Re: [coldsync-hackers] Titrax sync questions...

[Andrew Arensburger <arensb+CShackers at ooblick dot com>]

On Thu, 12 Jun 2003, Izzy Blacklock wrote:
> I've been thinking about a future need for my conduit.  Eventually I'm going
> to have to build a sync function.  What has me worried, is how this will work
> with multiple files.  I guess the root of my confusion is that I don't really
> know how the sync process works.  I have to admit, I haven't looked at any
> sync conduits yet, so perhaps I'm worrying over nothing.
>
> My biggest concern is on the timing and order of things.  Does a sync occur
> after each file is read from the palm?  ie, Does the palm expect to receive
> the sync'd version of a pdb right after it sends the pdb, or does it send all
> the pdbs then wait for attempts to write data back?  Does coldsync run the
> various conduits as it reads the pdb from the palm, or does it wait until it
> has them all then run the conduits then write them back to the palm?

	That's not quite how conduits work, or at least not sync conduits.
	Fetch and Dump conduits get complete PDB files to work with. Sync
conduits don't. The purpose of a Sync conduit is to talk to the Palm and
ensure that the database on the Palm is the same as the one on disk.
	For instance, the (default) generic conduit, which is a sync
conduit, downloads each modified record from the Palm in turn, compares it
to what's on disk, and then decides what to do: if the on-disk record
hasn't changed, it replaces the on-disk record with the on-Palm record; if
both have changed, and aren't identical (i.e., they've been modified in
different ways), it creates new records so that both the on-Palm and
on-disk databases have both versions of the record.
	Yes, this can get complex, but that's what a Sync conduit is
expected to do.

> I fear that the sync process is being run after each file is received.

	Not quite: part of the job of a Sync conduit is to read the
on-disk file and the on-Palm database.

> If
> that is the case, building a sync process for titrax would be problematic.  I
> would need to modify at 4 pdbs; TitrixNamesDB, TitraxDataDB, TitraxNoteDB,
> and datebookDB.  If I trigger on the Titrix pdb, how do I insure that the
> latest version of the rest of the files has already been fetched?  How do I
> get the conduit to write out the other pdbs after the TitrixNames conduit has
> finished?

	The only thing that leaps to mind is to have the conduit handle
all four databases. As each one runs, it creates or checks a temporary
file, and writes to it the current time and the database that it's
currently syncing. If all four databases aren't listed in the file, it
proceeds no further.
	Either that, or the first time it runs, it syncs all four
databases and writes a timestamp somewhere. The subsequent three times, if
the timestamp exists and is more recent than a few minutes, it doesn't
bother.
	You could probably play other tricks along the same lines. For
instance, if you sync all four databases the first time, then the next
three times, neither the on-Palm nor on-disk databases will have any dirty
records. You could use that as a flag.

-- 
Andrew Arensburger                      Actually, these _do_ represent the
arensb@ooblick.com                      opinions of ooblick.com!
                        Generic Tagline V 6.01
-- 
This message was sent through the coldsync-hackers mailing list.  To remove
yourself from this mailing list, send a message to majordomo@thedotin.net
with the words "unsubscribe coldsync-hackers" in the message body.  For more
information on Coldsync, send mail to coldsync-hackers-owner@thedotin.net.