'utf-8' codec can't decode byte... with accented username

[Tomalak]

I'm working on Windows 10, Mercurial 6.9.1, vscode-hg 1.9.0.

The .hgrc file in my home directory is ANSI-encoded ("ANSI" meaning "the configured Windows default encoding", in my case that's Windows-1252).

My Mercuial username contains an accented character.

For some reason, vscode-hg seems to load the .hgrc file in the UTF-8 encoding, or possibly run all of Mercurial in an UTF-8 context, I'm not sure what's going on there.

As a result, committing from the left VSCode sidebar fails - it aborts and rolls back with a Python decoding error ("'utf-8' codec can't decode byte... invalid start byte!"), pointing to my username.

Committing manually with hg commit -m "..." works, both from the VSCode-integrated console and from outside of VSCode.

Changing my .hgrc to UTF-8 encoding fixes the issue with committing from the VSCode sidebar, but breaks all other methods (command line, TortoiseHg).

How to reproduce

• change [ui]/username in your .hgrc so that it contains an accented character
• save the file in a single-byte encoding, such as latin-1 or Windows-1252
• (re)start VSCode
• try to commit a change

Unfortunately, I can't easily change the default encoding of my machine without causing other issues, so for the time being I'm stuck with Windows-1252.

Is here a way for me to change how Mercurial is run by vscode-hg in order to work around this issue?