Tickets :: Comment on [#8483] Contact deleted from mobile after sync because of a move into a new addressbook

* Your Email Address
* Spam protection	Enter the letters below: . . __..__. ..__. \ /(__ \| \| \|\| \| \/ .__)\|__\|\__\|\|__\
Comment	> Ok, > > > > I found the problem ! > > > > Real how to reproduce bug : > > > > 1. Create a new contact with mobile > > > > 2. Sync this contact in Horde (default addressbook A) > > sync line created in horde_syncml_map, timestamp = 1249678393 > > an add line is created in history table for the addressbook A => > turba:1b6e0d16bbb5eb2216da5d2265e95d0b:20090807225313.19014a63moznjl15@mail-test.exemple.fr, timestamp = > 1249678393 > > > > 3. Move this contact from the addressbook A to addressbook B (the > both are marked as synced in turba options) > > add line created in history table for addressbook B => > turba:localldap:20090807225313.19014a63moznjl15@mail-test.exemple.fr, > timestamp = 1249678421 > > delete line created in history table for addressbook A => > turba:1b6e0d16bbb5eb2216da5d2265e95d0b:20090807225313.19014a63moznjl15@mail-test.exemple.fr, timestamp = > 1249678421 > > > > 4. Sync again mobile with horde > > the contact is deleted from mobile !!! Expected result => nothing ! > > > > It's in indirectly bind to the file : > > > > lib/SyncML/Backend/Horde.php, line 199. > > It call the turba method "getActionTimestamp" to get the TimeStamp > for add action. > > But if we examine the function Turba api.php > _turba_getActionTimestamp($uid, $action, $sources = null), we see > that as the source parameter is null, it loop on sync books and as > soon as it find an entry in history for this contact, it return it ! > But there is 2 entries for this contact in 2 differents addressbook. > And it return the result for the bad addressbook... So the timestamp > is equals to sync timestamp and it finish by "Added to server from > client: 20090807225313.19014a63moznjl15@mail-test.exemple.fr ignored" > > > > Well, you have surely a best knowledge of Horde / Turba / SyncML so > maybe you can submit a patch for this bug. In all cases, I will try > to resolve it in theses days ! > > > > Thanks > >
Attachment
Watch this ticket

Saved Queries