Commit Graph
2544 Commits
Author SHA1 Message Date
jschoubben 6ea33475b6 B1 (H-A-412): a step's oversized end is read as its answer object
The delivery SDK's build-step part keeps a step's end over the steps state's
value as the object answer.<ask>, read through the build seat's verb answer;
OutputObject is gone. The builder's wait names the answer object it could
not read.

Builds: design 47 H2 §775, §791; H-A-412
2026-11-22 01:12:25 +01:00
jschoubben 905a055a1e B2: the builder serves the seat mesh-builds' drop and its queue's tools
A caller whose step ended by its bound could not stop the builder working on
its ask, and nobody could see what waited on the seat's work queue. The
builder now serves the seat's verb drop: an ask it runs is cancelled, one not
pulled yet is kept dropped in its state taken (dropped.<ask>, for
ask-taken-bound); both end the step could-not-run naming the drop and who
asked it. builder_asks reads the waiting asks through the runtime and the
taken ones and the drops from the state taken; builder_ask_drop asks the
seat's drop once the hand act is recorded.

Builds: design 47 H2 §777, §778; H-A-401, H-A-404, H-A-410
2026-11-22 01:08:30 +01:00
jschoubben 3a647a8e2d B3: an answered build ask is one journal line naming its ask
The line a build step's finish wrote named no ask, so an answer could not be
matched to the ask the runner put. BuildStep now takes the ask's id and its
finish writes "answered <ask> <pipeline> <movers> <step> <outcome>".

Builds: design 47 H2 §792
2026-11-22 00:52:34 +01:00
jschoubben 7dbd440a4e B1: the builder pulls the seat mesh-builds' work queue and takes each ask once
An ask put while the builder is down waited for nobody: the plain subscription
heard only what came while it ran. The builder now pulls every waiting ask from
the seat's one worker, takes it once by its id in its state taken (for
ask-taken-bound), says ask-taken at the pull, sends in-progress while it builds
and acknowledges once the step's end is said. Its publish to a registry seat is
a kept call carrying the publish's ask id. The manifest declares taken and the
kept-call pace and ask-taken-bound as runtime settings.

Builds: design 47 H2 §775, §777, §820; H-A-381
2026-11-22 00:45:40 +01:00
jschoubben 9acf583daf R3: a step's end is one journal line, said in one place; Resume puts no taken ask again
Every step-ended goes through sayStepEnded, which writes "step ended <pipeline>
<mover> <step> <outcome> in <duration>"; the cancel verb now says each step it
ends. Resume at a start puts again only an open step's ask no holder took.

Builds: design 47 H2 §784, §777 (H-A-397)
2026-11-22 00:35:13 +01:00
jschoubben f420a8213a R2: an ask's bound counts from the pull and ends its step, dropped
The bound key ask.<seat>.<ask> is armed at the put, again at the holder's
ask-taken (heard now for mesh-builds and mesh-wants too) with what is left
since taken-at, again from now at each step-progress, and at a start for every
open ask whose key is gone; it is deleted at the step's end. Past it the step
ends could-not-run, naming an untaken ask's seat holder as taking none, and its
ask is dropped through the seat's drop. The silent seats' bound, the second put
under AgainID, the ask2. keys and the put again at a holder's running go; a put
refused by a full work queue is said and put again at the kept-call pace.

Builds: design 47 H2 §777, §778 (H-A-349 DQ-4, H-A-395)
2026-11-22 00:30:56 +01:00
jschoubben e7d9849efa R4: pipeline-ended carries the pipeline's phases
taken (the record's start at its take-in), placed (the earliest step start),
ended, and per step kind the earliest start and latest end across movers, so
the gate reads where a pipeline's time went.

Builds: design 47 H2 §785
2026-11-22 00:22:50 +01:00
jschoubben f6e7f4210b B9: a fetch or a publish met away ends its run again, by again-delays, with no seat call
The build modules' shared steprun parts of §799: again-delays (15s,60s) read
into a Puller's Delay; a fetch's 401/5xx or connection fault and a publish's
no responders or exited bundle are transport faults whose why opens fetch: or
publish:; the assigner call that named a silent holder goes. The builder reads
an exited bundle as no responders and keeps the part in a key's fault words.

Builds: design 47 H2 §799 (H-A-390)
2026-11-22 00:20:17 +01:00
jschoubben 1058b0eec6 R1: the runner puts a step's ask and asks published as kept calls, paced, within ask-unanswered-bound
A put that met a transport fault is put again under the ask's own id at the
kept-call pace instead of failing the step at the first fault.

