Discrepancies in extracted photo-id images from dumps

[Peter Lebbing]

On 20/01/2019 16:05, Stefan Claas wrote:
> Thanks, but it is still unclear to me what content of the user id
> is taken. Here for example an old key from me:

I had a quick scan through the source code, but couldn't find it. But it
seems to me it's likely to be a hash of the User Attribute Packet (RFC
4880), which is an alternative form of a User ID. As such, it would be a
unique identifier of the image in question. Does it really matter?

> If i base32 decode the strings i get binary output which i don't
> understand.

It's 32 digits of base32, which makes it 32*2log(32) = 32*5 = 160 bits.
So my guess is it's a raw hash and as such not something to be
"understood" but high entropy data. Either a 160-bit hash (seems likely)
or a truncated longer hash.

Oh, note that the base32 is the one from RFC 6189, not the one from
3548. I saw that in the source. I played around a bit, but couldn't
quickly home in on what the hash actually is. Maybe you have better
luck (or more patience :-). Hashes are rather unforgiving of mistakes
you make in the input to it :-).

HTH,

Peter.

-- 
I use the GNU Privacy Guard (GnuPG) in combination with Enigmail.
You can send me encrypted mail if you want some privacy.  My key is
available at <http://digitalbrains.com/2012/openpgp-key-peter>

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: OpenPGP digital signature
URL: <https://lists.gnupg.org/pipermail/gnupg-users/attachments/20190120/2cadfb10/attachment-0001.sig>