On Tue, 2009-04-14 at 12:05 +0300, Juha Tuomala wrote:
On Monday 13 April 2009 10:35:39 Adam Williamson wrote:
> On Thu, 2009-04-09 at 23:15 +0200, Kevin Kofler wrote:
>
> > > Some contacts had got duplicated for some reason, but still,
nice effort.
> >
> > Unfortunately, as I've been told, duplication of data is an
inherent issue
> > with libopensync 0.22, and the main reason why they're working on
> > redesigning everything in 0.3x/0.40. Unfortunately, their redesign
is
> > breaking everything at the moment. :-(
>
> Yes, but it's more of a long-term issue (or if you get an incomplete
> sync - the problem is a 'slow sync' duplicates data). I wouldn't
expect
> it as a matter of course on a single-shot test sync. As I said,
though,
> it could actually have been duplicated prior to the sync, I didn't
think
> to check. I'll do some more extensive testing with various devices
in a
> bit.
- In normal use, it will at some point fall into slow sync, sooner or
later
and create duplicates¹. I don't think any user would want that.
- opensync 0.2x cannot detect² the SAME/SIMILAR cases reliably.
- Thus only because of these two issues, (plus hundred two smaller
ones)
opensync 0.2x is broken, by design, inside the core - regardless of
plugins.
- It's useless to ship completely broken (not low Quality, broken)
software.
- Fedora 11 will ship them anyway.
Only thing unclear to me is why?
As I said to you in private, in practice I have not observed this as a
major issue. I did extensive testing with several devices and sync
groups over a long period with 0.22 in Mandriva and did not observe any
duplication of data.
--
Adam Williamson
Fedora QA Community Monkey
IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org
http://www.happyassassin.net