[PATCH 2/5] dirmngr: hkp: avoid race condition when some hosts die

Werner Koch wk at gnupg.org
Wed Nov 2 08:34:51 CET 2016


On Tue,  1 Nov 2016 20:39, dkg at fifthhorseman.net said:

> hm, interesting.  If you can avoid blocking system calls, you can get
> the same benefit by just using an event-driven model, with no explicit
> threads at all, right?

Right.  The problem with event driven processing is that complex
operations require a lot of state and it is not anymore possible to
follow the processing flow in an audit.  Thus the use of an automatic
co-processing model (to decribe nPTh in a different way) makes things a
lot easier.

> It's not clear to me what features or fixes are required -- but i know
> that i want to help people have a functional, smoothly-running system
> that supports the modern GnuPG mini-server model without them noticing

I know.  However we need to prioritize what we do.  Even if I say these
patches should go into 2.3 they can also be backported to 2.2 to make it
into a Debian point release.

> fwiw, this dirmngr series is not a lot of "new code".  rather, it's a
> reduction (two patches in this series are simply dead code removal):

Maybe.  Removing code may also break things.  I have also noticed some
portability problem.

> I don't want these patches to get lost, and i'd love review.  however, i

I would suggest to open a feature request and attsch the patches.


Salam-Shalom,

   Werner

-- 
Die Gedanken sind frei.  Ausnahmen regelt ein Bundesgesetz.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 194 bytes
Desc: not available
URL: </pipermail/attachments/20161102/6fe705e4/attachment.sig>


More information about the Gnupg-devel mailing list