Builds: design 47 H2 §775
2026-11-22 00:15:07 +01:00
jschoubben 125d6382ea D6: the delivery seat's deliveries filters before it sizes
deliveries filters by repository, module, outcome and since while it reads,
pages by the cursor next and answers its deliveries with capped and total.
The seat's form rises to 2.

Builds: design 47 H2 §779, §811
2026-11-22 00:09:51 +01:00
jschoubben d2d3ebbf44 D5: the git seat's tags name their commits, and a holder says away
tags answers items {tag, commit}; trees' description says a missing directory
answers not-found; the error kind away joins refused, not-found and failed,
every answer carrying part and words with it; tag's description names the
<directory>/v<version> form. The seat's form rises to 2. The builder's,
bridge's and indexer's readers take the tag's name from the item.

Builds: design 47 H2 §776, §796; H-A-350 DMC-2, DMC-3, H-A-354
2026-11-22 00:08:50 +01:00
jschoubben 278310cc67 D4: the catalog seat pages registrations, says the seat list, takes a source alone
registrations filters by repository beside module while it reads, pages by
the cursor next and answers capped and total. The read verb seat-list answers
the seat list's object in seat-lists by digest, size and name. A registration
takes a source-only form under source.<owner>.<name>, with origin, one branch,
policy and check (a build seat and a step kind), naming no module; its history
keeps the same key. The seat's form rises to 2.

Builds: design 47 H2 §806, §811, §814, §815; H-A-349 DQ-9, DQ-10
2026-11-22 00:05:48 +01:00
jschoubben 1767d7b4e9 D3b: an install ends in a step-ended naming the step install
The build-step part's step enum and the delivery SDK's PipelineSteps gain
install; an install's end copies the module and the commit its ask names and
carries the archive, and the SDK's check takes it so.

Builds: design 47 H2 §800; H-A-367 (4), H-A-379
2026-11-22 00:04:05 +01:00
jschoubben ac99946fcb D4a: the registry's list takes module and a cursor, and pages at 768 KB
oci-artifact-registry's list filters by module beside name while it reads,
continues from the opaque cursor next and answers capped and total; the
seat's form rises to 2. The shared remove no longer judges its caller for a
retention removal.

Builds: design 47 H2 §779, §811, §813 (7); H-A-350 DMC-10
2026-11-22 00:02:21 +01:00
jschoubben 4749c686eb D3: the build-step part says output, current, not-run and install in its form 2
output takes the ask and answers the output's last 64 KB, its size, the
failing tests with their words, or no longer kept; its call and chunks go.
current answers the node, whether it takes asks and why not since when, its
capacity and every ask it runs with its asker. A step's end gains not-run and
archive, a step ask names the archive to unpack, and node-build-node accepts
install. The delivery SDK's neutral forms carry the same fields.

Builds: design 47 H2 §791, §800, §801; H-A-350 DMC-16, H-A-367
2026-11-22 00:00:41 +01:00
jschoubben 1f5b0f813d Merge origin/main into feat/47-H2 2026-11-21 23:53:44 +01:00
jschoubben 99b25a7db4 H-A-361: a hand-off places again the stayers not asked on a mover that left
On 21 Nov the order was the reverse of the first commit's: mesh-contracts'
old build was withdrawn first (21:50:16), so the stayers that awaited it
were placed not asked, and only then did the hand-offs take it away. A
mover that left is still a dependency of those stayers: their not-asked
steps held the pipeline's end as failures. At the hand-off the steps of
each stayer that ended not asked on a dependency that left, directly or
through a mover itself taken out, are taken out, so the placing that
follows places them again against the newest published package; a
dependency that truly did not build is named again at that placing. The
journal says which movers are placed again.

Builds: design 47 H H-A-361, H-A-53, H-A-223
2026-11-21 23:34:24 +01:00
jschoubben a559b0e7c9 D1b: the two seats' work queues carry their settings, derived once in mesh-sdk
mesh-builds' and mesh-wants' accepted kind step names max-age 24h, max-msgs
10,000, ack-wait 60s, in-progress 20s, max-ack-pending 1,000, max-deliveries 5
and a duplicates window of 2m; the grants part makes each seat's ask stream
through mesh-sdk's buscompose.SeatQueue and says a seat whose settings it
refuses, and its own copy of the derivation goes.

