Repository navigation
Conversation
alt-glitch
added
type/bug
teknium1
added a commit
that referenced
this pull request
Sep 15, 2026
Serialize room drives through their actual member completion, freeze input watermarks by retained entry identity, and share the same completion path with handoff continuations. Stop discards queued work without releasing an active owner early; rename follows the existing room binding. Observe stranded replies for the hard-cap duration plus grace after the foreground wait, and retain unresolved failures in collapsed Activity. Never automatically retry an ambiguous failed submit within the same drive. Adapted from the queue and boundary approach in #92041 by @enwaiax and harvest-budget approach in #107193 by @Finn763; #106502 by @wadib identified failed-submit watermark consumption. The implementation retains current numeric watermark storage, room lifecycle bindings and serial round limits. Related: #92003, #105247, #100026
teknium1
added a commit
that referenced
this pull request
Sep 15, 2026
Serialize room drives through their actual member completion, freeze input watermarks by retained entry identity, and share the same completion path with handoff continuations. Stop discards queued work without releasing an active owner early; rename follows the existing room binding. Observe stranded replies for the hard-cap duration plus grace after the foreground wait, and retain unresolved failures in collapsed Activity. Never automatically retry an ambiguous failed submit within the same drive. Adapted from the queue and boundary approach in #92041 by @enwaiax and harvest-budget approach in #107193 by @Finn763; #106502 by @wadib identified failed-submit watermark consumption. The implementation retains current numeric watermark storage, room lifecycle bindings and serial round limits. Related: #92003, #105247, #100026
Collaborator
|
Thanks @Finn763. Your cap-derived harvest approach landed in #111283 at 9326d9c. Native Electron with real held inference and explicitly accelerated observation clocks reproduced the old 60-tick stop and verified the late reply arrives after the extended watch. Closing this harvest-only PR as superseded. |
karljohannisson
pushed a commit
to karljohannisson/hermes-agent
that referenced
this pull request
Sep 15, 2026
Serialize room drives through their actual member completion, freeze input watermarks by retained entry identity, and share the same completion path with handoff continuations. Stop discards queued work without releasing an active owner early; rename follows the existing room binding. Observe stranded replies for the hard-cap duration plus grace after the foreground wait, and retain unresolved failures in collapsed Activity. Never automatically retry an ambiguous failed submit within the same drive. Adapted from the queue and boundary approach in NousResearch#92041 by @enwaiax and harvest-budget approach in NousResearch#107193 by @Finn763; NousResearch#106502 by @wadib identified failed-submit watermark consumption. The implementation retains current numeric watermark storage, room lifecycle bindings and serial round limits. Related: NousResearch#92003, NousResearch#105247, NousResearch#100026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What Problem This Solves
#105247 Issue 1:
harvestStrandedUntilSettled(group-rounds.ts) polled every5s for a fixed 60 tries = 5 minutes. A member turn that is visibly still
working legally runs up to
GROUP_TURN_HARD_CAP_MS(20 min). If the roomstayed quiet after the base timeout (no new send, no fresh harvest pass), a
reply landing after minute 5 was never delivered.
Root Cause
The harvest bound was a magic number disconnected from the turn lifecycle it
protects. The cap owner (group-turns.ts) and the watcher (group-rounds.ts)
never shared the constant.
Changes
group-turns.ts: exportGROUP_TURN_HARD_CAP_MS+ new puregroupStrandedHarvestMaxTries(intervalMs)= max(60, ceil(hardCap/interval)group-rounds.ts: use it instead of the hardcoded 60; doc comment updated.group-harvest-bound.test.ts: 3 cases (covers cap+margin, old floor,interval scaling). Sabotage-verified (bound forced to 60 -> 2 red).
Evidence
src/plugins/hermes-botssuite 64 files /587 tests green;
tsc -p . --noEmitclean.Scope
Partial: addresses the Issue 1 (harvest window) half of #105247. Issue 2
(silent member spawn/slot failures + retry + system notice) is not included:
no slot-busy error signature is observable from the desktop tree (zero hits
for slot/busy/pool-exhaustion strings in the plugin), so a transient-error
retry detector would be guesswork. Happy to follow up once the backend
error code is named.