Repository navigation
Fixed
- The
/remoteclose in the TUI followsthread_close's close contract (#14234). A close refused because the thread has messages it has not read names/remote close <session> --forceand lists those messages, or says the ids are unavailable on that device's engine. The picker then offers "Close anyway". A forced close lists the messages it dropped. The refusal'sdetails.reasonis the engine's own reason (turn_activeorwork_pending). (#14246, fixes #14245) (#14248) - Thread tools: a command for a thread whose session parked on another device after it was listed no longer fails
unsupportedwith advice to update omo on that device. The device service answers such a commandroute_unlisted(its host no longer lists that session, so nothing was sent to it).thread_closenow answersclosed: false, reason: "not_live", and the other thread tools answernot_resumable, as for a session that ended there. A create, or a reopen's session-file lookup, refused that way ishost_unavailable, with a retry. (#14237, fixes #14233) (#14248) - The cold-wake session host unit (and every other unit or plist the cold-wake install writes) is now written 0644 under any umask. Under umask 077 it was written 0600, because a file's creation mode is masked by the umask; the mode is now set explicitly before the file is moved into place. (#14249, fixes #14247) (#14248)
- A cold-wake unit or plist an earlier build already left 0600 is set back to 0644 the next time the install (or a refresh to a newer runtime) finds its content unchanged, with no reload or restart. A file this user cannot change the mode of (another owner's, or on a read-only filesystem) is left as it is and the install still succeeds. (#14249) (#14248)
Changed
thread_closeno longer drops messages a thread has not read unless you ask it to. Before, it always interrupted the session and silently dropped any messages queued behind its run. Now:- A busy thread with unread messages is refused
turn_active. The refusal names those messages (pending_deliveries) and gives theforceoption. thread_close({ thread, force: true })closes it anyway.- Every close result carries
dropped_deliveries, the messages it handed back unread (an empty list when there were none). - An idle thread still closes without
force, even when an interrupted run left messages stranded (#14169); those are listed too. - On a device whose engine cannot name its queued messages (older than code-yeongyu/senpi#7006), a default close never interrupts. It says the message ids are unavailable there, and
force: truestill closes it. omo remote closetakes--forceand reports the dropped messages as well. (#14242, fixes #14234) (#14248)
- A busy thread with unread messages is refused
Added
-
The desktop can ask OmO which model each category and agent runs. A new read-only extension request,
omo.modelRouting.resolve, answers the model every category (under its canonical name) and every task agent would spawn on, using the same resolvers ataskcall uses against the session's project config, model registry and sign-ins, and says whether that model came from your omo.json or from OmO's defaults. (#14025) -
822bd92 Merge pull request #14290 from code-yeongyu/release/v5.1.79-source-state
-
fd7c18a Merge pull request #14286 from code-yeongyu/ratchet-no-ignore
-
2d1b133 fix(lint-ratchet): the lint ratchet refuses a file oxlint skipped, so a config's ignorePatterns cannot hide one either
-
0bed730 fix(lint-ratchet): lint ratchet runs oxlint with --no-ignore, so an ignore file cannot hide a file from it
-
abfb5be Merge pull request #14248 from code-yeongyu/release-train/omo-20261119-thread-train-3
-
7559752 Merge pull request #14267 from code-yeongyu/fix/11845-progress-aware-spawn-budget
-
5e628bc Merge pull request #14268 from code-yeongyu/feat/no-bun-mock-module
-
c7367ed test(omo-native): review fixes for the progress-aware budget
-
07b5060 test(script): cover createRequire and members read off a loaded bun:test
-
42ef9ed fix(script): no-module-mock also sees bun:test loaded through createRequire
-
94f3fcc feat(script): ratchet module mocks in tests (no-module-mock)
-
e9940fb test(omo-native): bound daemon spawn waits by child CPU progress, not wall clock
-
8356fb8 Merge pull request #14259 from code-yeongyu/fix/14257-published-smoke-tarball-wait
-
1de932c fix(ci): fail fast on invalid smoke wait overrides; name sub-minute waits in seconds
-
a20f5a8 fix(ci): wait for latest's tarball to commit before the published install smoke packs it
-
50641b1 Merge dev 95e7eb5 (#14239, #14025) into thread train 3
-
95e7eb5 Merge pull request #14025 from code-yeongyu/feat/11344-model-routing-resolve-2
-
452b41d Merge pull request #14239 from code-yeongyu/p0/relcwd-owned-folder
-
608ebb5 Merge dev 5b42751 (release v5.1.78) into thread train 3
-
ff717ee Merge origin/dev (5b42751) into feat/11344-model-routing-resolve-2
-
5b42751 Merge pull request #14243 from code-yeongyu/release/v5.1.78-source-state
-
12bdc2c Merge #14249 (cold-wake unit mode 0644 under any umask, Fixes #14247) into thread train 3
-
24ed73e Merge #14227 (drop the dead no-endpoint check in device-service-core) into thread train 3
-
053403d fix(cold-wake): make the unchanged-unit mode repair best effort (#14249)
-
4543cc7 fix(cold-wake): repair the mode of an unchanged unit left 0600 (#14247)
-
a51b95f Merge current dev (train-3 prep)
-
bc80fcf docs(changes): fold the thread train 3 members' fragments into 14248.md
-
7bc8c2b docs(changes): note the cold-wake unit mode fix (#14249)
-
9136ad1 fix(train): move the two-CLI-generations comment back above typedFailure (#14246 review LOW, comment only)
-
3dadbb1 fix(cold-wake): write the session-host unit 0644 regardless of umask (#14247)
-
e323a36 Merge #14246 (L1/L2 follow-up of #14242, Fixes #14245) into thread train 3
-
645afac test(thread): train 3 resolution - #14237's route_unlisted close expects #14242's dropped_deliveries
-
c822134 fix(remote): /remote close carries the close contract (force, dropped ids, surface-named hint); unread refusal keeps the engine's reason (#14245)
-
f30e165 Merge #14242 (thread close unread, Fixes #14234) into thread train 3
-
84ce4a7 Merge #14237 (thread route unlisted, Fixes #14233) into thread train 3
-
81368b7 chore(thread): un-export two module-local constants (values unchanged) and format (#14242 static checks)
-
2244531 fix(thread): thread_close never drops unread deliveries unless forced; dropped_deliveries on every close (#14234)
-
a51ee84 test(thread): correct the cleanup comment (review L1 of #14239): bun:test already awaits an async hook; the wrapper exists to remove the owned folder
-
99994ce fix(thread): an endpoint-level route_unlisted (reopen lookup) is host_unavailable, one mapping for open and find_session (#14237 review L1)
-
e462f2a test(thread): the relative-cwd case owns its caller folder instead of assuming a repo-root cwd (#14238)
-
2c005dd fix(thread): answer route_unlisted as a session the device no longer runs, not 'update omo' (#14233)
-
916de04 refactor(device-service-core): drop the dead no-endpoint check after the endpoint lookup
-
199099f Merge origin/dev (630a673) into the branch; regenerate bundles once
-
76a9861 chore(build): regenerate bundles after merging dev
-
e0590e5 Merge dev into feat/11344-model-routing-resolve-2
-
847200f Merge remote-tracking branch 'origin/dev' into feat/11344-model-routing-resolve-2
Thank you to 1 community contributor:
bun add -g omo-ai