Builds: design 47 H2 §777; H-A-350 DMC-1, H-A-362
2026-11-21 23:23:04 +01:00
jschoubben 7606d515fd H-A-361: a hand-off places the movers that stay; a mover that left is not awaited
main.158 and main.163 ended failed seconds after newer pipelines took
movers away: the movers that stay still awaited the movers that left,
whose builds had ended superseded, so the next placing ended them "not
asked: its dependency did not build" and the end judgement counted
them as failures. A mover that left is no longer awaited (the stayer
builds against the newest published package, H-A-53), is never placed
in the older pipeline, and the hand-off places what stays at once and
says in the journal which movers stay.

Builds: design 47 H H-A-361, H-A-53, H-A-223
2026-11-21 23:17:59 +01:00
jschoubben b40745534f D2: the runner's seat says phases, unfinished and re-pinned, and pages show
A step keeps asked-at and taken-at; pipeline-ended carries phases and
unfinished; show takes step and state filters and the cursor next, and
answers capped and total; re-pinned is declared. The seat's form rises to 2.

Builds: design 47 H2 §778, §785, §786, §789, §821; H-A-349 DQ-13
2026-11-21 23:17:31 +01:00
jschoubben c1ec75789f Readers of mesh-contracts' regenerated seats take the form
The node enroller's and the watchdogs' generated decoders answer the form
they read (mesh-contracts K1); their readers here take it.

Builds: design 47 H2 §780
2026-11-21 23:15:31 +01:00
jschoubben 75208ffcba D1: mesh-builds and mesh-wants say ask-taken, take drop, and the want carries run
Both seats declare ask-taken (ask, node, at) on ask-taken.<ask> and gain the
act drop; mesh-builds declares version-reserved; the want value the
node-engine applies carries run, the composition run's pipeline sequence, and
mesh-wants' form rises to 2. The queue settings of the accepted kind wait on
mesh-sdk's seatdef and buscompose (H-A-362, D1b).

Builds: design 47 H2 §777, §778, §805, §813 (1), §821; H-A-349 DQ-1, H-A-350 DMC-15, H-A-355 (1)
2026-11-21 23:15:31 +01:00
jschoubben 6046ebe2e0 G1: the fields digest is kept per type
A type added or removed whole changes no reader's value; only a type whose
fields changed while the form did not fails the compare.

Builds: design 47 H2 §780; H-A-352 (2)
2026-11-21 23:15:31 +01:00
jschoubben ff7db33720 G1, D0: seat readers decode tolerantly and say their form
The seat generator gives every value, event and answer the wire field form
(left out while 1) and a tolerant decoder answering the form it read; each
contract carries its forms and a digest of its fields, and the compare fails
a definition whose fields changed while its form did not. Every delivery seat
is regenerated, and the readers take the form.

Builds: design 47 H2 §780; H-A-352
2026-11-21 23:11:10 +01:00
jschoubben 2df32b7892 T1: the teller's ask-again for want of room goes
The channel holds no room to wait for (A-424): a refusal for want of room is
said as any refusal is, and its waiting, its bound of refusals and their tests
go. The wait for asks the router did not take stays.

Builds: design 47 H2 §813 (10)
2026-11-21 23:03:08 +01:00
mesh-tools 7e597a623e Merge pull request 'The runner and the builder write a journal line per step end, ask placed and ask answered (H-A-320, H-A-333)' (#320) from fix/47-H2-journal-lines into main 2026-11-21 21:01:55 +00:00
jschoubben 6a847a8d9c H-A-333: the ended line at every said end, the answered line at every answer
The runner's ended line was written only for a heard step-ended and its run's
readers; the unanswered-bound end, the end at placing, a supersede, a re-pin,
the approval's end, the start pass and the verbs said their ends silently. It
is now written in Hearer.say, the one place that says step-ended, and at the
starter's two sayers that call stepEvent directly (repin, sayNotAsked). The
builder's answered line sat on four of buildEntry's returns; the ask's answer
is now one line in BuildStep's finish, naming the pipeline, the movers, the
step and the outcome, the per-entry existing/built-now line kept beside it.

