[solved] Problem with specific UTF-8 chars sequence in name inside ZIP archive.

[TomasT342]

When I try rename the file inside the ZIP archive to "pošťák.txt" (postman in Czech language), the name is changed to "po眠k.txt" while saving.
The names "pošťak.txt" and "posťák.txt" are saved without any problems.
Name "Ž pošťák.txt" is saved OK, name "Z pošťák.txt" is saved as "Z po眠k.txt"
The already existing name "pošťák.txt" inside the ZIP archive is displayed without problems.
Configuration for Zip: "Internal Zip packer", "Store Unicode names as UTF-8" (the change does not affect the problem)

In the previous version of TC (10.x) on the same computer, the name "pošťák.txt" is saved without any problems.
(Windows 10 22H2 build 19045.3324, TC 11.00)

[petermad]

I cannot reproduce this, when copy-pasting the names from your report to the new name inside the archive.

When I try rename the file inside the ZIP archive to "pošťák.txt"

The already existing name "pošťák.txt" inside the ZIP archive is displayed without problems.

These two lines seems to contradict each other - how can you rename a file to a name that is already present inside the archive?

[Gral]

I can cofirm this with Polish locale so the same codepage is used. Interesing that this invalid name is correct in DOS codepage - 852

The already existing name "pošťák.txt" inside the ZIP archive is displayed without problems.

These two lines seems to contradict each other - how can you rename a file to a name that is already present inside the archive?

Probably because this is another copy, and:

In the previous version of TC (10.x) on the same computer, the name "pošťák.txt" is saved without any problems.

[petermad]

Hmm, all I know that this zip-file: https://tcmd.madsenworld.dk/postmantest.zip is packed whith TC 11.00, and all the filenames looks like what I see in the first post of this topic.

[TomasT342]

I tried different renames again today and the problem is really strange.
It really isn't easily reproducible.

I named one particular file in the ZIP archive "Pošťák", when saved it changed to "Po眠k" (that's a mistake).
I changed its name again and just added a space at the end, the file was saved as "Po_k " (that's a mistake).
I changed its name again, removed the space, the file was saved as "Po_k" (it's fine now).
Now I have renamed it again to "Pošťák" and the name has been saved correctly.

Another interesting finding - if I open the archive in WinRAR, the file name appears to be fine. The file that I see in TC11 as "Po眠k" I see in WinRAR as "Pošťák".

There is some problem there at the beginning, but it manifests itself randomly and can probably be removed by repeated renaming.
I am adding a link to the ZIP archive in which the problem occurred:
https://anonymfile.com/2oOxA/brin-d-pok.zip

[jbcb]

Confirmed.

postman.png

[ghisler(Author)]

This happens because the character sequence is valid UTF-8, and some packers don't set the UTF-8 flag when packing.
Please add the line
PreferUtf8ForZip=0
to your wincmd.ini, section [Packer].

Here is a thread with a similar problem:
viewtopic.php?p=429308#p429308

Unfortunately there is no solution, the name is valid both in local encoding and UTF-8.

[TomasT342]

Thanks, this solution is suitable for me.

[petermad]

Moderator message from: petermad » 2023-08-15, 22:47 UTC

Moved to "Total commander (English)" from "TC11.x bug reports (English)"