Hardware: Mac17,6, Apple M5 Max, 18 CPU cores, 36 GiB RAM, arm64
Task title: Dev mimarisi inceleme promptu
Thread ID: 01a0432b-0b4f-78a1-b80e-e5d384a07e58
Project branch: dev
Problem:
Some older messages remain visible in the task, but several newer completed
user/assistant turns are absent from the task history UI after reopening the
same long-running task.
Expected:
All completed user and assistant turns should appear chronologically when
the task is reopened.
Actual:
The task view displays an older section of the conversation but omits five
later completed UI-related turns.
Evidence:
The missing user messages, their message IDs, turn IDs, and corresponding
task_complete events still exist in the local rollout data. This indicates
that the turns were persisted and suggests a task history indexing or
rendering issue rather than data loss.
A compaction event occurred shortly after the first affected prompt. This is
included as a time correlation only; I cannot confirm that compaction caused
the issue.
Reproduction:
Open a long-running Codex desktop task.
Submit several consecutive follow-up prompts and allow them to complete.
Navigate away from or close the task.
Reopen the same task.
Observe that older turns remain visible while some later completed turns
are absent.
I’ve attached:
A sanitized persisted-turn timeline
A sanitized rollout event excerpt
The screenshots exclude credentials, absolute local paths, reasoning records,
tool traffic, hardware identifiers, serial numbers, UUIDs, and account
metadata.
I also have a sanitized JSON evidence report and a ZIP archive containing the
supporting files. I can share them privately with OpenAI staff if requested.
You are welcome to discuss Codex issues here. However, the official place to report and track Codex bugs is the OpenAI Codex GitHub issue tracker.
I asked ChatGPT to search for the closest related issue, and it identified:
If that issue describes the same problem you are experiencing, consider adding a reaction to the GitHub issue itself, rather than only reacting to this forum topic. GitHub reactions can help the maintainers gauge how many users are affected.
If it does not match your problem closely enough, search the existing Codex issues for a better match. If you cannot find one, consider opening a new issue and include as much relevant information as possible, such as your Codex version, operating system, environment, reproduction steps, expected behavior, actual behavior, and any relevant logs or error messages.
reaction to the GitHub issue itself, rather than only reacting to this forum topic. GitHub reactions can help the maintainers gauge how many users are affected.