Builds: design 47 H H-A-320 (1), H-A-333
2026-11-21 21:58:36 +01:00
jschoubben dc83ef3e27 H-A-320 (1): a journal line per step end, step ask and build answer
A working 36-mover pipeline was read as stalled twice: the runner wrote no
line when a step ended or was asked, and the builder none when it answered a
step ask. The runner now writes "<pipeline> <mover> <step> ended <outcome>
(<node>, <duration>)" at each step end it writes (the asking step, every
reader of its test run, and each reader EndRun ends), and "<pipeline> <mover>
<step> asked on <seat>" for each ask put (the re-ask paths keep their own
line). The builder writes "<module>@<fingerprint> build step answered
<outcome> for <pipeline> (existing build / built now)".

Builds: design 47 H H-A-320 (1)
2026-11-21 21:50:48 +01:00
mesh-tools 600c13b9e9 Merge pull request 'Design 47 H: the runner redoes a cut placing, places on a late end, names a stalled pipeline, and reads a prior publish from the build seat (H-A-311, H-A-313, H-A-326, H-A-328, H-A-331)' (#319) from fix/47-H1-runner-placing-redone into main 2026-11-21 20:48:59 +00:00
mesh-tools 5396dc9b11 Merge pull request 'The teller's want-of-room test bed answers the seat-holder ask; one taker of registration-changed (H-A-295)' (#318) from fix/47-H1-teller-test-want-of-room into main 2026-11-21 20:48:56 +00:00
mesh-tools 425dbaef99 Merge pull request 'Design 47 H: a passed pipeline whose movers' versions arrive late still writes its wants (H-A-270, H-A-273, H-A-277)' (#317) from fix/47-H1-end-taken-when-versioned into main 2026-11-21 20:48:49 +00:00
jschoubben 714c2c1ccc H-A-331: the published answer carries the publish time and digest where kept; a withdrawn number is never published
The verb published answered no time, and after a release no digest,
while the seat form says both come when published. The builder now
keeps when it heard or confirmed each publish on the mark, and each
published digest on a kept release; marks and releases written before
answer without them. A number a take-over withdrew is not published.
The runner drops a pipeline's kept answers at its end.

Builds: design 47 H H-A-331, H-A-328
2026-11-21 21:38:50 +01:00
jschoubben e583dce741 H-A-326 (1), (3): a mover's wait on a package is a read first, an event second
main.137 stood twenty minutes with nats and mesh-vault waiting on
mesh-grants packages published at 20:00, before the wait was written,
so no publish event came. The runner now asks the build seat's
published when it writes the wait and places the mover when the
package is published, saying so in the journal apart from an event.

Builds: design 47 H H-A-326 (1), H-A-326 (3), H-A-328
2026-11-21 21:25:03 +01:00
jschoubben 5e76fdf725 H-A-328: the seat mesh-builds answers published; the builder serves it
A package published before a pipeline existed is in no record of the
runner, so the builder, which owns publishes, answers whether an
artifact built at a fingerprint is published, from its reservations
(marked from the registries' record) and its releases.

Builds: design 47 H H-A-328, H-A-326 (1), H-A-274
2026-11-21 21:25:03 +01:00
jschoubben 438c8740a6 H-A-326 (2): a pipeline waiting only on packages is named stalled
A step held on a package's publish counted as open, so a pipeline whose
only movers waited on publishes that had come before their wait was
never named by runner_health (main.137, twenty minutes).

Builds: design 47 H H-A-326 (2), H-A-311 (3)
2026-11-21 21:18:32 +01:00
jschoubben 05d9f90bdb H-A-311 (3) review: the health reads records keeping revisions; a required step waiting is open
runner_health, show, list and stands refused every pipeline record
carrying the runner's own kept-revisions ("unknown field
kept-revisions" on main.131 and main.137): the field is now dropped
before the seat's strict decode. A required step waiting (an approval,
a pin-held step) now counts as open, as the end's judgement holds it.

Builds: design 47 H H-A-311 (3), H-A-313, H-A-285, H-A-309, H-A-322
2026-11-21 20:53:04 +01:00
jschoubben f22945c26a H-A-311 (3): runner_health names a running pipeline with nothing open and nothing placed for five minutes
main.137 stood seventeen minutes with nothing open while runner_health
said healthy. The tools program says each such pipeline in its journal
once per key (H-A-313).

Builds: design 47 H H-A-311 (3), H-A-313
2026-11-21 20:44:43 +01:00
jschoubben e137df7189 H-A-311 (2): a step-progress or step-end heard late still places what its end unlocks
A late progress or end moved nothing, so movers waiting on what the
step's end gave stayed unplaced: main.137 placed nothing after its
lint's late progress at 20:17:55.

