Seahorse embeds filename "-&25"

[Mark Preetorius]

Binary package hint: seahorse

If you decrypt files encrypted with seahorse and try to restore the embeded
file name (default behavior of PGP 7.0.3, gpg knows the option
--use-embeded-filename), you always get "-&25" als filename of the decrypted
file.

When encrypting with gpg on the command-line with the same options, the problem
does not occur.

Tested in seahorse 2.26.1 (Ubuntu 9.04)

Steps to reproduce:
1. encrypt file with seahorse
2. decrypt it with PGP 7's default settings or gpg --use-embeded-filename

Actual results:
decrypted file is always named "-&25"

Expected results:
foo.bar.pgp should be decrypted to foo.bar (or another embeded filename if
foo.bar was renamed after encryption)

Does this happen every time?
yes

Other information:
setting an empty filename by adding

set-filename ""

to the gpg.conf is a workaround, but an ugly one.

I also posted this as a bug in the upstream-package here: http://bugzilla.gnome.org/show_bug.cgi?id=588432
Is that the correct proceeding?

[Andreas Moog (ampelbein)]

Thank you for your report. It is generally not necessary to report issues to GNOME and here. But as Launchpad offers the possibility to link another bugtracker to a report, no harm is done. See the "Also affects project" link, it is used to tell LP what other project is affected.

[Patrick Toomey (ptoomey3)]

This bug appears to be do to the seahorse failing to use gpgme_data_set_file_name() on the plaintext data before it is encrypted. This results in some clients (ex. PGP) naming the decrypted file incorrectly. I have attached a patch that should fix the behavior.

[Patrick Toomey (ptoomey3)]

Of course I realized about 2 minutes after submitting that the string returned from g_file_get_basename() must be freed after use with g_free. This patch should be used in place of the previously submitted patch.

[Patrick Toomey (ptoomey3)]

Of course I realized about 2 minutes after submitting that the string returned from g_file_get_basename() must be freed after use with g_free. This patch should be used in place of the previously submitted patch.

[Patrick Toomey (ptoomey3)]

I submitted a link back to the patch on the seahorse mailing list. One of the upstream maintainers said he would test the patch when he gets a chance.

[Sebastien Bacher (seb128)]

Thank you for your work, seems the patch is close to be accepted upstream but wait on a reply to a question, could you update the bug?