[Csync2] sqlite3 and _FILE_OFFSET_BITS=64

Art -kwaak- van Breemen ard+csync2 at telegraafnet.nl
Fri Jun 13 12:28:44 CEST 2008


On Thu, Jun 12, 2008 at 09:29:13PM +0200, Lars Ellenberg wrote:
> On Thu, Jun 12, 2008 at 07:10:31PM +0200, Art -kwaak- van Breemen wrote:
> > On Thu, Jun 12, 2008 at 04:20:31PM +0200, Art -kwaak- van Breemen wrote:
> > > I've noticed that there was no (working) sqlite3 support.
> > > So I've continued working on the patch I found on:
> > > http://tmp.askask.com/2008/01/csync2-sqlite3-wip.patch
> > 
> > Hmmm, only now I found another patch which seems correct:
> > http://lists.linbit.com/pipermail/csync2/2008-January/000349.html
> 
> so, which one shall I integrate?
The patch from
http://lists.linbit.com/pipermail/csync2/2008-January/000349.html
seems much more sane than my patch, but it contains more changes
than just sqlite3.

Both patches means that sqlite < 3 will not be supported any
more.

> did anyone else test these things?
> 
> please step forward and say "worked for me", if so.
> I don't have too much time to play on this right now,
> so I have to rely on your feedback.

Both did not "worked for me" yet, I did not do a sync to other
systems with my patch, and I haven't tested that other patch.
I will however continue with the other patch, and remove the
pieces that are not sqlite3 specific.

My biggest question for now: do I need to make the patch backward
compatible with sqlite? I personally think it's essential to
support sqlite to get csync2 with sqlite3 support adopted with
major distributions.

The second problem I've encountered was the:
_FILE_OFFSET_BITS=64 .
I am not really sure what the best place is to put it. But I am
pretty confident that it needs to be there, since csync2 stops to
work on 32 bits systems and large files. I wouldn't want to sync
a 4G file, but if it is there it should cope with it.

Regards,
Ard van Breemen


More information about the Csync2 mailing list