[OT] New keyserver

Richie Laager rlaager at wiktel.com
Wed Sep 18 00:46:01 CEST 2002

Hash: SHA1

> -----Original Message-----
> From: V. Alex Brennen [mailto:vab at cryptnet.net] 
> Sent: Tuesday, September 17, 2002 12:48 PM
> To: Jan-Benedict Glaw
> Cc: gnupg-devel at gnupg.org; flo at rfc822.org; 
> guckes at math.fu-berlin.de; pgp-keyserver-folk at flame.org; 
> cks-devl at mail.freesoftware.fsf.org
> Subject: Re: [OT] New keyserver
> On Tue, 17 Sep 2002, Jan-Benedict Glaw wrote:
> > 4. Is a new keyserver really what folks around would like to see?
> > Is anybody more like "It's my baby, go away and do something
> > else?" Or should we take current pksd and send patches as long as
> it's not 100%
> > what's needed?
> > 
> > *Please* comment on this!

I'm always open to patches for PKS. PKS was just moved to SourceForge
for continued development. (The SF project name is "pks".) We've
consolidated almost all of the known patches and are preparing for a
0.9.5 release candidate. The discussion for this has been taking
place on pgp-keyserver-folk at flame.org. To subscribe, send a message
to pgp-keyserver-folk-request at flame.org with the word subscribe in
the body. (Perhaps we should have a separate mailing list for PKS,
but that's another topic altogether.)

> Well, CKS has most of the features I think you're looking for
> already  implemented.  It uses postgres, and has the beginnings of
> database  abstraction so that it can go against Berkeley db, or
> even oracle.  

Have you considered using unixODBC? It seems to me that ODBC would
make it very portable across databases. Of course, this would require
the use of SQL, right? Is that an issue?

> It seems a shame to re-duplicate all of that work.  Progress on CKS
> has been slow lately, but I haven't gotten any patches.  If 
> you where to put time into CKS, we'd be able to get a powerful 
> stable keyserver faster.  And we could refocus the duplicated 
> efforts on new features like respecting the 'no-modify' flag.

I feel that there are good reasons to duplicate code and bad reasons
to duplicate code, and that code duplication shouldn't be a major
factor in using/developing PKS vs. CKS. In fact, I don't see a reason
that we should develop one or the other. Why not keep both under
development. Choice is good.

> If there is something you don't like about CKS that is preventing
> you from working on it, please let me know and I will most likely
> be willing to change it.  I haven't rejected any patches submitted
> by other yet.  If other people are willing to get very involved,
> I'll set up CVS on my server for the project.

Nothing is preventing me from working on CKS. In fact, I might.
However, I had some trouble checking the code out of CVS on Savannah.
When I used "subversions.gnu.org" as the host instead of
"savannah.nongnu.org", it worked. Any ideas on what went wrong?

Richard Laager

Version: PGP 7.0.4


More information about the Gnupg-devel mailing list