Commit Graph
3043 Commits
Author SHA1 Message Date
jschoubben 5c98c86d70 Answer A-299: the view's answers take one key form per asks kind; the kinds are seven
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 20:00:56 +01:00
jschoubben b75c04f038 Answer A-298: one shared follower of the seat-holders view in the catalog seat's client
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 20:00:56 +01:00
mesh-tools 023a355cdb Merge pull request 'Design 47 G: §717 (the views are named by their seats) and the lead's answers A-294 to A-297' (#588) from design/47-G-lead-answers-16 into main 2026-11-12 17:59:25 +00:00
jschoubben b03f5193b8 Design 47 G §717: a follower hears its seat's holder from the seat-holders view
mesh/delivery merged: source.main moved to novox/hq@023a355cdb2c
mesh/check being carried to b03f5193b8c4 on 023a355cdb2c | from 7247b92fd6b1
A seat's view is said on its holder's module subject, so a follower must know
the holder; asking for it every 5 seconds was polling and kept a holder that
moved. It now follows the seat-holders view as A-295's holder function does,
reached from the catalog half's claim, and re-follows when the holder changes.

Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 18:54:41 +01:00
jschoubben 80863372dc Design 47 G §717: the status view's and the assigner's views are named by their seats
Developer J27 met, building A-210, a status view that claims no seat and an
assigner holding two seats, both followed by the teller by module name. ADR
0540 §1 already names the seat mesh-delivery-view; G defines it with its one
view, puts the assigner's asks and wants on mesh-wants, and has a follower find
the holder in the background so no start stops for it.

Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 18:54:41 +01:00
jschoubben 01fdde3701 Answer A-297: the runner's step ask id takes the usual form, repeatable per placement
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 18:54:41 +01:00
jschoubben ef76908e0d Answer A-296: seat forms in the manifests, own buckets from the credential's module
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 18:54:41 +01:00
jschoubben eed5be9244 Answer A-295: a callers list names a seat, resolved from the seat-holders view the SDK follows
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 18:54:41 +01:00
jschoubben 12c9c1efcf Answer A-294: mesh-wants' step keeps no token; step-progress names the caller its definition lists
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 18:54:41 +01:00
mesh-tools 7247b92fd6 Merge pull request 'Design 50: the own resolver's file waits longer than its pass over a silent mesh resolver (ADR 0678, issue 1748)' (#587) from design/50-resolver-wait into main 2026-11-12 17:51:40 +00:00
jschoubben ce0287dd2d Keep design 50 to what ADR 0678 asks, and hold the wait at 3 s per try
mesh/delivery merged: source.main moved to novox/hq@7247b92fd6b1
mesh/check being carried to ce0287dd2d87 on 7247b92fd6b1 | from 0d04a16560e4
A restart of resolved and a timed ask at the own resolver answered in 44-46 ms
with the first mesh resolver silent, so no read catches the pass today: the
value stays ADR 0678's until a read of resolved's server switches exists, and
the proof is the Go build steps at the next silence. The phases table and the
as-built notes go back to main: a to-be design states no live state.
2026-11-12 18:06:41 +01:00
jschoubben 067de23181 Give the own resolver's file a wait longer than its pass over a silent mesh resolver
ADR 0678: on a machine holding node-resolver the file names the own resolver
alone, so a program gave up after about 2 s while resolved still waited on a
silent first mesh resolver (issue 1748). Design 50 section 2 now gives the
file timeout:3 attempts:2 and how the value follows the measured pass; the
new milestone part resolver-wait holds the lines that build it, putting the
design in the shape of a main file and milestone parts.
