openai / openai/codex

Windows Desktop: pending follow-ups stall after completion; manual resume processes only the last of three

Open
#42,890 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

agent app bug windows-os
Dominant language
Rust
Stars
125k
Forks
19.4k
PR merge metrics
PR metrics pending

Description

Environment
  • Codex desktop app on Windows.
  • Observed September 4, 2026.
  • Exact desktop app version not captured.
What happened

Pending follow-up messages do not automatically start a new turn after the current assistant turn finishes.

In the reported three-message case, none of the three pending messages fired after completion. After manually resuming/steering, only the last message was processed. The assistant answered that message and ended its turn again, without processing the earlier two.

The pending cards expose a Steer control. This report describes the visible behavior; the internal distinction between queued and steered messages has not been verified.

Steps from the observed sequence
  1. While the assistant is working, add several follow-up messages so they appear as pending cards.
  2. Let the current assistant turn finish normally.
  3. Observe that the pending messages remain and no next turn starts.
  4. Manually resume/steer a pending message.
  5. Observe that only the last of three messages is processed, followed by another completed assistant turn; the earlier messages are not processed.
Expected behavior

Pending follow-ups should dispatch as intended after completion, without requiring manual intervention. Multiple accepted messages should be handled in the intended order, or the UI should clearly indicate any replacement or rejection.

Actual behavior

The queue stalls at the turn boundary. Manual intervention processes only the last message in the reported three-message case. This report does not establish that the earlier messages were permanently deleted.

Evidence

The reporter captured screenshots showing pending cards both while the assistant was working and beneath a completed assistant response. A test message eventually arrived after manual intervention; that does not demonstrate automatic queue recovery. The three-message/last-only behavior is the reporter's observation, not an independently instrumented reproduction.

Related reports
  • #34026 — Windows desktop follow-ups trapped in the queue after completion. A stale “thinking” state has not been confirmed in this case.
  • #23100 — Queued messages do not dispatch until a task is selected. Focus dependence has not been tested here.

Root cause is unknown; no backend dispatch trace has been inspected.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No source files, tests, or entry points are named. Start by reproducing the Windows sequence with three pending follow-ups, then compare the behavior with related reports #34026 and #23100. Done means pending messages dispatch automatically after completion in the intended order, or the UI clearly indicates replacement or rejection.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
desktop, operating-systems
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
48/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.