103 Commits

Include renames

Author	SHA1	Message	Date
jschoubben	a76910a0c0	Say only the merge tool's two out of date refusals a warning, and never step a severity down (issue 1275, ADR 0442's cases)	2026-10-19 20:51:58 +02:00
jschoubben	dc86203d29	Say a refusal for an out of date verdict a warning for ten minutes, urgent after (issue 1275, operator's amendment, AC6)	2026-10-19 20:51:58 +02:00
jschoubben	301cd2287f	Answer gitea_commit_statuses in the bus test's runtime, and say a held, refused member once (issue 1275, review round 1)	2026-10-19 20:38:48 +02:00
jschoubben	524e3405e5	Ask a group's Approve only when every member may merge, and withdraw the ask when a required status turns (issue 1275) ... An Approve given hours after its ask must not rest on a check made at the ask: a required status that turns not a success while the ask waits takes the ask back, and its late answer merges nothing.	2026-10-19 20:32:22 +02:00
jschoubben	6e6d1f3e47	Merge a group whole or not at all: re-read every required status first, stop at the first refusal, say a held or half merged group urgent (issue 1275, 1276) ... An approved group was merged member by member, each on its own, so a refusal of the first did not stop the second. A group now merges in one pass in its order, after every member's required statuses are read again at its head, and a held or half merged group stays raised. A refusal naming the member's own repository check as missing asks that check (issue 1276).	2026-10-19 20:32:15 +02:00
jochen	a32b7af3fd	Ask a refused merge again when its checks turn green on the forge (issue 1144)	2026-10-18 23:50:25 +02:00
jschoubben	6954467292	mesh-delivery bus test waits for the group's composed check, not a member's own check asked at once (issue 1115)	2026-10-18 12:33:02 +02:00
jschoubben	5f002441b7	mesh-delivery takes an empty answer to a check ask as a refusal (issue 1115)	2026-10-18 12:28:03 +02:00
jschoubben	fc9ddd9185	mesh-delivery adopts a standing verdict only when every required check stands, at every look, and counts an ask with no id as none (issue 1115)	2026-10-18 12:25:53 +02:00
jschoubben	1760b94824	mesh-delivery waits for nothing: a check is asked at once and a standing verdict adopted (issue 1115)	2026-10-18 12:22:31 +02:00
jschoubben	30d3bb9466	A pull request announced at a live head takes its delivery over when the other one closes or is edited (issue 1114)	2026-10-18 11:53:28 +02:00
jschoubben	bc1acddaa4	mesh-delivery hears the forge's own pull-request closed event as a close unmerged (issue 1107)	2026-10-18 04:55:48 +02:00
jschoubben	b0ec4f6f14	A group member refused for a pending status of its own asks its own check once, not the composed check (issue 1049)	2026-10-18 04:27:42 +02:00
jschoubben	7383bc3170	A stale composed pass counts members judged since it was asked, whichever verdict lands first (issue 1047, review round 2)	2026-10-18 01:20:18 +02:00
jschoubben	ebd29223f3	Prove over a real bus that an edit of the description asks the Approve again (issue 1079, review round 1)	2026-10-18 01:18:43 +02:00
jschoubben	e8a74765f8	A composed check that lands judged against a trunk that has moved is out of date and asked again (issue 1047, review round 1)	2026-10-18 01:18:13 +02:00
jschoubben	6556e61485	Read a pull request's review record again when its description is edited, and ask the Approve again if it was unanswered (issue 1079)	2026-10-18 01:12:40 +02:00
jschoubben	93365be98f	A new pull request at a commit whose delivery was stopped and unmerged takes that delivery over (issue 1083)	2026-10-18 01:10:51 +02:00
jschoubben	814c1a991c	A trunk move asks at once the check of a group member whose Approve is not asked yet (issue 1047)	2026-10-18 01:10:12 +02:00
jschoubben	983e4a6444	A closed group's member whose merge was refused asks the composed check again at once, and a refusal loop past ten minutes is said (issue 1082)	2026-10-18 01:09:37 +02:00
jschoubben	da094008ea	A group's composed verdict names each member's siblings at the heads it composed, so a sibling's own merge is no move (issue 1049)	2026-10-18 01:08:26 +02:00
jschoubben	c898a937a1	Order the events of the failed-backup bus test: each after the module took the one before (issue 1068)	2026-10-17 22:57:44 +02:00
jschoubben	82693a441c	A release the controller refuses to take leaves the delivery held for its failed backup (issue 1068)	2026-10-17 22:53:43 +02:00
jschoubben	4799f79242	Read the plan's backups as the controller says them: machine, module, state, words (issue 1068)	2026-10-17 22:45:46 +02:00
