#1235 Warn before creating ZIP with non-ASCII characters in filenames

[Jaan Vajakas]

I think 7-Zip should show a warning dialog before creating a ZIP file, if the ZIP file to be created would contain non-ASCII characters in the names of the files it contains.

The main reason why people use ZIP instead of more advanced formats like 7z is portability and easiness for less advanced computer users: ZIP files can be opened on more platforms without installing additional software. If the user needs a special software configuration to extract the files correctly (as is the case when non-ASCII filenames are used, unless the receivers of the file are known to use the same codepage and all the filenames can be encoded in that codepage), this purpose is defeated.

More details:

It seems that currently (as of 7-zip 9.38 beta), when creating ZIP files with 7-zip in Windows 7, 7-zip handles filename encoding as follows: if all the characters in the name of a file to be compressed can be encoded in the system's encoding (more precisely, the OEM (MS-DOS) codepage, e. g. IBM859), then the filename is encoded using that codepage; otherwise the filename is encoded using UTF-8. For example, when I create a ZIP file with three files é.txt, β.txt and βé.txt, then 7-p encodes the first filename in the ZIP using the OEM codepage and the other two using UTF-8.

When I extract the same archive using Windows Explorer on the same Windows 7 system where I created the ZIP file with 7-Zip, only the first file é.txt is extracted with the correct name. If I extract it with 7-Zip on the same computer, all three files are OK. If the ZIP is opened on a computer using a codepage where the code of é means something else, I guess both 7-Zip and Windows Explorer would decode the first filename incorrectly.

[Jaan Vajakas]

Typo in the bug description: "7-p encodes" -> should be "7-Zip encodes" (i. e., the 7-Zip GUI, e. g. using the context menu in Windows Explorer)