[avahi] Announcing libavahi-ui
lennart at poettering.net
Thu Feb 8 08:10:52 PST 2007
On Thu, 08.02.07 15:58, Sjoerd Simons (sjoerd at luon.net) wrote:
> > Hmm, I don't think this is a good idea. Firstly, the commit() function
> > does something similar. (and does it the real way, i.e. it is indeed
> > atomic) Admittedly though, it works only once after creation of the entry
> > group. Secondly, Avahi accumulates multiple changes into a single
> > packet anyway. i.e. you have at least 50ms or so to pass your TXT
> > changes to Avahi and they will be sent out to the network as a single
> > packet.
> Yeah, the freeze and thaw stuff is purely for the Second case. While it's
> great that avahi optimizes so changes are batched, freeze/thaw does actual give
> you guarantees instead of assuming that you'll be fast enough (Oh and probably
> saves some dbus round-trips too).
Hmm, your code makes guarantees? How's that if the underlying Avahi
code cant give you those?
> > > Any feedback is very much appreciated :)
> > I forgot to paste the proper URLs in my original mail:
> > http://davyd.livejournal.com/152749.html
> > http://oracle.bridgewayconsulting.com.au/~davyd/misc/main.c.html
> Ah this seems somewhat more high-level, while my code is really quite a thin
> wrapper around avahi's native API. Both have their advantages and disadvantages
> Did you contact Davyd for feedback or ... ?
Nope, I didn't. I wasn't sure if your code was already at a point
where you want public feedback.
Lennart Poettering; lennart [at] poettering [dot] net
ICQ# 11060553; GPG 0x1A015CC4; http://0pointer.net/lennart/
More information about the avahi