Repository navigation
chore(deps): bump bodhi-realtime-agent to pick up GoAway CLOSED-state guard - #409
Merged
Merged
Conversation
…e guard) Picks up the upstream fix from sonichi/bodhi_realtime_agent#4: handleGoAway now early-returns when sessionManager.state !== 'ACTIVE', preventing the unhandled SessionError when a late GoAway arrives after a reconnect has already torn down the transport. Observed frequency in our voice-agent.log (per notes/bodhi-closed- reconnecting-fatal.md): 71× CLOSED → RECONNECTING + 2× RECONNECTING → RECONNECTING FATALs. All resolved upstream in this commit. Lockfile-only change — package.json pin is `github:sonichi/bodhi_realtime_agent` (unchanged), npm install re-resolved to the new HEAD commit de64333. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
sonichi
commented
Apr 17, 2026
sonichi
left a comment
sonichi
left a comment
Owner
Author
There was a problem hiding this comment.
LGTM. Cold-reviewed:
- Bumps
node_modules/bodhi-realtime-agentfrom33a08c0tode64333inpackage-lock.json. Verifiedde64333is the squash-merge commit of bodhi PR #4 (fix(voice-session): guard handleGoAway against late-arriving GoAway after CLOSED). - 1 line change, no functional sutando code touched.
- Pulls in the state-guard that closes the production-observed FATAL class (24/24 of 5000 voice-agent log lines were the same
Invalid transition: CLOSED → RECONNECTING).
Ready to ship. Voice-agent restart will pick it up — confirms the upstream fix in our dist after merge.
sonichi
pushed a commit
that referenced
this pull request
Apr 17, 2026
Mini overnight deliverable #3 of 4. Runs ~8 checks in <20s and prints PASS/WARN/FAIL for each. Exits 1 on any FAIL, 2 on any WARN, 0 on all-PASS so it integrates cleanly with other scripts / CI. Checks: voice-agent port 9900 listening + log mtime <60s bodhi state machine 0× CLOSED→RECONNECTING FATALs in current log (post-#409) conversation-server port 3100 /health = ok ngrok tunnel curl public URL, expect 200-range presenter-mode sentinel present + future expiry Claude quota >=10% remaining disk space >5 GB free memory-sync last sync <6h Designed for the 5-10 min pre-talk window. Owner runs it once, sees the summary, triages the WARNs before going on screen. Local test: 5 pass, 3 warn, 0 fail. Warnings are all actionable: - ngrok tunnel unreachable (SSL handshake fails — separate bug, non-blocking for remote ICLR talk which uses Zoom audio routing, not phone) - presenter-mode INACTIVE — expected, run before talk - memory-sync 9h — run `bash src/sync-memory.sh` before talk Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This was referenced Apr 17, 2026
sonichi
added a commit
that referenced
this pull request
Apr 17, 2026
* feat: stage-readiness.sh — single-command pre-flight check for ICLR talk Mini overnight deliverable #3 of 4. Runs ~8 checks in <20s and prints PASS/WARN/FAIL for each. Exits 1 on any FAIL, 2 on any WARN, 0 on all-PASS so it integrates cleanly with other scripts / CI. Checks: voice-agent port 9900 listening + log mtime <60s bodhi state machine 0× CLOSED→RECONNECTING FATALs in current log (post-#409) conversation-server port 3100 /health = ok ngrok tunnel curl public URL, expect 200-range presenter-mode sentinel present + future expiry Claude quota >=10% remaining disk space >5 GB free memory-sync last sync <6h Designed for the 5-10 min pre-talk window. Owner runs it once, sees the summary, triages the WARNs before going on screen. Local test: 5 pass, 3 warn, 0 fail. Warnings are all actionable: - ngrok tunnel unreachable (SSL handshake fails — separate bug, non-blocking for remote ICLR talk which uses Zoom audio routing, not phone) - presenter-mode INACTIVE — expected, run before talk - memory-sync 9h — run `bash src/sync-memory.sh` before talk Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(stage-readiness): portable stat + df + remove dead ngrok conditional Three real bugs surfaced by MacBook's cold-review on a Mac-with-Homebrew- coreutils dev environment (the common setup — MacBook's Apr 26 pre-flight will run under this shell). 1. `stat -f %m` BSD syntax fails silently on GNU coreutils: GNU emits filesystem info to stdout AND exits non-zero. The original `BSD || GNU` fallback in `$(...)` concatenated both outputs, then `set -u` arithmetic expansion died with "File: unbound variable" and silently aborted remaining checks. Fix: new `stat_mtime()` helper tries GNU first (works on Linux + coreutils-on-Mac), BSD as fallback, explicit `echo 0` as second fallback. Order matters — GNU first is correct because GNU's presence is the bug trigger; pure BSD Macs without coreutils work cleanly either way. 2. `df -g` is BSD-only. GNU df doesn't support `-g`, so the disk check silently printed nothing and was skipped entirely. Fix: POSIX `df -k` returns KB universally; convert to GB via arithmetic. 3. Dead `[[ "$probe_url" != */* ]]` conditional on the ngrok probe: the pattern matches any URL containing `/` — i.e., every full URL, because `https://` has slashes — so the branch always triggers regardless of path. The intended "append /health only if no path" semantic doesn't work. Fix: just unconditionally append `/health` since all Twilio-bound ngrok webhooks need it. Tested on this Mac: `bash scripts/stage-readiness.sh` → 6 pass, 2 warn (ngrok SSL handshake + presenter-mode INACTIVE are expected pre-talk), exit 0. Credit: MacBook's cold-review-log.md entry on #412 caught (1) and (2); (3) I'd already flagged in build_log pass 435 but hadn't fixed. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Chi <wangchi@Chis-Mac-mini.local> Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
