openai / openai/codex

Codex App browser process crashes during browser-use/webview lifecycle — reproduced twice with ResizeObserver storm

Open
#46,472 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

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

Description

What version of the Codex App are you using (From “About Codex” dialog)?

26.915.31029

What subscription do you have?

Pro

What platform is your computer?

Microsoft Windows NT 10.0.26200.0 x64

What issue are you seeing?

I submitted this through Codex /feedback before opening this issue.

Feedback ID:
01a04b89-b43f-72a0-a569-45e2e96477a9

The feedback dialog confirmed receipt but also reported:

"Your feedback was received, but some diagnostic files could not be uploaded"

Because of that warning, I preserved the evidence locally rather than assuming all diagnostics reached OpenAI.

Preserved evidence includes:

  • Complete Codex desktop logs surrounding both incidents
  • Both Crashpad sidecar JSON artifacts
  • Relevant Windows Event Logs / EVTX evidence
  • Windows AppModel Runtime lifecycle evidence
  • Crashpad directory inventories
  • Screenshots of the browser-use state after restart
  • Reconstructed incident timelines
  • SHA-256 manifests for the evidence
  • Original incident archive plus a separately hashed supplemental archive

No matching September .dmp has yet been located. An older Codex Crashpad .dmp from April 2026 exists in a separate Codex Crashpad tree, so the absence of a September dump is recorded only as an observation; I am not claiming that no dump was ever generated.

I do not believe the evidence currently establishes that the tested POS application itself is defective or caused the crashes. The repeated pattern appears more relevant to investigation of Codex browser-use / Electron / embedded webview lifecycle behavior.

I can provide sanitized portions of the preserved evidence if maintainers need them.

P.S. If finding two same-day Codex browser-process crashes, correlating the failures down to the second, locating the Crashpad evidence after Windows/WER reported nothing, and preserving the diagnostics that /feedback couldn't fully upload earns store credit, I'll happily accept one month of the $200 Pro plan. 😂

Worth a shot.

What steps can reproduce the bug?

Running codex on my current POS software under GPT 5.6 Extra High

What is the expected behavior?

No crashes would be best

Codex-Remote-2026-09-18-Supplement-v2.zip
Codex-Remote-2026-09-18-Supplement-v2.zip.sha256.txt

Additional information

Potentially related existing reports:

  • #43347 — Closing the last in-app Browser Use tab crashes the desktop app
  • #42006 — Windows desktop app crashes when in-app Browser route is torn down
  • #38308 — Multiple embedded browser tabs accumulate WebView2 processes and crash Codex when a task finishes

“Codex About reports version 26.915.31029 (released Sep 17, 2026). Windows identifies the installed MSIX package as OpenAI.Codex_26.915.3509.0_x64__2p2nqsd0c76g0.”

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 by comparing this report with related issues #43347, #42006, and #38308, then review the supplied Crashpad, Windows Event Log, and lifecycle evidence. Confirm a reproducible browser-use/webview crash on Windows and establish a clear lifecycle cause and regression test or verification path as the definition of done.

Written by the indexing model from the issue text.

Assessment

Tech stack
electron
Domain
desktop, operating-systems
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Needs clarification
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.