jschoubben	8f7aeeeb81	Hold a delivering delivery for a failed backup before the send; release asks the backups again (issue 1068)	2026-10-17 22:45:03 +02:00
jschoubben	47e42806ff	Leave a composed member's gate to the group's verdict, reword the mark leftovers, restore main's merge-check.sh (issues 1022, 978)	2026-10-17 21:33:03 +02:00
jschoubben	2d3b61ec45	Merge a group in one pass on its one Approve, with no order and no turns (issue 978) ... A group has no order: mesh-delivery asks the controller for none, reads no after: line (it says once that the line is not read), merges every member in one pass once the composed check passed and the Approve covers the heads, and gives the plan its word once every member is merged. A failed member stops no other. The composed verdict sets each member's mesh/merge-gate (pending, saying why, until one stands), the groups answer says composed_gate, and a trunk move asks the out-of-date members' checks at once and merges on the same Approve (issue 1047).	2026-10-17 20:54:16 +02:00
jschoubben	668444785d	Ask the Approve as soon as the checks pass, with no ready mark (issue 1022) ... The review finishes before the push and its record is a Reviewed-by line in the description, so the mark repeated what the push meant. mesh-delivery asks the operator when a pull request's checks pass, keeps an ask through a trunk move by asking the check again at once, and hands the merge tool the approved heads, which replace the forge review of each sibling.	2026-10-17 20:44:54 +02:00
mesh-tools	652b00b25b	Say on every directory how it treats files it does not know, and check it (issue 1024, ADR 0422) ... Every directory and archive resource carries `unknown`: move-out for a home folder whose program writes nothing there (the node-engine, delivered in mesh-host #117, moves the rest to ~/.local/state/mesh-moved/<module>/ with a list), module for ~/.ssh and ~/.claude, program elsewhere; `knows` keeps what a program or the operator relies on (powerlevel10k's compiled files, i3's scripts and the asus/g14 drop-ins, dunst's slack rule, betterbird's records, node-env's forge credential, issue 1033). Two shared folders are no longer declared. Each module's README says it. checks/owned holds every manifest to it; checks/sshdir stays while main's merge check runs it (issue 1035). Tests that pin a manifest follow the new key only. Reviewed in three rounds.	2026-10-17 16:23:03 +00:00
jschoubben	3ec2881a27	issue 1025: the merged announcement is an owed effect, and a group's merge is tested	2026-10-17 15:41:43 +02:00
jschoubben	4c0ed75090	issue 1025: mesh-delivery announces each merge on the Approve as the event merged, with its warrant	2026-10-17 15:40:00 +02:00
jschoubben	52f70f42b7	issue 975: a merge heard again changes nothing, and a verdict naming no trunk commit is not written to the key	2026-10-17 14:10:14 +02:00
jschoubben	2c8c1543ba	issue 975: the pull request's key written by compare-and-set, the ready flag on a ready mark's check, build-agent reads the key	2026-10-17 13:48:54 +02:00
jschoubben	27ca8bbcff	issue 975: the acceptance over a real bus: a ready mark and its Approve are merged within seconds of their events	2026-10-17 13:28:21 +02:00
jschoubben	0809a1f982	issue 975: a test that fails on a ticker and on a timer that names no bounded step it retries	2026-10-17 13:25:33 +02:00
jschoubben	b5225d3a2f	issue 975: no forge poll; the open pull requests are read at start and when a merge check's status, a new head or a ready mark says they changed	2026-10-17 13:25:03 +02:00
jschoubben	d675f5e7c6	issue 975: no tick and no reconcile loop; a look after each event, one timer for the steps that wait on the clock	2026-10-17 13:23:25 +02:00
jschoubben	95e11464fe	issue 975: ready marks, verdicts and Approves ask the merge as events, not at the next minute	2026-10-17 13:19:01 +02:00
fix-deliv-awg@novox	4cccd817a0	End a group's ask and Approve when any member is pushed (issue 948, ADR 0398 §3)	2026-10-16 21:36:06 +02:00
fix-deliv-awg@novox	f61e00aeaa	Ask the operator to Approve a merge into main and merge only on the warrant (issue 948, ADR 0398 §3)	2026-10-16 21:29:55 +02:00
fix-deliv-awg@novox	417f69567c	Merge only into main: a ready mark on another base is left to its author (issue 948, ADR 0398 §4)	2026-10-16 21:26:53 +02:00
jschoubben	2d1b20364f	mesh-delivery: an empty-head ready mark is cleared only after a minute, and not right after mark-ready; clearer wording (issue 913, issue 873)	2026-10-16 14:35:58 +02:00
