Gitbash start kimi failed

[LaurenceLong]

What version of Kimi Code CLI is running?

1.22.0

Which open platform/subscription were you using?

kimi code

Which model were you using?

No response

What platform is your computer?

windows

What issue are you seeing?

$ kimi
+-----------------------------------------------------------------------------+
| |
| 'gbk' codec can't encode character '\u2590' in position 0: illegal multibyte sequence
See logs: C:\Users\Laurence.kimi\logs\kimi.log

What steps can reproduce the bug?

Launch gitbash
start kimi

What is the expected behavior?

No response

Additional information

No response

[shji47]

you can use this code in git bash to get sys.stdout.encoding for python
python python -c "import sys; print('stdout:', sys.stdout.encoding); print('stderr:', sys.stderr.encoding)"
the output in my git bash is

('stdout:', None) 
('stderr:', None)

Therefore, git bash use the system default character encoding, which is usually 963(GBK), so it reports this error. If you want to force python to use UTF-8 encoding for stdout, you can add the following code in /ui/shell/console.py

import sys
import io

sys.stdout = io.TextIOWrapper(sys.stdout.buffer, encoding='utf-8', errors='replace')
sys.stderr = io.TextIOWrapper(sys.stderr.buffer, encoding='utf-8', errors='replace')

but i recommand you use cmd or powershell to run kimi

[Br1an67]

This is the same root cause as #773 — the terminal encoding is not UTF-8 (GBK in Git Bash instead of ASCII in cmd.exe, but the effect is identical).

Our PR #1497 already fixes this by reconfiguring sys.stdout/stderr to UTF-8 at startup on Windows, regardless of the terminal's default encoding. Once #1497 is merged, this issue should also be resolved.