liususan091219
pushed a commit
to liususan091219/sutando-fork
that referenced
this pull request
May 28, 2026
…e guard) (sonichi#409) Picks up the upstream fix from sonichi/bodhi_realtime_agent#4: handleGoAway now early-returns when sessionManager.state !== 'ACTIVE', preventing the unhandled SessionError when a late GoAway arrives after a reconnect has already torn down the transport. Observed frequency in our voice-agent.log (per notes/bodhi-closed- reconnecting-fatal.md): 71× CLOSED → RECONNECTING + 2× RECONNECTING → RECONNECTING FATALs. All resolved upstream in this commit. Lockfile-only change — package.json pin is `github:sonichi/bodhi_realtime_agent` (unchanged), npm install re-resolved to the new HEAD commit de64333. Co-authored-by: Chi <wangchi@Chis-Mac-mini.local> Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
liususan091219
pushed a commit
to liususan091219/sutando-fork
that referenced
this pull request
May 28, 2026
…i#412) * feat: stage-readiness.sh — single-command pre-flight check for ICLR talk Mini overnight deliverable sonichi#3 of 4. Runs ~8 checks in <20s and prints PASS/WARN/FAIL for each. Exits 1 on any FAIL, 2 on any WARN, 0 on all-PASS so it integrates cleanly with other scripts / CI. Checks: voice-agent port 9900 listening + log mtime <60s bodhi state machine 0× CLOSED→RECONNECTING FATALs in current log (post-sonichi#409) conversation-server port 3100 /health = ok ngrok tunnel curl public URL, expect 200-range presenter-mode sentinel present + future expiry Claude quota >=10% remaining disk space >5 GB free memory-sync last sync <6h Designed for the 5-10 min pre-talk window. Owner runs it once, sees the summary, triages the WARNs before going on screen. Local test: 5 pass, 3 warn, 0 fail. Warnings are all actionable: - ngrok tunnel unreachable (SSL handshake fails — separate bug, non-blocking for remote ICLR talk which uses Zoom audio routing, not phone) - presenter-mode INACTIVE — expected, run before talk - memory-sync 9h — run `bash src/sync-memory.sh` before talk Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> * fix(stage-readiness): portable stat + df + remove dead ngrok conditional Three real bugs surfaced by MacBook's cold-review on a Mac-with-Homebrew- coreutils dev environment (the common setup — MacBook's Apr 26 pre-flight will run under this shell). 1. `stat -f %m` BSD syntax fails silently on GNU coreutils: GNU emits filesystem info to stdout AND exits non-zero. The original `BSD || GNU` fallback in `$(...)` concatenated both outputs, then `set -u` arithmetic expansion died with "File: unbound variable" and silently aborted remaining checks. Fix: new `stat_mtime()` helper tries GNU first (works on Linux + coreutils-on-Mac), BSD as fallback, explicit `echo 0` as second fallback. Order matters — GNU first is correct because GNU's presence is the bug trigger; pure BSD Macs without coreutils work cleanly either way. 2. `df -g` is BSD-only. GNU df doesn't support `-g`, so the disk check silently printed nothing and was skipped entirely. Fix: POSIX `df -k` returns KB universally; convert to GB via arithmetic. 3. Dead `[[ "$probe_url" != */* ]]` conditional on the ngrok probe: the pattern matches any URL containing `/` — i.e., every full URL, because `https://` has slashes — so the branch always triggers regardless of path. The intended "append /health only if no path" semantic doesn't work. Fix: just unconditionally append `/health` since all Twilio-bound ngrok webhooks need it. Tested on this Mac: `bash scripts/stage-readiness.sh` → 6 pass, 2 warn (ngrok SSL handshake + presenter-mode INACTIVE are expected pre-talk), exit 0. Credit: MacBook's cold-review-log.md entry on sonichi#412 caught (1) and (2); (3) I'd already flagged in build_log pass 435 but hadn't fixed. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Chi <wangchi@Chis-Mac-mini.local> Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
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
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.
Summary
Bumps `bodhi-realtime-agent` lockfile pin from `33a08c0` → `de64333` to pick up upstream PR bodhi#4 (just merged): `handleGoAway` now early-returns when `sessionManager.state !== 'ACTIVE'`, preventing the unhandled `SessionError: Invalid transition: CLOSED → RECONNECTING` FATAL when a late GoAway arrives after the reconnect has already torn down the transport.
Why
Per `notes/bodhi-closed-reconnecting-fatal.md`, current logs show 71× CLOSED → RECONNECTING + 2× RECONNECTING → RECONNECTING FATALs. These are "staying alive" rejections — they don't crash the process, but each represents a silent transport churn that likely contributes to "Gemini didn't respond" latency spikes even when voice-agent looks healthy. All resolved upstream in `de64333`.
Scope
Lockfile-only change. `package.json` pin is `github:sonichi/bodhi_realtime_agent` (unchanged, no version pin). `npm install bodhi-realtime-agent` re-resolved to the new HEAD commit. +1/-1.
Note per `feedback_no_npm_install_during_rename_sweeps`: this is an intentional lockfile bump of a single known dependency, not a sweep — no other lockfile entries changed.
Test plan
🤖 Generated with Claude Code