Encoding problems with non-ASCII characters with Git 2.44.0

[inosik]

• I was not able to find an open or closed issue matching what I'm seeing

There's something wrong w.r.t. encoding in Git Bash. Non-ASCII characters (German Umlauts in my case) appear garbled in the terminal output.

Git 2.43.0:

$ git add -p
diff --git a/test.txt b/test.txt
index a3ea8e6..0e54281 100644
--- a/test.txt
+++ b/test.txt
@@ -1 +1,2 @@
 äöü
+change
(1/1) Stage this hunk [y,n,q,a,d,e,?]?

Git 2.44.0:

$ git add -p
diff --git a/test.txt b/test.txt
index a3ea8e6..0e54281 100644
--- a/test.txt
+++ b/test.txt
@@ -1 +1,2 @@
 ├ñ├Â├╝
+change
(1/1) Stage this hunk [y,n,q,a,d,e,?]?

Setup

• Which version of Git for Windows are you using? Is it 32-bit or 64-bit?

$ git --version --build-options

git version 2.44.0.windows.1
cpu: x86_64
built from commit: ad0bbfffa543db6979717be96df630d3e5741331
sizeof-long: 4
sizeof-size_t: 8
shell-path: /bin/sh
feature: fsmonitor--daemon

• Which version of Windows are you running? Vista, 7, 8, 10? Is it 32-bit or 64-bit?

$ cmd.exe /c ver

Microsoft Windows [Version 10.0.19045.3930]

• What options did you set as part of the installation? Or did you choose the
  defaults?

I installed Git using Scoop, which just extracts PortableGit-2.44.0-64-bit.7z.exe.

• Any other interesting things about your environment that might be related
  to the issue you're seeing?

Nothing I can think of.

Details

• Which terminal/shell are you running Git from? e.g Bash/CMD/PowerShell/other

Bash inside of the Windows Terminal App. But this is also reproducible using CMD, PowerShell and git-bash.exe.

• What commands did you run to trigger this issue? If you can provide a
  Minimal, Complete, and Verifiable example
  this will help us understand the issue.

git add -p

• What did you expect to occur after running these commands?

Non-ASCII characters should be displayed properly .

• What actually happened instead?

Non-ASCII characters are garbled.

• If the problem was occurring with a specific repository, can you provide the
  URL to that repository to help us with testing?

No specific repository, but here's a Gist with the test file.

[dscho]

I can reproduce this, and work around it by using MSYS=disable_pcon:

$ MSYS=disable_pcon git add -p
diff --git a/test.txt b/test.txt
index a3ea8e6..0e54281 100644
--- a/test.txt
+++ b/test.txt
@@ -1 +1,2 @@
 äöü
+change
(1/1) Stage this hunk [y,n,q,a,d,e,?]?

$ MSYS=enable_pcon git add -p
diff --git a/test.txt b/test.txt
index a3ea8e6..0e54281 100644
--- a/test.txt
+++ b/test.txt
@@ -1 +1,2 @@
 äöü
+change
(1/1) Stage this hunk [y,n,q,a,d,e,?]?

I can also work around it by calling chcp 65001:

$ cmd //c chcp 65001
Active code page: 65001

me@work MINGW64 ~/repros/umlauts-4851 (main)
$ MSYS=enable_pcon git add -p
diff --git a/test.txt b/test.txt
index a3ea8e6..0e54281 100644
--- a/test.txt
+++ b/test.txt
@@ -1 +1,2 @@
 äöü
+change
(1/1) Stage this hunk [y,n,q,a,d,e,?]?

I guess that we'll want to always change the code page to 65001 in the enable_pcon mode?

[Kirill]

I found this bug too.
Folder in Explorer

This screenshot I see in Git bash, Terminal, FAR Manager

I downgrate to 2.21.0 and it's work correctly.

I update to 2.44.0, but bug lost (I try reproduce in newer version after downgrade, but I see again correct filename in last version)

[inosik]

@dscho MSYS=disable_pcon doesn't seem to do anything in my case:

$ MSYS=disable_pcon git add -p
diff --git a/file.txt b/file.txt
index a3ea8e6..d88e171 100644
--- a/file.txt
+++ b/file.txt
@@ -1 +1,2 @@
 ├ñ├Â├╝
+Ôé¼
(1/1) Stage this hunk [y,n,q,a,d,e,?]?

chcp 65001 does work.

[MrMarvel]

Same here utf-8 filenames wrong encoded in git 2.44.0.

It was working in git 2.42.0

[zzfhadr]

same +1

but if use git bash it will work... but i don't want to use it.. not convinient for copy cand paste

[stenretni]

I added chcp.com 65001 > /dev/null to my .config/git/git-prompt.sh as a workaround. So far so good.

[drq001out]

In Chinese, VS Code & CMD, same problem!!!

[inosik]

Same behavior with Git 2.45.0.

[dscho]

@inosik I have to be honest: due to shifts in priorities at my day job, I am stretched a little too thin to work on this. Maybe you can? It would require a little bit of C++ knowledge (not C# or F#) to work on the MSYS2 runtime, which is a bit tricky to navigate, I'd try my best to assist with guidance.

[inosik]

I never did any C++ coding, but I could give it a shot. Can you tell me where I should start looking?

[dscho]

@inosik the first thing would not even be C++, but to verify that dropping the usr\bin\msys-2.0.dll from v2.43.0 into, say, a Portable Git v2.44.0 "fixes" the problem.

[inosik]

verify that dropping the usr\bin\msys-2.0.dll from v2.43.0 into, say, a Portable Git v2.44.0 "fixes" the problem.

Moving msys-2.0.dll around didn't fix the problem, but copying mingw64/bin/git.exe from Git 2.43.0 to 2.45.0 did.

[dscho]

@inosik thank you for testing! This led me into the right direction. It still took over a day to bisect, but at least now I have a fix in #4968. Could you please verify that this works for you? You should be able to extract the git.exe from the PR build's artifacts, once the build finishes.