on both of the thinning's forgets;
past that it fails with restic's words as before.
2026-10-22 18:30:16 +02:00
jschoubben 889fa09ea4 Prove a done build's artifacts resolve in a real artifact store, in both build modules (issue 1587, S9 S10, ADR 0498's check)
The integration tests start the distribution registry (the store's own program) in the
test with its in-memory storage and the real PushBlob: the built bundle and archive each
answer a reference ending in its digest, the key the builder would write is valid and
deliverable, and the store serves bytes of that digest at the reference. The registry is
imported by these tests alone (go list -deps of the programs does not name it).
2026-10-22 18:28:52 +02:00
jschoubben 3a25b6e1f8 mesh-build-node: push each done bundle to the artifact store from its own process and answer its reference (issue 1587, S5-S8, ADR 0498 §1 §2, lead decision L9)
After the build ended the module pushes the packed bundle with mesh-sdk's PushBlob under
<module>/<artifact> by its digest, a candidate's as a trunk build's, and the step answers
artifact-store://<module>/<artifact>/blobs/<digest> instead of the archive's path. A push
that fails fails the step not stored: <artifact>: <the store's words>; the packed bytes are
never written to the node. The manifest requires artifact-store and places its address.
2026-10-22 18:25:14 +02:00
jschoubben 22992da15d mesh-build-node: emit a package that builds before its tests, so tests importing its own output find it (issue 1869, part of issue 1309's run step)
A package with a build or prepare script and a tsconfig.json is built with tsc -p . in the
shut-off run step, at both levels, instead of the --noEmit type-check; the fetch's
--ignore-scripts never runs those scripts. mesh-sdk's tests import ../dist and failed
ERR_MODULE_NOT_FOUND. The result names the step build.
2026-10-22 18:23:25 +02:00
jschoubben b6247f837d mesh-build-go: push each done archive to the artifact store from its own process and answer its reference (issue 1587, S1-S4, ADR 0498 §1 §2, lead decision L9)
After the build ended the module pushes the packed bundle or archive with mesh-sdk's
PushBlob under <module>/<artifact> by its digest, a candidate's as a trunk build's, and the
step answers artifact-store://<module>/<artifact>/blobs/<digest> instead of a path. A push
that fails fails the step not stored: <artifact>: <the store's words>. The packed bytes are
never written to the node. The manifest requires artifact-store and places its address.
2026-10-22 18:22:17 +02:00
jschoubben 9c5b7e14e1 mesh-build-go: fold the siblings' commits into every digest and echo them, and test and digest the integration level under integration,isolation (lead decision L5, the checker's agreement; issue 1309)
The fetch writes beside.json naming the siblings it placed; the digest adds a line per
sibling to every package's text and answers them as beside, so a core repository's verdict
is kept only for the siblings it was judged with. DigestVersion 3, once for both changes:
the integration level's tags are integration,isolation, so the bubblewrap proofs run in
every check on a build node.
2026-10-22 18:18:03 +02:00
jschoubben c0873728e2 Replace the deleted manifest reads check by each build module's state announcement (ADR 0495 §4-§5, issue 1790) 2026-10-22 18:17:04 +02:00
jschoubben 86ba54b06e Take the SDK at the milestone's mesh-sdk head 8892a117 in pacman, which holds main's SDK and the milestone's (issue 1840)
mesh/repo-check head 86ba54b06e09 merged into trunk b5f812270eae beside mesh-controller#439@abe356a2 — pass: its merge-check.sh passed, each part in its to…
mesh/check failed: unit: i3/cmd/i3-tools.TestTheMainFileAndEveryModulesContributionLoadTogether: manifest_test.go:247: expected the launcher, clipboar…
mesh/merge-gate head 86ba54b06e09 merged into trunk b5f812270eae — pass: group feat/upgrade-safety composed: every machine composes with the change as it d…
mesh/delivery covered after its failure: a later walk that contains it was delivered: plan-1792694058124013462 (novox/mesh-catalog at 2f05e707 into main,…
mesh/delivery-group group feat/upgrade-safety delivered: every member is delivered
2026-10-22 18:14:25 +02:00
jschoubben af91a7e660 Merge remote-tracking branch 'origin/main' into feat/upgrade-safety 2026-10-22 18:14:16 +02:00
jschoubben d45767bb2f Keep another module's path out of i3's fixture (issue 1814, S2) 2026-10-22 18:13:58 +02:00
jschoubben d108487396 mesh-build-go: read a pinned release compressed with xz as the build agent does, so betterbird's archive gets a done key (issue 1590)
The xz reader is github.com/therootcompany/xz, declared in go.mod as build-agent declares it,
with build-agent's framing walk copied (xz_framing.go and its tests): too many blocks or
streams are refused before anything is decoded, a dictionary over 64 MiB is refused. The
same release tar under gzip and under xz packs to the build agent's archive.
