[sync] Client and server timestamps don't match

[jochem at mondrian-it.nl]

Hi all,

While testing the funambol windows mobile pluging 3.0.31 against horde's
syncml backend, I get the following INFO level entry in the log:

client requested sync with anchor ts 16252726 but server has
timestamp16232634 on file. Enforcing SlowSync

The time on the PDA and the server are (almost, only a few seconds off)
identical, and I've waited long enough between syncs to rule out that the
small difference might be a problem. 

I've tried disabling the check in the code in
/usr/share/php/SyncML/Command/Alert.php (the warning is on line 79) so that
the difference is ignored, but this causes duplicate entries to show up on
the PDA.

With the check enabled, the server is forcing a slow sync, which causes all
changes on the server to be lost, and all items from the pda to be copied to
the server.

I'm trying to synchronize calendar items using SIF Format. Using iCalendar,
my test item from the pda isn't synced to the server, even though the sync
is reportedly completed successfully. It does generate a number of php
notices when using iCal:

Mar 08 14:10:05 HORDE [warning] [horde] SyncML: Caught output: <br />
<b>Notice</b>:  Undefined index:  AllDayEvent in
<b>/usr/share/php/SyncML/Device/Sync4j.php</b> on line <b>309</b><br /> <br
/> <b>Notice</b>:  Undefined index:  Start in
<b>/usr/share/php/SyncML/Device/Sync4j.php</b> on line <b>323</b><br /> <br
/> <b>Notice</b>:  Undefined index:  End in
<b>/usr/share/php/SyncML/Device/Sync4j.php</b> on line <b>326</b><br />  [on
line 211 of "/usr/share/php/SyncML.php"]

For both SIF and iCalendar, this following php Notice is also caught:

Mar 08 14:12:22 HORDE [warning] [horde] SyncML: Caught output: <br />
<b>Notice</b>:  Only variable references should be returned by reference in
<b>/usr/share/php/Horde/DataTree.php</b> on line <b>345</b><br />  [on line
211 of "/usr/share/php/SyncML.php"]

I'm using today's (Thursday March 8, 2007) HEAD release, running on PHP
4.4.0 and Apache 2 & SuSE 10.0.

I'll try and dig into the problem a bit more, see if I can find a solution,
and if I do I'll post it here. In the mean time, any suggestions are
welcome.

Regards,

Jochem