[BUG] In the inline terminal, the output of any command exhibits character corruption.

[IM-HCC]

Problem (one or two sentences)

In the inline terminal, the output of any command exhibits character corruption. I have identified the cause of the problem: my local character encoding is GBK, but the decoding method used by RooCode is UTF-8, which is responsible for the occurrence of character corruption.

Under the section "Custom Instructions for All Modes" in roocode, it is specified that character corruption will no longer occur after executing the command "chcp 65001".

The above information was generated by a translation software. If there is anything unclear, please feel free to let me know. My intention is truly to resolve this issue. Finally, I would like to express my gratitude to all the developers who have contributed to this project.

Context (who is affected and when)

Reproduction steps

Adjust the local Windows system encoding to GBK.
In rocode, enable the Agent to use the chcp command to avoid character corruption.

Expected result

chcp �����ҳ: 936

Actual result

Active code page:936

Variations tried (optional)

No response

App Version

3.40.0

API Provider (optional)

None

Model Used (optional)

N/A

Roo Code Task Links (optional)

No response

Relevant logs or errors (optional)

[roomote]

I've analyzed this issue and here's my proposed implementation plan:

Problem Analysis:
The character corruption occurs because:

1. Windows systems with non-UTF-8 encodings (like GBK/code page 936) output terminal text in their native encoding
2. Roo Code decodes all terminal output as UTF-8 (via TextDecoder() in ExecaTerminalProcess.ts and VSCode terminal stream handling)
3. The mismatch between GBK-encoded bytes and UTF-8 decoding causes character corruption

Proposed Solution:
Force UTF-8 encoding on Windows terminals to ensure consistent UTF-8 output. This aligns with the workaround you discovered (chcp 65001).

Implementation Details:

1. Modify Terminal.getEnv() in src/integrations/terminal/Terminal.ts:

   • On Windows, add environment variables to force UTF-8 encoding:
     • PYTHONIOENCODING=utf-8 (for Python scripts)
     • Set console code page environment for PowerShell/cmd

2. For VSCode terminals on Windows:

   • Since chcp 65001 works as a workaround, we can automatically configure the terminal to use UTF-8 by setting appropriate environment variables or terminal profile settings

3. For Execa terminals (ExecaTerminalProcess.ts):

   • Already sets LANG=en_US.UTF-8 and LC_ALL=en_US.UTF-8
   • On Windows, may need additional handling since Windows ignores LANG/LC_ALL

4. Alternative approach:

   • Add a user setting to configure terminal encoding if automatic detection is not reliable

This approach will ensure consistent UTF-8 encoding across all terminal types on Windows, eliminating the character corruption issue without requiring users to manually run chcp 65001.

Please mention @roomote in a comment saying that you approve this plan, and I'll proceed with the implementation.

[IM-HCC]

I think it would be appropriate to add a user setting to configure the terminal encoding. I hope that adding this feature will not have any negative impact on the overall functionality of the software.