Builds: design 47 H H-A-311 (2), H-A-286, H-A-288
2026-11-21 20:43:02 +01:00
jschoubben 6c66bf9a5a H-A-311 (1): a placing cut by a concurrent write is done again from the record read again
A placing whose compare-and-set was refused because the pipeline was
written meanwhile was dropped and never done again: main.137 stood
seventeen minutes with eleven steps placed until a restart.

Builds: design 47 H H-A-311 (1), H-A-288
2026-11-21 20:41:29 +01:00
jschoubben b79fd19c84 Teller: hear registration-changed on the views reader's one subscription
Since mesh-sdk heard a seat's event on its bare subject too (H-A-166), the teller's
own hub handler of registration-changed overlapped the one the views reader makes
for the catalog half's followers: the hub refused the reader's, said "a message has
one taker", and registered none of the reader's adopted events after it, so the
catalog half's registrations and source branches were their first views' alone.
The widenings now hear the event through ProcessReader.AdoptEvent beside the
followers; a failed handling is given back for the bus to ask again, as before.
The carry test now fails when any of the reader's events is refused.

issue 1960 (H1), test stability.

Builds: design 47 G A-199, A-222, §532, §335
2026-11-21 19:24:22 +01:00
jschoubben 70bb8f7f51 Teller test bed: answer who holds operator-channel, so a catch-up does not hold the asks' watch
The want-of-room test failed most runs since 20 Nov: the teller's picture of the
assigner's asks hands the first ask on twice (its first view, then again), and the
second sight of an open asking catches up with the router's record, which first asks
the seat mesh-module-assigner's assignments who holds operator-channel. The bed
neither granted nor answered that call, so it waited out its call timeout on the
watch's goroutine, and the second ask, written meanwhile, was not raised within the
test's 10 s. The bed now grants and answers the call as the live seat does, naming
ops-router, so the catch-up reads ops-router_asks at once (20 of 20 runs pass, were
0 of 20 locally).

issue 1960 (H1), test stability.

Builds: design 47 G §570 (issue 1610 AC1, the refused ask asked again), §717
2026-11-21 19:21:24 +01:00
jschoubben 2eaa41a420 H-A-292, H-A-293: builder_publish_queue reads one page; a build is retired only within its series
The review found two faults. The tool paged on through a state of any size, one replay of the bucket per page; it now
reads one page of each of its own states and, when the state holds more, answers that page capped, with the keys read
and the state's size from its info, never asking a second page and never refusing. And the retiring of superseded done
builds keyed the newest reservation and release by module alone, so a standing 0.2 would retire a done 0.1.x; both are
now found per module and series (MAJOR.MINOR).

Builds: design 47 H H-A-292, H-A-293, H-A-291 (1), H-A-290 (5)
2026-11-21 18:59:58 +01:00
jschoubben 90ff867e7a lock the test fakes' maps
Two tests failed under -race on main too: the re-ask state fake's maps were written by a try on the watch's goroutine while
the test read and deleted them, and the operator-ask test read the build store unlocked and waited on the builder while
the build its answer starts could still be counting itself in. The fake's maps are locked and read through it, the store
is read under its lock, and the builder is waited on only once that build has claimed its key.

Builds: design 47 H H-A-290 (the module's tests pass with -race, as its checks ask)
2026-11-21 18:59:58 +01:00
jschoubben 77fb6c4ab1 H-A-291: a start publishes only each module's newest done build; a live enqueue cuts a start list short
A start asked a publish again for every unreleased done build, older ones whose digests a newer reservation or release
of the same module made dead included, and a live publish queued during a long start pass waited for its end. A done
build a newer reservation or release (the builder's own record) supersedes is now retired at the start and at the pass:
its key dropped, its publish item ended, said in the journal; the start asks only each module's newest. A live enqueue
marks its holder, and a pass about to send a start-class item stops there, lets the sends in flight end and lists again,
so the live one goes next; a start-class enqueue never does.

Builds: design 47 H H-A-291 (1), (2), H-A-290, H-A-275
2026-11-21 18:59:58 +01:00
jschoubben bda61ae64c H-A-290: the builder's publish queue sends live publishes first, in passes, never a withdrawn build
The tool runner's release waited forty minutes behind old versions: every start asked a publish again for every done
build, and each send relisted and reread the whole queue. A publish whose build ended in this run now goes before any
asked again at a start (the class kept on the key's value, a live key of a build before the start made a start key at
the start); a pass lists a holder's keys once and sends them four at a time; a withdrawn or taken-over build is dropped
with its publish item ended at enqueue, at the pass and at the start; a send that returns unsent says why, and every call
it makes honours its one bound; builder_publish_queue reads its states in pages and answers counts per holder and class,
the oldest and newest queue time, a list capped at fifty and one module's items with their class.

