Page MenuHome GnuPG

gpgtar Problem with utf8 string/length of filenames
Closed, ResolvedPublic


An archive (.tar.gpg) encrypted with gpg4win version 3.1.12 could not be decrypted with versions 3.1.12/11 (Windows 10).
Input error: Error while running C:/Program Files (x86)/GnuPG/bin/gpgtar.exe: gpgtar: error creating ….(Path+filename 220 characters) No such file or directory
Suspected cause: Very long filenames with special characters and umlauts (ä,ö,ü...) were used.
Decryption attempt with version 4 (Windows 10) also failed:
filename.tar.gpg → Eingabe von C:/Program Files (x86)/GnuPG/bin/gpgtar.exe --utf8-strings --openpgp --skip-crypto ...:
Entschlüsseln fehlgeschlagen: Input/output error.
Eingabefehler: Fehler während der Ausführung von C:/Program Files (x86)/GnuPG/bin/gpgtar.exe: gpgtar: Ungltige Option "--utf8-strings"
Step by step solution attempt under powerShell did work:

  1. decoding the tar file worked. gpg -tar -x filename.tar.gpg -> filename.tar
  2. tar -x -v -f filename.tar -> creates the original directory tree

Suggested solution: Do not abort if files cannot be created, but continue and create all that work.
If there are restrictions regarding the file names (character set, length), these should be documented. Cleopatra should check these before encoding.


3.1.12 and 4

Event Timeline

werner claimed this task.
werner added a project: gpgtar.
werner added a subscriber: werner.

The bug with the long filenames has been fixed but it is not yet released. Release will be in gpg4win 4.0.1 See T5754.

Full command line Unicode support is available in gpg4win 4.0.0

We can't ignore such a bug because it would mean undetected data loss in unattended environments.

Thanks for the good report.