[VS Code][26.908.40401] Existing chats fail to open with "Codex hit a snag" after update; rollback restores access
Nobody has claimed this yet.
- Dominant language
- Rust
- Stars
- 125k
- Forks
- 19.4k
- PR merge metrics
- PR metrics pending
Description
What version of the IDE extension are you using?
- Affected version: 26.908.40401.
- Working rollback version supplied by the affected user: 26.26.908.31748 (verbatim). The repeated leading
26.may be a transcription typo for26.908.31748; the installed version string has not been independently verified. - Observed on 2026-09-14.
What subscription do you have?
Not provided.
Which IDE are you using?
Visual Studio Code. Exact VS Code version not provided.
What platform is your computer?
The supplied logs contain a Windows-style filesystem error, 系统找不到指定的路径。 (os error 3). Exact Windows version, architecture, and local/WSL/Remote-SSH setup have not been confirmed.
What issue are you seeing?
After updating Codex and restarting VS Code, existing conversations remain visible in the Chats list, but clicking an existing conversation displays "Codex hit a snag". The error screen does not provide any additional diagnostic details.
The affected user confirmed that rolling back from 26.908.40401 to the version recorded above resolved the problem and restored access to the existing conversations.
This is a user-reported upgrade regression with a successful rollback workaround. The root cause has not been established.
What steps can reproduce the bug?
Observed sequence on the affected installation:
- Have existing Codex conversations in the VS Code extension.
- Update the extension to 26.908.40401 and restart VS Code.
- Open the Codex Chats list. Existing conversations are still listed.
- Click an existing conversation.
- The conversation fails to open and the UI displays "Codex hit a snag" without useful error details.
- Roll back to the user-reported version 26.26.908.31748 (see the version-string caveat above).
- Existing-conversation access works again.
This has not been independently reproduced on a clean profile or a second machine. New-conversation behavior was not confirmed.
What is the expected behavior?
Existing conversations should open after an extension update and restart. If opening a conversation fails, the UI should expose an actionable error or diagnostic reference instead of only a generic recovery screen.
Additional information
Selected log excerpts supplied during troubleshooting follow. Conversation and request identifiers have been redacted. Timestamps are copied as supplied; timezone was not captured.
2026-09-14 18:50:45.310 [warning] [IpcClient] Received broadcast but no handler is configured method=thread-stream-following-status-requested
2026-09-14 18:50:45.323 [info] Conversation created conversationId=[redacted] durationMs=16 initializationDurationMs=15 inputsDurationMs=1 mode=default preparationDurationMs=0 requestDurationMs=0 requestPreparationDurationMs=0 threadStartKind=default usedPrewarmedThread=true
2026-09-14 18:50:45.336 [warning] [IpcClient] Received broadcast but no handler is configured method=query-cache-invalidate
2026-09-14 18:50:45.392 [warning] [IpcClient] Received broadcast but no handler is configured method=query-cache-invalidate
2026-09-14 18:50:45.459 [error] Request failed conversationId=none durationMs=2 error={"code":-32603,"message":"系统找不到指定的路径。 (os error 3)"} failureReason=null id=[redacted] method=fs/readDirectory pendingCountAfter=0 priority=interactive queueWaitMs=0 source=filesystem spanId=null timeoutMs=0 traceId=null
2026-09-14 18:50:45.464 [warning] Failed to start visualization turn diff snapshot error={}
The filesystem message translates to "The system cannot find the path specified."
Diagnostic limitation: These excerpts include Conversation created, and the failing fs/readDirectory request has conversationId=none. They do not identify the missing path or directly establish that this request caused the existing-chat error screen. No webview exception stack was captured. The IPC warnings and snapshot warning are included as surrounding evidence, not asserted causes.
This report was prepared with Codex assistance from the affected user's observations and supplied logs, and submitted at the user's request. No conversation contents, credentials, or project paths are included.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by reproducing the failure in the VS Code Chats list on version 26.908.40401 and inspect the supplied IPC and filesystem logs, especially the fs/readDirectory request. Done means existing conversations open after the update, or the failure exposes an actionable diagnostic; the missing path and root cause still need to be established.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- rust, vscode
- Domain
- developer-experience, frontend
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Needs clarification
- Newbie friendliness
- 35/100