edit/unsend targeting a self-chat's echo row is a silent no-op (only the sent row syncs to the iPhone)

Environment: macOS 27.0.1, imsg 0.16.0.

What we found (building a bridge that sends into a self-chat): a self-chat stores each sent message twice — a sent row (is_from_me=1) and a received echo row (is_from_me=0), with distinct guids (chat↔message linkage via ck_chat_id; our self-chat carried ~39 such pairs).

imsg edit / imsg unsend targeting the sent row works end-to-end: the edit delivers to the iPhone, the unsend removes the iPhone's copy.
Targeting the echo row is a silent no-op: exit 0, no error, nothing happens on the iPhone.

Anyone scripting against guids collected from history/watch (which emits both rows) can easily hold the echo guid and wonder why their edit/unsend did nothing.

Suggestion: auto-resolve to the canonical sent row when the target is a self-chat pair member, or at least print a hint/error naming the twin guid. A docs note is a fine fallback.

Codex review: this still needs some work.

Summary
This issue describes a concrete released failure, and current main still accepts edit/unsend targets without checking sent-message eligibility; no fixing PR owns the work.

Reproducibility: Source proves unchecked GUID dispatch and queued acknowledgments, supported by the reporter's release-level comparison; current-main iPhone behavior and native eligibility still need verification, and no live commands or tests were run.

Next step
Verify native sent/echo eligibility and establish failing edit/unsend regressions before adding shared received-target rejection.

Review details

Best possible solution:

Reject ineligible received targets with guidance to select the sent GUID, and retarget automatically only if a verified native twin relationship exists.

Do we have a high-confidence way to reproduce the issue?

Source proves unchecked GUID dispatch and queued acknowledgments, supported by the reporter's release-level comparison; current-main iPhone behavior and native eligibility still need verification, and no live commands or tests were run.

Is this the best way to solve the issue?

A shared helper eligibility check would cover CLI and RPC within the existing sent-message contract; matching twins by text or nearby timestamps could select the wrong message.

AGENTS.md: found and applied where relevant.

Remaining risk / open question:

The reported iPhone no-op has not been independently reproduced on current main.
A reliable native sent-message eligibility signal and an authoritative relationship between distinct echo and sent GUIDs remain unverified.

Codex review notes: model internal, reasoning medium; reviewed against d3c706e6c38c.

Label changes

Label changes:

No label changes.

Label justifications:

P2: The advanced self-chat mutation workflow has limited scope and a reported workaround of choosing the sent GUID.
impact:ux-friction: Successful-looking acknowledgments provide no guidance when an echo target leaves the intended message unchanged.

Evidence reviewed

Acceptance criteria:

Establish failing native edit and unsend regressions for received targets with passing sent-target controls.
make test-helper
make lint
make test
On configured macOS, verify CLI and RPC report actionable echo-target errors and sent-target mutations reach the iPhone.

What I checked:

Release-level report and discussion: The reporter identifies imsg 0.16.0 on macOS 27.0.1 and compares distinct sent and echo GUIDs: only sent targets affect the iPhone. Both discussion comments are automated, providing no independent reproduction; the fetched timeline contains no linked fixing PR.
Current default branch: GitHub identifies main as the default branch, and its revision matches the clean review checkout. (d3c706e6c38c)
Unchecked mutation dispatch: Edit resolves the supplied GUID and invokes a native selector; unsend resolves its part and invokes retractMessagePart:. Neither validates direction or resolves a sent twin, and both return queued true after a nonthrowing invocation. (Sources/IMsgHelper/MessageMutation.inc:33, d3c706e6c38c)
Lookup preserves requested identity: Lookup loads the requested GUID or searches native objects for that exact identity; wrapper normalization and part selection contain no sent/echo pairing mechanism. (Sources/IMsgHelper/MessageLookup.inc:51, d3c706e6c38c)
Sent-message contract and shared entry points: CLI help describes editing and retracting sent messages, but forwards the supplied GUID and prints queued; RPC reaches the same helper and returns ok true after invocation. (Sources/imsg/Commands/BridgeMessagingCommands.swift:66, d3c706e6c38c)
Read surfaces and documentation: History queries include both directions, while mutation examples use generic message GUIDs without explaining self-chat echo eligibility. (Sources/IMsgCore/MessageStore+Queries.swift:26, d3c706e6c38c)

Likely related people:

Peter Steinberger: Raw commit cb58dec adds Sources/IMsgHelper/MessageMutation.inc:33 relative to its recorded parents. This identifies author metadata, not feature responsibility or a PR merger. (role: source-line author; confidence: high; commits: cb58dec17663; files: Sources/IMsgHelper/MessageMutation.inc)

How this review workflow works

ClawSweeper keeps one durable marker-backed review comment per issue or PR.
Re-runs edit this comment so the latest verdict, findings, and automation markers stay together instead of adding duplicate bot comments.
A fresh review can be triggered by eligible @clawsweeper re-review comments, exact-item GitHub events, scheduled/background review runs, or manual workflow dispatch.
PR/issue authors and users with repository write access can comment @clawsweeper re-review or @clawsweeper re-run on an open PR or issue to request a fresh review only.
Maintainers can also comment @clawsweeper review to request a fresh review only.
Fresh-review commands do not start repair, autofix, rebase, CI repair, or automerge; those execution lanes are retired.
Maintainers can comment @clawsweeper explain to ask for more context, or @clawsweeper stop to stop active automation.

Reviewed October 19, 2026, 1:59 PM ET / 17:59 UTC.

The fix rejects a received self-chat echo before invoking any native mutation API and tells callers to choose the sent copy (is_from_me=true). The synthetic regression reproduced 11 failed assertions before the fix and passes afterward across all three edit selectors and unsend. Native macOS 27 inspection confirmed the BOOL ownership selector exists on all relevant message types. No live messages were modified. Landed in PR #356.

Commit: e01f113
CI: https://github.com/openclaw/imsg/actions/runs/39309551551

fix: harden message mutations and contact lookups

Navigation Menu

Repository navigation

edit/unsend targeting a self-chat's echo row is a silent no-op (only the sent row syncs to the iPhone) #350

68 remaining items

Metadata

Assignees

Labels

Type

Fields

Projects

Milestone

Relationships

Development

edit/unsend targeting a self-chat's echo row is a silent no-op (only the sent row syncs to the iPhone) #350

Description

Activity

clawsweeper commented on Oct 16, 2026

68 remaining items

steipete commented on Oct 19, 2026

Metadata

Metadata

Assignees

Labels

Type

Fields

Projects

Milestone

Relationships

Development

Issue actions