Search code examples
gitterminalcharacter-encodingcommand-line-interfacelocale

'git log' output encoding issues in Windows 10 CLI terminal

Problem

How can I make git log command output properly displayed in the Windows CLI terminal?

Example

Git commands sequence leading to the problem

As you can see, I can type diacritical characters properly, but on git log, the output is somehow escaped. According to the UTF-8 encoding table, the codes between angle brackets (< and >) from the output correspond to the previously typed git config parameters.

I have tried to set LESSCHARSET environment variable to utf-8 as suggested in one of the answers for similar issue, but then the output is garbled:

'git log' output after setting LESSCHARSET=utf8

I know .git/config is encoded properly with UTF-8 as it's handled by gitk as expected.

Proper gitk output

Here is the locale command output if necessary:

LANG=
LC_CTYPE="C.UTF-8"
LC_NUMERIC="C.UTF-8"
LC_TIME="C.UTF-8"
LC_COLLATE="C.UTF-8"
LC_MONETARY="C.UTF-8"
LC_MESSAGES="C.UTF-8"
LC_ALL=

The output is the same also in pure Git Bash:

Enter image description here

so I believe the problem is shell-independent and relates to Git or its configuration itself.

Solution

  • Okay, I experimented a bit and found out that Windows Git commands actually need UNIX variables like LC_ALL in order to display Polish (or other UTF-8 characters) correctly. Just try this command:

    set LC_ALL=C.UTF-8

    Then enjoy the result. Here is what happened on my console (font "Consolas", no chcp necessary):

    Windows console CMD

    Update:

    • Well, in order for Windows commands like type (display file on console) to work correctly, you do need chcp 65001.
    • And if you prefer commands from Git Bash like cat you profit from the aforementioned set LC_ALL=C.UTF-8.

    Windows console CMD, part 2

    Update 2: How to make the changes permanent

    As user mono blaine said, create an environment variable LC_ALL and assign it the value C.UTF-8, either globally or for your own user profile only (sorry for the German screenshot):

    Create environment variable

    Next time you open a command processor console (cmd.exe) you should see the variable value when issuing the command echo %LC_ALL%. In PowerShell you should see it when issuing $env:LC_ALL.

    The simplest way to make the UTF-8 code page permanent ist to open regeedit and add a new value named Autorun of type string to section HKEY_LOCAL_MACHINE\SOFTWARE\Microsoft\Command Processor and assign it the value chcp 65001.

    Registration editor

    Henceforth, this command will be executed each time you open a new cmd.exe console. You even see its output in the new window: "Aktive Codepage: 65001." (or similar in your respective language).

    Oh, by the way: In order to display a UTF-8 encoded file correctly in PowerShell you can use Get-Content -encoding UTF8 file.txt or cat -encoding UTF8 file.txt (cat being an alias for Get-Content in PowerShell).