openai / openai/codex

Review panel duplicates a changed file and misreports change totals

Open
#45,959 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

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

Description

Summary

The Codex Review panel can show the same changed file multiple times and present an inconsistent change count for a single turn.

Observed behavior

During one turn that edited three files, the Review panel showed:

  • Last Turn: +233 -273
  • docs/bugs/003-ai-checkout-link-inline/fix-attempt-006.md: +34 -0
  • docs/bugs/003-ai-checkout-link-inline/status.md: +5 -3
  • EmSite/assets/js/checkout.js: +0 -270
  • EmSite/assets/js/checkout.js: +192 -0
  • EmSite/assets/js/checkout.js: +2 -0

The same checkout.js file appears three times. The review summary is therefore presented as separate file entries rather than one aggregated file-level change, and the displayed change totals are misleading when compared with the actual set of changed files.

Expected behavior

  • Show each changed file once in the Review panel.
  • Aggregate additions and deletions for that file.
  • If separate edit operations or patches need to be shown, nest them under one file entry instead of duplicating the file.

Steps to reproduce

  1. Start a Codex task in a Git repository.
  2. Apply multiple edits or patches to the same file during one turn, along with edits to other files.
  3. Open the Review panel and select the last-turn changes.
  4. Observe that the repeatedly edited file appears as multiple file entries and that the change summary is not presented consistently as an aggregated file-level result.

Environment

  • Codex Desktop app Version 26.908.70816
  • Windows
Image

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

Start at the Review panel's last-turn changes entry point and trace how repeated edits are converted into file entries and summary totals. Reproduce the three-edit case from the issue, then verify that each changed file appears once with aggregated additions and deletions, while any separate patches remain nested under that file.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
desktop
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
68/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.