2026-11-12 18:05:05 +01:00
mesh-tools 0d04a16560 Merge pull request 'Design 47 G: the lead's answer A-293' (#586) from design/47-G-lead-answers-15 into main 2026-11-12 16:24:57 +00:00
jschoubben 7c533965bf Amend A-293: a later done end deletes the refused key a failed end wrote
mesh/delivery merged: source.main moved to novox/hq@0d04a16560e4
mesh/check checking 7c533965bfa6 on 0d04a16560e4
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 17:11:35 +01:00
jschoubben f99c0c9e29 Amend A-293: the newest build is the last trunk build-ended event heard for the module
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 17:09:10 +01:00
jschoubben 66d37b8f30 Answer A-293: in by-entries mode a failed newest build is said as a refused key on the builder's build-ended event
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 17:05:37 +01:00
mesh-tools 852a0d1163 Merge pull request 'Design 47 G: the lead's answers A-291 and A-292' (#585) from design/47-G-lead-answers-14 into main 2026-11-12 16:00:57 +00:00
jschoubben cbbf170718 Answer A-292: a node-engine's failed say is said again within the link, loudly
mesh/delivery merged: source.main moved to novox/hq@852a0d1163d3
mesh/check checking cbbf17071862 on 852a0d1163d3
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 16:53:42 +01:00
jschoubben 3c357ee1f1 Answer A-291: a process receives a chunked seat answer through one SDK helper
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 16:46:26 +01:00
mesh-tools 43612bf1be Merge pull request 'Design 47 G: the lead's answers A-287 and A-288' (#582) from design/47-G-lead-answers-13 into main 2026-11-12 15:34:11 +00:00
mesh-tools 5f35cd2e32 Merge pull request 'Design 47: the lead's answers for a placed module whose build failed (a small fix on main, issue 1966)' (#584) from design/47-unpublished-placed-module into main 2026-11-12 15:28:10 +00:00
jschoubben 048545b266 Answer A-290: a step answer over the steps state's bound is kept as an object the key names
mesh/delivery merged: source.main moved to novox/hq@43612bf1be0a
mesh/check checking 048545b266bb on 43612bf1be0a
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 16:27:32 +01:00
mesh-tools 7792e96670 Merge pull request 'Design 39 parts: the lead's answers D3 (a list of a place a secret is kept, for §15) and D4 (the database §13 and §14 read)' (#583) from design/39-list-secret-place into main 2026-11-12 15:27:07 +00:00
jschoubben 34660dd256 Correct A-289: the unpublished count holds no package, since no reachable case leaves a reserving build's package unpublished
mesh/delivery closed: a newer head of the same pull request
mesh/check being carried to 34660dd2565f on 7792e9667042 | from 63ccbe512bc0
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 16:17:06 +01:00
jschoubben 84faadbe38 Answer A-289: a build that reserves no version leaves its package done; a package is published by its own language's seat
mesh/delivery closed: a newer head of the same pull request
mesh/check checking 84faadbe382e on 63ccbe512bc0
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 16:12:45 +01:00
jschoubben 11d257516a Design 47 fix: U8, the refused key is written on the failed build's event
mesh/delivery merged: source.main moved to novox/hq@5f35cd2e324a
mesh/check checking 11d257516a51 on 5f35cd2e324a
2026-11-12 15:56:00 +01:00
jschoubben 99d01964ef Design 47: the lead's answers for a placed module whose build failed, a small fix on main 2026-11-12 15:49:05 +01:00
jschoubben da6c3ab163 Answer A-287 and A-288: a kept announcement names every declared state; a refusal's test holds its meaning
mesh/delivery closed: a newer head of the same pull request
mesh/check pass da6c3ab16305 on 63ccbe512bc0 | unit 43 run, 0 kept, 1 skipped | integration 43 run, 0 kept, 1 skipped | composition passed in 52s on 4…
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 15:47:09 +01:00
jschoubben 1995bd1247 Design 39 parts: the lead's answer D4, the database §13 and §14 read is the one postgres made
mesh/delivery merged: source.main moved to novox/hq@7792e9667042
mesh/check being carried to 1995bd1247be on 7792e9667042 | from 63ccbe512bc0
2026-11-12 15:45:59 +01:00
mesh-tools 63ccbe512b Merge pull request 'Design 39 bus-and-switch: the lead's answer C5, how a carried switch is read' (#581) from design/39-carried-switches into main 2026-11-12 14:42:06 +00:00
jschoubben 1a6a72dc02 Design 39 parts: the lead's answer D3, a list of a place a secret is kept, for §15's check 2026-11-12 15:41:47 +01:00
jschoubben 45ee24759b C5 cites the principles that settle it: failure is loud, the simplest thing
mesh/delivery merged: source.main moved to novox/hq@63ccbe512bc0
mesh/check checking 45ee24759b07 on 63ccbe512bc0
2026-11-12 15:37:49 +01:00
jschoubben 8326a3b57d Design 39 bus-and-switch: the lead's answer C5, how a carried switch is read 2026-11-12 15:37:25 +01:00
mesh-tools e81a8d62a5 Merge pull request 'Design 47 G: the lead's answers A-284 to A-286' (#580) from design/47-G-lead-answers-12 into main 2026-11-12 14:17:58 +00:00
