Strange Key ID problem. : subkey id is reported where primary key id should be reported as missing?

[ishikawa]

Dear Werner Koch,

Thank you for your prompt feedback.


On (06/29/2009 04:03 PM), Werner Koch wrote:
> On Mon, 29 Jun 2009 07:56, chiaki.ishikawa at ubin.jp said:
>
>> (After writing this summary, I now think it is a bug, but am not sure
>
> Yes it is a bug.  But not in GnupG but in tghe keyserver software.
>

Oh, I didn't realize this.
It is very difficult to figure this out as this stands.

>> So I tried to obtain the key using gpg with arguments such as
>> --keyserver pgp.mit.edu -recv-keys 17785FE8
>> --keyserver pgp.nic.ad.jp -recv-keys 17785FE8
>> to no avail.
>
> Both keyservers run old and non-OpenPGP compatible software - Do not use
> them.
>
> I tried this too (using gpg2 but that doesn't matter)
>
>    $ gpg2 --recv-key 17785FE8
>    gpg: requesting key 17785FE8 from hkp server keys.gnupg.net
>    gpg: key 812347DD: public key "Mozilla Software [...]
>    gpg: Total number processed: 1
>    gpg:               imported: 1
>
> So everything is fine.  The key was taken from one of the SKS keyservers
> to which keys.gnupg.net resolve.  If you now look at the key:
>
>    $ gpg2 --list-key 17785FE8
>    pub   1024D/812347DD 2007-07-17 [expires: 2009-07-16]
>    uid                  Mozilla Software Releases<releases at mozilla.org>
>    sub   1024D/17785FE8 2007-07-17 [expires: 2009-07-16]
>    sub   2048g/1B0EC2E7 2007-07-17 [expires: 2009-07-16]
>
> You will notice that 17785FE8 is a signing subkey.  The old pks software
> can't handle this and breaks the key.
>
> We do have this problem for years now and at least the admins of
> pgp.mit.edu don't care about it and keep on running this broken
> software.
>
> I am now considering whether we should detect
>    Server: pks_www/0.9.6
> and reject this one as entirely broken.  Maybe just a warning if
> --expert is used.

I think that maybe it is a good idea to
have this rejection.
Then at least, the user of gpg will try to move to the
sites that run newer software.
(I wonder though why my setup didn't access keys.gnupg.net automatically.
Maybe the other PCs on which gpg worked to fetch such keys
were configured properly to access newer sites whereas the
said PCs in my post were configured lately with gpg and
so I had to type in keyserver names manually (I didn't have the
time to edit the configuration file. I just tried to verify the
archives on the fly.)

Also, such messages *may* encourage the admins of sites
that run old software to move on to the newer software when
a system upgrade is attempted. Just a chance.

So I am for such a change.

Thank you again for your explanation.

Sincerely,
Chiaki Ishikawa


> Shalom-Salam,
>
>     Werner
>