Builds: design 47 H H-A-290 (1)-(5), H-A-269, H-A-271, H-A-275
2026-11-21 18:59:58 +01:00
jschoubben d51d389c1b Runner: a pipeline with nothing open ends at once (47-H1-runner-end-when-empty)
main.117 judged its end at 17:45:19 and never wrote it: every mover's widens
was read inside the end's compare-and-set loop, 126 movers each cut on
grants-for, so each try took about eleven minutes and every write was refused
as changed until the restart, whose start pass judged no end.

The reads an end needs are made once, before the compare-and-set, shared by
every mover and every retry; the runner's start pass ends each running
pipeline with nothing open before the queue is placed; every end says its
outcome in the journal.

Builds: design 47 H H-A-288, H-A-289, H-A-223, H-A-243
2026-11-21 18:18:32 +01:00
jschoubben 9a49b0f4f7 Give a kept mover's asks ids of its kept revision and place it from that revision (H-A-286)
Review round 1: a kept mover's ask ids still came from the pipeline's start
time, so a pin-held ask or a composition took the id of an ask the runner
had just dropped for the old fingerprint, and a restart never matched the
steps placed again; and a kept mover placed later read the branch's held
view, so a lagging view gave the new fingerprint the old pins. Every ask id
of a kept mover's steps (build, tests, validate, pin-held asks,
composition, rerun, restart) now mixes its kept revision into the id's
sum, and every placing of a kept mover (at once, later, at a restart, a
rerun) reads its entries at its kept revision by number, waiting while
that revision cannot be read. The test's starter gets a clock that moves.

Builds: design 47 H H-A-286 (amending H-A-284; with H-A-283, H-A-285)
2026-11-21 17:58:07 +01:00
jschoubben 47f81051be Keep a same-source re-pin's mover in its pipeline and build it again there (H-A-283)
A module importing packages its own pipeline releases was re-pinned at
each publish; each re-pin revision of the same source took it out of the
running pipeline, and each newer build withdrew the older reservation, so
nothing ever published. A revision of the same source now moves no mover
out of the open pipeline of that source: the pipeline keeps it at its new
fingerprint, ends every step of it superseded (said at the old
fingerprint, its asks dropped, its runs left) and places them again from
the re-pin revision's entry under new asks, the build first, still naming
the pipeline's revision. The revision the mover is kept at rides in the
record's kept-revisions, beside the movers, where the pin check reads its
entry. A revision whose every mover is kept starts no pipeline; one of a
new source moves movers as before.

Builds: design 47 H H-A-283, H-A-284, H-A-285 (with H-A-223, H-A-244, ADR 0682)
2026-11-21 17:47:54 +01:00
jschoubben ece9d6b18f Builder: a published reads each reservation once, and only the unbuilt ones on a miss
Live, a published at a version fourteen reservations share cost three seconds:
the fallback by version read every reservation there, and the journal line
read them all again, so the backlog drained at fifty a minute.

The published now hands the reservations it read (each once, a second pass
taking what the first read) to its journal line. The fallback reads the new
index key of the reservations at a version whose build has not recorded every
digest (kept at reserve and where a digest is recorded, written again at the
start), the only ones a published can be kept on; the line counts the rest from
the index by version without reading them. Before that pass the handler says
in progress to the bus.

Builds: design 47 H H-A-280, H-A-271
2026-11-21 16:45:25 +01:00
jschoubben 5628ca2919 Builder: a build whose number a registry holds elsewhere is built again, never moved
Every artifact carries its version from its build (a Go module zip's paths, a
program's stamped version, an image's and a bundle's build environment), so
no digest survives a change of number: the move that kept digests would have
published a binary saying 0.1.37 under another number.

The try, on a number a registry holds at another digest, withdraws it on the
reservation (a gap) and raises version-held-elsewhere naming both digests,
then asks the build of the same fingerprint again through the build-again
path of a done key whose version is not released. Its reserve never takes a
withdrawn number again, skips every number a registry holds, and the build
records fresh digests and runs its own publish; the takeover ends the old
number's publish item and bound.

Builds: design 47 H H-A-275, H-A-279, H-A-226
2026-11-21 15:56:04 +01:00