feat(kimi-web): render cron fire notices in the web chat - #1426
Conversation
Show scheduled-reminder fires as distinct notice cards in the web chat, both live and after a reload, instead of hiding them. Each card carries a humanized schedule, a dimmed job id, and a collapsible prompt body.
🦋 Changeset detectedLatest commit: 3310d08 The changes in this PR will be included in the next version bump. This PR includes changesets to release 1 package
Not sure what this means? Click here to learn what changesets are. Click here if you're a maintainer who wants to add another changeset to this PR |
commit: |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 98d1eeb523
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
- Give synthesized cron messages a fresh promptId so a fire mid-turn cannot be reconciled into the optimistic user echo and hide it. - Add cron.fired to KNOWN_AGENT_CORE_TYPES so event.-prefixed frames reach the projector and render live too.
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 9c1058106e
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
Slice the first line to the collapse limit with an ellipsis when a single-line prompt exceeds it, so the collapsed state actually truncates and the expand toggle is not a no-op for common one-line reminders.
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 0d75ac0530
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
…r echoes Skip optimistic-echo reconciliation for user messages whose origin is cron_job or cron_missed: their prompt text can coincide with a still- optimistic user message, and the loose content match would otherwise replace the user's turn with the cron notice instead of appending.
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: ae6e4937a6
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
A cron injection steered into an active turn lands inside that turn's message sequence, between a tool use and its result. Treating it as a hard user-turn boundary flushed the pending assistant group, so the next tool result had no group to fold into and the tool rendered without output. Embed such in-turn cron notices as a block inside the assistant group instead, and only render a cron at a turn boundary as its own turn.
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: e76994dee3
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
Embedding whenever a group was pending was too broad: on REST snapshots without prompt ids the whole transcript shares one group, so an idle cron fire merged into the previous assistant answer and its own reply kept going in that group. Embed only while the group has a running tool (a cron sandwiched between a tool use and its result); flush to its own turn otherwise.
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 6c246f94a5
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
The synthesized cron message carried a cron_pr_ promptId that the web client caches into promptIdBySession for Stop/abort. Because it is not a real daemon prompt id, it clobbered the active promptId, so Stop first aborted a nonexistent prompt and only recovered via the error fallback. Omit the promptId; the reducer already skips optimistic-echo reconciliation for cron-origin messages, so it is not needed for de-dup.
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: af5cc8d204
ℹ️ About Codex in GitHub
Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".
# Conflicts: # apps/kimi-web/test/lib-logic.test.ts
# Conflicts: # apps/kimi-web/src/composables/messagesToTurns.ts
1 participant
Related Issue
None — see Problem below.
Problem
The web UI did not render cron (scheduled-reminder) fires at all. The live
cron.firedevent was dropped on the client, and the persisted cron-injected user message was hidden when converting messages to turns, so a scheduled fire was completely invisible in the web chat — users only saw the agent's subsequent reply with no indication it had come from a reminder. The TUI already renders these as a distinct notice.What changed
Render cron fires as their own in-transcript notice card in the web chat, both live (by projecting the
cron.firedevent into a synthetic user message) and after a reload (by recognizing thecron_job/cron_missedmessage origin). Each card shows a humanized schedule (e.g. "Every 5 minutes"), a dimmed job id, and a collapsible prompt body; stale or missed fires use the warning accent. Added zh/en strings and unit tests for the turn conversion, the event projector, and cron humanization.Checklist
gen-changesetsskill, or this PR needs no changeset.gen-docsskill, or this PR needs no doc update.