Mailing List Archive: OpenPGP smartcard and KMail (KDE 3.4)

OpenPGP smartcard and KMail (KDE 3.4)

Mar 30, 2005, 7:14 AM

Post #1 of 3 (2890 views)

Hi

I've filed a bug with KMail (http://bugs.kde.org/show_bug.cgi?id=102478) and
was told that you're the ones to ask.

I try to encrypt/sign a mail (OpenPGP/MIME) with KMail. The private key
resides on a OpenPGP smartcard; GnuPG version is 1.4.0 from Debian unstable.
KDE version is 3.4 (from http://pkg-kde.alioth.debian.org/kde-3.4.0/).

KMail never asks for the smartcard's PIN and hangs. Until I kill the gpg
process (in this case "gpg --no-sk-comment --status-fd 41 --no-tty --charset
utf8 --enable-progress-filter --command-fd 43 --sign --detach --armor -u
0F2B3C20E772CFE9").

Using a key that isn't on the smartcard works.

Is this a misconfiguration or a bug? What can be done?

Thomas Hühn

Re: OpenPGP smartcard and KMail (KDE 3.4) [ In reply to ]

wk at gnupg

Mar 31, 2005, 12:46 AM

Post #2 of 3 (2765 views)

Permalink

On Wed, 30 Mar 2005 16:14:33 +0200, Thomas HÃ¼hn said:

> KMail never asks for the smartcard's PIN and hangs. Until I kill the gpg

I assume you are not using the gpg-agent.

It is either a bug in GPGME's passphrase callback code or even one in
Kmail. Please run KMail this way:

GPGME_DEBUG=5:/tmp/gpgme.dbg: kmail

and send me gpgme.dbg by PM. Passphrases or PINs entered durng this
session are visible in gpgme.dbg but as it doesn't ask you it
shouldn't be a problem.

Salam-Shalom,

Werner

Re: OpenPGP smartcard and KMail (KDE 3.4) [ In reply to ]

bernhard at intevation

Apr 1, 2005, 3:33 AM

Post #3 of 3 (2772 views)

Permalink

On Thu, Mar 31, 2005 at 09:46:04AM +0200, Werner Koch wrote:
> On Wed, 30 Mar 2005 16:14:33 +0200, Thomas Hühn said:
>
> > KMail never asks for the smartcard's PIN and hangs. Until I kill the gpg
>
> I assume you are not using the gpg-agent.

Just a side note: At the Freiheitspart in Berlin two weeks ago,
I also had a hang _with_ the gpg-agent when the card was not inserted.
Even as I reinserted, I needed to kill it.

Also the pin was not cached.

There was no time to further investige, so this is just an
observation, not a real bug report.