Desktop: live threaded reply renders as a top-level channel row until channel switch
Maintainers usually reply within 1 day
Nobody has claimed this yet.
Assessment
- Difficulty
- 3/5
- Estimated time
- 1-2 days
- Newbie friendliness
- 72/100
- Issue type
- Bug
- Clarity
- Mostly clear
- Activity status
- Active
- Tech stack
- typescript
- Domain
- desktop
Research direction
Start with useLiveChannelUpdates in desktop/src/features/channels/useLiveChannelUpdates.ts, then compare its merge path with useChannelSubscription in desktop/src/features/messages/hooks.ts and buildMainTimelineEntries in desktop/src/features/messages/lib/threadPanel.ts. Reproduce the live threaded-reply steps and trace the final render path; done means non-broadcast replies appear only in the thread panel and remain absent from the main timeline after live delivery.
Written by the indexing model from the issue text.
Description
Filed from Leo Wandersleb's account by his Buzz agent (Bob) on his behalf; Fable 5.1 helped authoring this issue.
Summary
When a threaded reply arrives live (relay push) in the channel the user is currently viewing, Desktop renders it twice: once in the thread panel (correct) and once as a full top-level row in the channel timeline, under the NEW divider. The root message's thread summary row updates correctly at the same time (1 reply · last reply just now), so the same event is counted as a reply and rendered as a channel row simultaneously.
The extra channel row disappears as soon as the user switches to another channel and back. Reloading history never reproduces it. It is purely a live-arrival artifact.
Observed on Buzz Desktop v0.5.23 (Linux, .deb, "You're on the latest version"), replying agent on buzz-acp/buzz-cli, but the reply event is an ordinary human-shaped kind 9.
What the relay holds (verified)
Raw REQ for all kinds in the channel since the root shows exactly one content event for the reply:
- kind 9, tags:
["h", <channel>],["e", <root-id>, "", "reply"],["p", <root-author>] - no
broadcasttag - identical tag shape to the human replies in the same thread
So this is not a second event and not an "also send to channel" reply.
Rendering tell
In the duplicated channel row the @Name mention renders as plain text; in the thread-panel copy the same mention renders as a mention chip. The two copies are produced by different render paths.
Where I looked (main at 4ab4f7860, identical at desktop-v0.5.23 / b9392d9d7)
buildMainTimelineEntries(desktop/src/features/messages/lib/threadPanel.ts) keepsparentId == null || isBroadcastReply(...)only. Unchanged since April. By this filter the reply cannot become a main-timeline entry.useChannelSubscriptionappendMessage(desktop/src/features/messages/hooks.ts) routes replies tothreadRepliesKey(channelId, rootId)and returns early unless broadcast. Correct.useLiveChannelUpdates(desktop/src/features/channels/useLiveChannelUpdates.ts, ~L340) merges every live event, including threaded replies, into the flatchannelMessagesKey(channelId)cache viamergeTimelineCacheMessages, with noparentIdcheck. This is the only live path that puts the reply into the channel's message list.- On channel switch,
projectChannelWindowMessagesre-flattens the window store overchannelMessagesKey. The window store never received the reply (the subscription path excluded it), so the row vanishes. That matches the observed behaviour exactly.
I could not find the final hop that turns the cached reply into a rendered top-level row (the memoised filter above should drop it), so the duplicate may come from a component I did not read. The useLiveChannelUpdates merge is the strongest suspect and is inconsistent with the subscription path either way.
Repro
- Open channel A in Desktop and keep it in view.
- From another client, post a top-level message in A, then a threaded reply to it (
buzz messages send --reply-to <root>works). - The reply appears in the thread panel and as a new top-level row in A's timeline under
NEW. - Switch to channel B and back to A: the extra row is gone.
Expected
A threaded reply without broadcast=1 never renders as a main-timeline row, live or from history.
Filed from Leo Wandersleb's account by his Buzz agent (Bob) on his behalf; Fable 5.1 helped authoring this issue.
- Dominant language
- Rust
- Stars
- 35.3k
- Forks
- 4.7k
- Avg merge
- 2d 15h
- Merged PRs (30d)
- 175
Getting set up
- Ships a Dockerfile or Docker Compose file
- Has a pull request template
- Read the contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from block/buzz
-
Difficulty 1/5 Under an hour Newbie friendliness 80/100
Maintainers usually reply within 1 day
-
Mobile: opening an attachment names the file after the link text instead of the imeta filenamePossibly taken @wenhaoone claimed this 2 days ago. Open
Difficulty 2/5 1-3 hours Newbie friendliness 90/100
Maintainers usually reply within 1 day
-
Desktop: private project home channels show no lock icon in the sidebarPossibly taken @Bartok9 claimed this 4 days ago. Open
Difficulty 2/5 1-3 hours Newbie friendliness 86/100
Maintainers usually reply within 1 day
-
buzz-pair: panics on wss:// relays — rustls CryptoProvider never installedPossibly taken @Bartok9 claimed this 5 days ago. Open
Difficulty 2/5 1-3 hours Newbie friendliness 92/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 74/100
Maintainers usually reply within 1 day
Similar issues
-
Difficulty 2/5 1-3 hours Newbie friendliness 72/100
Maintainers usually reply within 1 day
-
[Bug]: Bedrock request metadata forwarding does not work for /embeddingsPossibly taken A pull request linked to this issue is open or already merged. Openbug llm translation
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
Maintainers usually reply within 1 day
-
pytest plugin: a crashed xdist worker aborts the whole session with INTERNALERRORPossibly taken @hazelxue claimed this today. Open
Difficulty 2/5 1-3 hours Newbie friendliness 78/100
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 85/100
Maintainers usually reply within 1 day
-
skillfs: one malformed chat-log line aborts the entire skill-usage analysis (skill_usage_from_chat_logs.py)Possibly taken @zjncs claimed this today. Opencomponent:skillfs
Difficulty 2/5 1-3 hours Newbie friendliness 82/100
agentic-os-org/ANOLISA#6116 · 1 comment ·
Maintainers usually reply within 1 day