7za.exe 9.20 cannot create archive with non english letters filename (utf-8)

Question

I'm struggling to have 7za.exe create an archive including non latin characters. The encoding is utf-8, characters are cyrillic. I have a folder with the 4 files:

7za.exe privet.txt Кириллица.txt test.py

where the content of test.py is the following:

#!/usr/bin/env python
# -*- coding: utf-8 -*-

SOURCE_FILE = "Кириллица.txt"
DEST_ARCHIVE = "Кириллица.7z"

import subprocess

subprocess.call('7za a -bd -y privet.7z privet.txt', shell=True)

cmd_str = '7za a -bd -y %s %s' % (DEST_ARCHIVE, SOURCE_FILE)
subprocess.call(cmd_str, shell=True)

While I can create privet.7z from privet.txt, I cannot create Кириллица.7z from Кириллица.txt (an empty archive named Кириллица.7z is instead generated).

The output from 7za.exe is:

C:\BEPPE\STAMPARE\TEST_7za_cyrillic>python test.py
7-Zip (A) 9.20  Copyright (c) 1999-2010 Igor Pavlov  2010-11-18
Scanning
Creating archive privet.7z
Everything is Ok
7-Zip (A) 9.20  Copyright (c) 1999-2010 Igor Pavlov  2010-11-18
Scanning
ÐsиÑ_иллиÑ┼а.txt:  WARNING: Impossibile trovare il file specificato.
Creating archive ÐsиÑ_иллиÑ┼а.7z
WARNINGS for files:
ÐsиÑ_иллиÑ┼а.txt : Impossibile trovare il file specificato.
----------------
WARNING: Cannot find 1 file

Can anybody help me with this? I also tried from a batch script test.bat with the following content:

7za.exe a -bd -y privet.7z privet.txt
7za.exe a -bd -y Кириллица.7z Кириллица.txt

but the result was the same.

Answer 1

Handling UTF-8 (Unicode, really) is not easy, it wouldn't surprise me in the least that handling of UTF-8 (or even non-ASCII) filenames just hasn't been done. The messed up name of the file created hints at serious problems in the handling of filenames (a compress/uncompress program is at the mercy of possibly malicious contents). I'd handle that as toxic materials, looking for a more robust compressing solution.

Answer 2

This works for me:

1. Download Notepad++ (even the <900KB minimalist package will do)

2. Go to the Settings / Preferences / New Document tab and set the default Encoding to UTF-8 without BOM:

   

3. Open a new document and type the following command in it:

   cmd /u /c "chcp 65001 >nul && 7za.exe a -bd -y Кириллица.7z Кириллица.txt"
   

4. Ensure that Notepad++'s status bar says ANSI as UTF-8 and save the document as a .BAT file:

   

5. Open a Command Prompt and run the batch file

Also see here if you want the Command Prompt window to display Unicode characters.