Summary | Request: Support for task relations in SyncML |
Queue | Nag |
Queue Version | 2.3.1 |
Type | Enhancement |
State | Accepted |
Priority | 2. Medium |
Owners | |
Requester | felix.leimbach (at) gmx (dot) net |
Created | 02/15/2009 (5932 days ago) |
Due | |
Updated | 04/17/2010 (5506 days ago) |
Assigned | |
Resolved | |
Milestone | |
Patch | No |
in the AS protocol. There are also quite a few clients (iPhone
included) that have NO support for tasks whatsoever.
SyncML TODO+Cal application and that will still require SyncML.
So I'll come up with a patch for this request when I find the time.
whether it is useful at all, since from what I hear horde 4 will
have a new Active Sync component:
"Horde 4 will also sport more mobile device features with an
ActiveSync server component entering the main development tree in
time for the final release." (from
http://www.techworld.com.au/article/342104/horde_open_source_groupware_preps_version_4_release)
devices support ActiveSync (most notable blackberry devices) and not
all devices support SyncML, so we do still need to support both.
the AS protocol. There are also quite a few clients (iPhone included)
that have NO support for tasks whatsoever.
it is useful at all, since from what I hear horde 4 will have a new
Active Sync component:
"Horde 4 will also sport more mobile device features with an
ActiveSync server component entering the main development tree in time
for the final release." (from
http://www.techworld.com.au/article/342104/horde_open_source_groupware_preps_version_4_release)
Do you anticipate that task relations will work out of the box in Horde 4?
the UID attribute.
but we would definitely look at a patch.
written properly. Only when syncing they are not.
Synthesis client does not belong in that category.
I want to create a patch to selectively disable that work-around for
known-good clients. I'm currently digging into the lib/SyncML/ folder
and my only uncertainty is:
I don't see the RELATED-TO field being removed anywhere.
Any clue as to why they are missing?
Also, are you interested in such a patch?
create GUIDs
needs to assign and send the UID (not GUID) and the client just uses
them.
them UIDs that are simply incrementing integers. These are of course
not globally unique IDs so we can't rely on them.
written properly. Only when syncing they are not.
http://forum.synthesis.ch/showpost.php?p=3204&postcount=7
So this seems like a horde bug to me, or am I missing something?
SyncML clients.
needs to assign and send the UID (not GUID) and the client just uses
them.
written properly. Only when syncing they are not.
I found this analysis of the problem by the synthesis developers helpful:
http://forum.synthesis.ch/showpost.php?p=3204&postcount=7
So this seems like a horde bug to me, or am I missing something?
GUIDs, but simply some incrementing number for this field. Thus we
don't write or read UID fields during import/export, because they
won't be used anyway, and wouldn't match the UIDs we use in the SyncML
protocol.
the vTodo format the UID and RELATED-TO attributes are set correctly.
But when I sync via a SyncML client and sniff the network traffic I
see that the UID and RELATED-TO are missing as described in the
synthesis forum post linked below.
Any clues?
Milestone ⇒
State ⇒ New
Patch ⇒ No
Queue ⇒ Nag
Summary ⇒ Request: Support for task relations in SyncML
Type ⇒ Enhancement
Priority ⇒ 2. Medium
However, when syncing tasks with SynML those relations are not
preserved.
That is because horde does not fill the UID and RELATED-TO fields properly.
The effect is, that nested tasks loose their relations when synced.
A Synthesis developer has tracked down and described the issue in a
forum post:
http://forum.synthesis.ch/showthread.php?p=3204#post3204