jschoubben 41dd30fe43 Amend A-286: a check step that errors after its fetch merged still answers its fetch
mesh/delivery merged: source.main moved to novox/hq@e81a8d62a591
mesh/check checking 41dd30fe436e on e81a8d62a591
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 15:16:53 +01:00
jschoubben c7850092fa Correct A-285: the group's packages are group-packages, since packages already holds the Go versions
mesh/delivery closed: a newer head of the same pull request
mesh/check checking c7850092fa08 on e1a9a0f2d238
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 15:16:30 +01:00
jschoubben 9dfde6fca1 Answer A-285 and A-286: a fetch carries the group's packages as a list; a failed test never hides the validator
mesh/delivery closed: a newer head of the same pull request
mesh/check checking 9dfde6fca119 on e1a9a0f2d238
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 15:15:44 +01:00
jschoubben a8da15caff Answer A-284: a group's check builds a TypeScript member against the group's own packages
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 15:08:26 +01:00
mesh-tools e1a9a0f2d2 Merge pull request 'Design 47 G: the lead's answers A-281 to A-283' (#579) from design/47-G-lead-answers-11 into main 2026-11-12 14:04:49 +00:00
jschoubben de676868c6 Correct A-282: the step bound moves with the step by a revision-checked update, never a delete
mesh/delivery merged: source.main moved to novox/hq@e1a9a0f2d238
mesh/check being carried to de676868c65a on e1a9a0f2d238 | from a412c0760ba4
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 14:36:17 +01:00
jschoubben da9b672747 Answer A-282 and A-283: a step bound is created once; a template state's max-mib reaches its bucket
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 14:36:17 +01:00
jschoubben 14f908926e Answer A-281: the switch runs once and marks it, so §716's deliveries never become records at a start
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 14:36:17 +01:00
mesh-tools a412c0760b Merge pull request 'Design 25 bus-state: the lead's answer L1, a merged candidate's manifest is newly accepted' (#578) from design/25-merged-verdict into main 2026-11-12 12:28:30 +00:00
jschoubben ff30ac166e Design 25 bus-state: the lead's answer L1, a merged candidate's manifest is newly accepted
mesh/delivery merged: source.main moved to novox/hq@a412c0760ba4
mesh/check being carried to ff30ac166ecc on a412c0760ba4 | from a45b31fdaf2e
2026-11-12 13:20:17 +01:00
mesh-tools a45b31fdaf Merge pull request 'Design 39 parts: the licence manager is built of its parts, and the finished carry-over and its database leave' (#576) from design/39-parts into main 2026-11-12 12:08:26 +00:00
mesh-tools 5235c00630 Merge pull request 'Design 47 G: §716 (a group's members and bases before the take-in) and the lead's answers A-274 to A-280' (#577) from design/47-G-lead-answers-10 into main 2026-11-12 12:05:39 +00:00
jschoubben 6141d7b5d1 Design 39 parts: the lead's answers D1 (the vault reading) and D2 (§17's wording)
mesh/delivery merged: source.main moved to novox/hq@a45b31fdaf2e
mesh/check checking 6141d7b5d177 on a45b31fdaf2e
2026-11-12 12:21:10 +01:00
jschoubben 2132bb711d Merge §716 (group members before take-in) into the lead answers; A-279 in its form after §716
mesh/delivery merged: source.main moved to novox/hq@5235c0063024
mesh/check being carried to 2132bb711d4e on 5235c0063024 | from f9bdd3b1d1ce
Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 11:55:00 +01:00
jschoubben f74f9b19be Design 47 G §716: the gate keeps a group's members and their bases from the announcement
Before any take-in the gate held no group: its records open at the take-in, and from G it writes
`merges` no more, so the forge bridge's §715 read of each member's base through `groups` had
nothing to answer from. ADR 0666 §13 has the gate form the group; §716 keeps it in the states
`deliveries` and `groups` the gate already holds (§613), heard on the seat git's pull-updated,
pull-merged and pull-closed, and says who deletes each key and how `groups` answers.
A-236 now points to it, and A-279 is carried over from the answers branch in the form §716 makes
it, its Approve-entry form dropped.

Claude-Session: https://claude.ai/code/session_01NcbgfGKaxw984XyAPfLm4h
2026-11-12 11:54:16 +01:00
jschoubben ac294e7b86 Design 39 parts: build the licence manager as four parts of one manifest and let the finished carry-over leave
mesh/delivery closed: a newer head of the same pull request
mesh/check pass ac294e7b8628 on f9bdd3b1d1ce | unit 43 run, 0 kept, 1 skipped | integration 43 run, 0 kept, 1 skipped | composition passed in 91s on 4…
The operator asked for the manager built as clean as possible, one manifest of several parts. The seat's verbs stay
on the runtime of its own, since a process account answers none; the tools that only read get a bundle of their own
so they answer while the manager fails; the carry-over ran live, so its code and the store database's and grant-key's
requirements leave now and retire on the operator's Approve. The glossary names the two senses of part.
2026-11-12 11:49:19 +01:00