[sync] Turba and Syncml status?

[Karsten Fourmont]

 > Also, please remember that Turba is currently in release candidate
 > status. So patches that change significant chunks of code should be
 > held off on for now.

OK. I'm not going to do fance image support right now anyhow. Without 
that we only need a few things:

1) additional fields in the database: instead of
         'homeAddress' => 'object_homeaddress',
basically this:
         'homeAddress'  => 'object_homeaddress',
         'homeCity'     => 'object_homecity',
         'homePostalCode' => 'object_homepostalcode',
         'homeCountry' => 'object_homecountry',
so an address is nicely split into the various fields. (Probably a good 
thing anyhow)

This requires changes to the database, config/sources.php and 
config/attributes.php. As this is all configuration data it should be 
easy to provide the required data as a _seperate_ config example. Shall 
I put it directly into the turba package or just somewhere into syncml/docs?

2) changes to api.php. This is a bit more of an issue. Currently 
turba_exports reads like this:

// Need to reference some sort of mapping here to turn the UID
// into a source id and contact id:
// $source = $uid['addressbook'];
// $objectId = $uid['contactId'];
return PEAR::raiseError('Turba needs a UID map');

And this will obviously do nothing, only complain :-(

Syncml already has a UID map to relate client (PDA) UIDs to horde UIDs. 
Currently it stores a turba ID like this:
turba:localsql:a7852bb273265eb258ed2ea53692497c

So what format am I supposed to pass the uid in? Will something like
$uid= array('addressbook' => 'localsql' , 'contactId' => 
'a7852bb273265eb258ed2ea53692497c') ;
for the above example do?
Can I remove the comments and the raiseError statement?

The same applies for turba_replace as well.

I'll post a patch-proposal to the list.

Cheers
  Karsten





Chuck Hagenbuch wrote:
> Quoting Karsten Fourmont <fourmont at gmx.de>:
> 
>> I tried to include support for images into turba. However that turned 
>> out to be a bit tricky (rdbms indepented code for binary attributes, 
>> php pages that
>> return image data, stuff like that). So I better leave it out for now 
>> rather
>> than further delay things.
> 
> 
> Also, please remember that Turba is currently in release candidate 
> status. So
> patches that change significant chunks of code should be held off on for 
> now.
> Send it to the list if you want a judgement call on whether or not it 
> changes
> too much...
> 
> -chuck
>