jschoubben	741b729ad4	mesh-delivery: mark-ready labels then comments the head in the order the forge counts; an empty-head clear says why (issue 913, issue 873)	2026-10-16 14:33:50 +02:00
mesh-tools	8efe499db9	A delivering group says success on mesh/delivery-group, so its later members can merge (issue 902)	2026-10-16 09:12:53 +00:00
mesh-toolsandClaude	c7c784161d	Put a group's composed check out of date when a member is rechecked (issue 867, ADR 0323 §6) ... After mesh-host#114's recheck passed, group 57-config-and-provisioning kept its failed composed verdict for the same heads and was never asked again. A recheck now marks the member's group's composed check out of date, answered or still running, so it is asked again once every member is judged again and ready; a late answer from the old run is dropped. Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-16 01:17:58 +00:00
mesh-toolsandClaude	de9d29dc65	Keep a group member's siblings and refuse its merge until the group passed and every sibling is reviewed (issue 867, ADR 0382) ... mesh-delivery's groups names each member's pull request, head, base and whether it is open (member_details), so the controller can read a member's siblings beside its check. The forge's holder keeps the siblings each verdict read, puts that verdict out of date when a sibling gets a new head, closes, or its base moves by anything but its own merge, and the merge tool merges such a member only when mesh/delivery-group passed at the heads read and each sibling head is approved. The statuses name the siblings, or say they could not be learned. Co-Authored-By: Claude <noreply@anthropic.com>	2026-10-15 23:27:13 +00:00
jochen	ab739d02dd	mesh-delivery: a merge that starts over a stopped duplicate is made from the forge's word too (issue 783)	2026-10-15 15:01:23 +02:00
jochen	9e8521afb5	mesh-delivery: a stopped duplicate's delivery follows the pull request that merges it (issue 783) ... A delivery keyed by commit stayed stopped forever once an earlier, closed duplicate pull request with the same head had stopped it, even though a later pull request went on to merge that very commit: the merge only set merged_as on the stopped record, and the walk that carried it waited for this owner's word with no delivery left to give it. PullUpdated and PullMerged now start such a delivery over, under the same id, for the pull request that announces or merges it next, naming both pull requests in the transition's why. Any other stop (a person's, a group's, the same pull request's own close) is untouched.	2026-10-15 14:50:57 +02:00
jschoubben	373f339750	Give a walk the word of a stopped delivery that waits for it (hq issue 753) ... A delivery its group stopped for a member delivered since waits for a later walk of its repository that contains its merge (issue 747). When that walk carried no living delivery, nothing gave it its word and it waited forever. The delivery now takes its turn on the walk that answers its merge, in its group's order. A walk owed such a word stays under the orphans' timer, so past the grace without it `stalled` lists it with why (held, or the controller's refusal), and a walk no living delivery holds names the ended deliveries it carries.	2026-10-15 00:30:26 +02:00
jschoubben	6f7f07635e	Cover a failed or stopped delivery once a later finished walk contains it (hq issue 747) ... A group member whose walk failed in an outage stayed failed after a later walk of its repository, containing its merge, was done; its group then stopped every later walk of the other repository, and the catalogue could deliver nothing. A failed or stopped delivery now reads covered when a done walk of its repository, made after its merge, whose commit contains it (the forge's word, as issue 362's containment), built every module it moves. A group counts it as delivered, and a member the group stopped for it no longer holds a walk back: the group stops a walk only for a member whose change is still undelivered. A stop covered under issue 480 now likewise makes its group read reached, not stopped. While the forge is asked, the walk is held back rather than stopped, for at most ten minutes from the first ask: a forge that keeps erroring or cannot say then lets the walk be asked to leave the stopped merge out, as before. The hold is said by waits, stalled and stalled walks. The pass reads kept walks each tick, so today's stuck deliveries clear after a restart.	2026-10-14 23:19:06 +02:00