- move all filtering into SQL query in calendar_so, to be able to correctly return N rows starting from row M
- re-enabling propfind iterator again for calendar (fetching events in chunks of 500), to lower memory footprint
Please note: changed SQL queries used for CalDAV do not take changed participants (or status) in exceptions into account
- merged: r34529, r34584, r34592, r34594, r35948
- Lightning pops up alarm, until Sequence/etag get updated: if user has no edit rights on an other users calendar, etag never got updated, now we update it
- fixed user was not able to add alarms via CalDAV, if he had no edit rights for event (was always possible in web UI)
- alarms from other users calendars are not included any more, as they make no sense but a lot of trouble
- fixed wrong condition on adding alarms, causing some alarms no being saved
Replacement code for in r29270 removed filter
$cal_filters['query']['cal_reference'] = 0;
only works, if all recurrences and the master are returned in a single chunk, otherwise events get returned multiple times"
resource of the series master. Now the status of single recurrences of a
serie are send to the clients as (virtual) exceptions as Jaytrax&Joerg
implemented it already for SyncML.
The implementation is unfortunately a little different, as CalDAV
differs from SyncML and I dont know the SyncML part that well. Maybe we
can re-unify the code again together.
Tested so far with Tb3/Lightning1.0b and a little with iPhone.
Please let me know, if you run into problems with other clients.
Patch is mostly created by script in egroupware/doc/fix_depricated.php in separate commit.
I do NOT advice to apply this patch to a production system (it's commited to trunk!), as the automatic modified regular expressions have a good change to break something ...
