import of filenames with special characters fails

[fschaeffler]

In some of our Mercurial repositories, we have filenames with special characters. This brings some issues as they're not properly handled. For the test case here, I used the German ö (o with two dots at the top).

Note: ö has the octal value of 366 in ISO/IEC 8859.

Mercurial Repo:

• src/file.txt
• src/file-with-special-character-ö.txt

Created Git Repo:

• src/file.txt
• 'src/file-with-special-character-'$'\366''.txt'

Import Verification:

[...]
Only in repo-mercurial/src/file-with-special-character-ö.txt
Only in repo-git/src/file-with-special-character-�.txt

[fschaeffler]

What I found out so far is that the original Mercurial repository files were created not with utf8 but with latin1.

[kilork]

I have special test repository which uses cp1251 filename encoding, it was created in Windows. Actually, the file path handling logic is one of most untransparent in this project, I would be glad for any suggestions on how to improve this.

To fix this error we need to have minimal reproduction example, something with maybe just one commit which should be enough.