The server echoes a sent message part by part. The reducer replaced an
optimistic part with its server echo only while the FIRST part of the
message was still optimistic, so once the text echo had landed the file
echo no longer qualified and was appended instead: an attached image
showed twice until the next page fetch rewrote the parts.
Any remaining optimistic part of the same type is now a replacement
candidate. The scan only runs when a part with a new id arrives, which is
once per part during assistant streaming.
Testing: reducer test extended past the first echo to the file echo; ui
type-check and lint; verified the event order against a live OpenCode
(message, text part, file part within ~70ms of the accepted request).
The server echo of an optimistic user message carries different part
ids, so the reducer dropped the optimistic part and appended the server
one at the end, and MessageBody keyed user parts by part id. The key
change remounted the text subtree (blank frame, markdown re-parse,
truncation state reset) and the append reordered text against file
parts.
Replace the optimistic part in place and key user parts positionally.
Keeps optimistic prompt state in the session directory
Routes live assistant part updates using upstream event payloads
Adds regressions for startup session switch races
Canonicalize session message/part materialization across load, prefetch, reconnect, and recovery paths so OpenChamber restores session snapshots through one consistent merge flow.
Preserve live assistant streaming text when stale or delayed snapshots arrive, while still replacing optimistic user parts with confirmed server snapshots to avoid duplicated user messages.
Narrow recovery triggers to explicit incomplete snapshot signals instead of broad session-event fallbacks, reducing unnecessary session refetches during active streaming.
Keep turn windowing aligned with parented assistant replies and add regression coverage for materialization gaps, stale snapshot protection, optimistic user replacement, reconnect recovery, and turn grouping.
Bohdan Triapitsyn