Skip to content

fix(engine): preserve replay sequence after send failure - #334

Draft
willwashburn wants to merge 1 commit into
mainfrom
codex/message-routing-hardening
Draft

fix(engine): preserve replay sequence after send failure#334
willwashburn wants to merge 1 commit into
mainfrom
codex/message-routing-hardening

Conversation

@willwashburn

@willwashburn willwashburn commented Aug 18, 2026

Copy link
Copy Markdown
Member

Prevents an agent's higher-sequence durable delivery from being sent after its lower-sequence replay fails. Other agents on the same node continue replaying normally.\n\nAdds conformance coverage for partial provider failure and records the patch-level behavior in the root and engine changelogs.\n\nValidated with the focused delivery suite (53 tests), engine build, and engine lint.

Review in cubic

@coderabbitai

coderabbitai Bot commented Aug 18, 2026

Copy link
Copy Markdown

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 2bab660f-b2e8-45fd-8e3d-a5d299a728ab

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

  • 🔍 Trigger review

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@miyaontherelay

Copy link
Copy Markdown
Contributor

Assessment: still needed, not superseded

Checked whether relay#1579 ("join Relaycast messages into completed replay",
merged 2026-08-20) supersedes this. It doesn't — different subsystem
entirely. #1579 is about SessionClient.replaySession() joining a
Relayhistory turns journal with the Relaycast conversation for
human-readable session replay. This PR is about deliverPendingToNode's
durable-delivery redelivery — preventing a higher-sequence frame from
reaching a provider before a lower-sequence one that failed to send,
which would otherwise put a permanent gap in one agent's ordered message
stream. No code or data overlap between the two.

Current state: draft, but MERGEABLE/CLEAN against main, and every
check is green (Container amd64/arm64, Lint/Build/Test, Rust SDK). It's
just been sitting untouched since 2026-08-18 — 4 days with no reviewer
and no one hitting "ready for review."

Recommend: un-draft and merge, not close. A correctness fix for message
ordering that's green and conflict-free shouldn't go stale waiting on
someone to notice it.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants