17 check-ins that include changes to files matching 'internal/work/*'
|
2026-09-27
| ||
| 11:38 | Merge trunk into background-jobs-follow-the-lease check-in: cf04732cdd user: dmestas tags: background-jobs-follow-the-lease | |
| 11:27 | update node-identity-is-persisted from trunk closed check-in: f61888abaf user: zeitgeist tags: node-identity-is-persisted | |
| 11:23 | update ci-race-on-landings from trunk closed check-in: 3bfe3ae2d8 user: zeitgeist tags: ci-race-on-landings | |
| 11:05 | update visibility-list-refreshes-on-a-ttl from trunk closed check-in: 637868ecfb user: dmestas tags: visibility-list-refreshes-on-a-ttl | |
| 11:04 | update ci-attempt-of-and-settled-verdict from trunk closed check-in: a8270cc634 user: dmestas tags: ci-attempt-of-and-settled-verdict | |
| 11:04 | update ops-history-flush-off-request-path from trunk closed check-in: f558d02eb6 user: dmestas tags: ops-history-flush-off-request-path | |
| 11:02 | land ingest-lane-shed-is-backpressure check-in: 5ad81f099d user: dmestas tags: trunk | |
| 05:45 | A stopped gap repair is not recorded as tried, and only the lane's own refusal of a round counts as a shed check-in: 1d1710a957 user: dmestas tags: ingest-lane-shed-is-backpressure | |
| 05:30 | A round that becomes a rebuild leaves the ingest lane for good, and rounds take the store lock before the lane, so a parked rebuild can no longer deadlock the lane check-in: 053062a349 user: dmestas tags: ingest-lane-shed-is-backpressure | |
| 04:44 | Ingest rounds shed at the lane back off and are journalled once past a deadline, and a round queued for a rebuild slot leaves the lane while it waits check-in: 4892dad217 user: dmestas tags: ingest-lane-shed-is-backpressure | |
|
2026-09-12
| ||
| 17:19 | merge trunk into landing-addresses closed check-in: 014bb9663b user: dmestas tags: landing-addresses | |
| 16:30 | land lane-held check-in: a71af4e934 user: dmestas tags: trunk | |
| 16:18 | ops: the lanes panel lists what each lane is holding, with verb, subject and age closed check-in: 39d8bd654e user: dmestas tags: lane-held | |
|
2026-09-06
| ||
| 12:38 | landing: decision controls skip the override validation that silently blocked them closed check-in: ef78b6d72f user: dmestas tags: fleet-edit-mirror | |
|
2026-09-03
| ||
| 17:51 | land error-context check-in: fe6e4659ca user: dmestas tags: trunk | |
| 15:26 | Errors carry what a reader needs to act: internal/fault, fingerprint grouping, cost and fields on the journal check-in: 44d79cf15b user: dmestas tags: error-context | |
|
2026-08-17
| ||
| 21:26 |
work: bound bulk operations; footer times keep their precision
Nothing in quarry limited concurrency — every zit subprocess ran in the request goroutine with the arrival rate as the only bound. work.Lane is a concurrency limit plus a queue cap that SHEDS past it: a request accepted into a queue it cannot be served from consumes the capacity that would have served someone else. Wired to archive first, which buffers 256 MiB per request on an anonymous route. That is latent rather than live — archive 503s on zit#97 — which is why the bound belongs there now, so the endpoint is born bounded rather than acquiring a limit after someone finds it. Footer precision: '<1ms' was true and useless once most pages went sub-millisecond, collapsing the interesting range into one label. timing.Format keeps ~3 significant figures across five orders of magnitude, so 900µs and 90µs no longer read the same. Fossil has the mirror-image bug: three decimals of seconds prints everything faster than a millisecond as 0.000s. check-in: 45bb75ade2 user: dmestas tags: trunk | |