2026-10-22 18:13:49 +02:00
jschoubben 027ea030fa Qualify a short base image name before buildah reads it (issue 1824) 2026-10-22 18:13:09 +02:00
jschoubben 6eb9eb53de Commit the lock files of mssql, mongodb, model-usage and mesh-catalog, and refuse a Node module without one (issue 1868, ADR 0469 §5) 2026-10-22 18:12:32 +02:00
jschoubben eed6be0f4a mesh-build-go: run go test, go build and go list in bubblewrap, so a change's code reaches no key, login or network (issue 1309, S1-S6, ADR 0448)
Each run step goes through mesh-sdk's buildstep.Isolated: no network, none of the account,
the workspace read-write, the Go caches beneath a throwaway layer unless the build step is
told publish (a trunk build). The integration level runs its store, its bus and its tests in
one shut-off process (this program again, level <spec>), so they reach each other and
nothing else. build_go_test takes a -run pattern (a try's tests). The module declares the
bubblewrap package and refuses a step, unhealthy, while bwrap is missing or older than 0.10.
The proof test runs a probe test through the real unit level.
2026-10-22 18:12:01 +02:00
jschoubben 51900367a0 mesh-build-go: announce the states it keeps from its code, so readers are granted from its word (issue 1790, ADR 0495 §2 §5)
Kept (toolchains, keys, steps, each read by another module today) is said with stdio.Keeps
before the module serves; announcetest.Check holds the announcement equal to the code. Pinned
to mesh-sdk feat/r4e-sdk 9cb4d59. The manifest test's call of the deleted
StateReadsNothingDeclares still has to go (the edit is waiting on the operator's permission).
2026-10-22 18:12:01 +02:00
jschoubben 77596f37a3 mesh-build-image: announce the states it keeps from its code, so readers are granted from its word (issue 1790, ADR 0495 §2 §5)
Kept (keys, steps, each read by another module today) is said with stdio.Keeps before the
module serves; announcetest.Check holds the announcement equal to the code. Pinned to
mesh-sdk feat/r4e-sdk 9cb4d59. The manifest test's call of the deleted
StateReadsNothingDeclares still has to go (the edit is waiting on the operator's permission).
2026-10-22 18:12:01 +02:00
jschoubben 56f842c3af mesh-build-node: run the type-check, the test script, tsc and esbuild in bubblewrap, so a change's code reaches no key, login or network (issue 1309, S1 S7 S8 S9, ADR 0448)
Each run step goes through mesh-sdk's buildstep.Isolated: no network, none of the account,
the workspace read-write, the toolchain read-only, the npm cache beneath a throwaway layer
unless the build step is told publish (a trunk build). The module declares the bubblewrap
package and refuses a step, unhealthy, while bwrap is missing or older than 0.10. The proof
test runs a probe script through the real test step.
2026-10-22 18:08:12 +02:00
jschoubben 075b61901c mesh-build-node: announce the states it keeps from its code, so readers are granted from its word (issue 1790, ADR 0495 §2 §5)
Kept (toolchains, keys, steps, each read by another module today) is said with stdio.Keeps
before the module serves; announcetest.Check holds the announcement equal to the code. The
module is pinned to mesh-sdk feat/r4e-sdk 9cb4d59.
2026-10-22 18:08:06 +02:00
jschoubben 0270027bf9 Take the SDK at the milestone's mesh-sdk head f8b8f3a6 in pacman, which the delivery group's members must pin exactly (issue 1840) 2026-10-22 18:07:40 +02:00
jschoubben e85b7da63b Let i3's contribution test read its fixture, not the siblings' manifests (issue 1814, S2, ADR 0495 §1) 2026-10-22 18:05:04 +02:00
jschoubben c112df0c1c Announce the states the forge module keeps (issue 1790, ADR 0495 §2 §5) 2026-10-22 18:03:06 +02:00
jschoubben 8e791896f1 Announce the states the delivery gate keeps (issue 1790, ADR 0495 §2 §5) 2026-10-22 18:02:11 +02:00
jschoubben b07a101caf Answer each named path's git object id from gitea_commit_trees (issue 1758, S1-S2, ADR 0499 §3) 2026-10-22 17:59:23 +02:00
jschoubben 72effa1fb1 Copy an upstream artifact into the artifact store by its pinned digest (issue 1587, S11-S13, ADR 0498 §4) 2026-10-22 17:57:05 +02:00