http://bugzilla.moblin.org/show_bug.cgi?id=2425
--- Comment #5 from pohly <patrick.ohly(a)intel.com> 2010-03-04 02:38:37 PST ---
(In reply to comment #4)
> True. However, I learned in the meantime that the binfile
support is not
> enabled for SyncML servers. Once we turn SyncEvolution also into a server, we
> need to revisit this issue.
... so now we need to solve this issue. Without it, our suspend/resume tests
against our own server should fail.
Only in some very constructed situation, and one which is not currently covered
by any of our tests. Here's where the two anchors make a difference:
1. client and server in sync
2. item added on server
3. client and server sync
4. suspend after sending that new item to client
5. a second item added on server
5.1 resume => send changes since step 4
5.2 two-way sync instead of resume,
using anchors from step 1 => send changes since
step 1
Our server fails in 5.2, because it will always only send changes since the
last sync (step 4 in this example).
The relevance of this is debatable. It depends on clients actually being able
to roll back to a previous state. Our own client can't do that, so it would
have to do a slow sync in 5.2 instead of resuming.
More important is support for blobs. Large items that are not transferred
completely at the time of the suspend are stored as a blob.
--
Configure bugmail:
http://bugzilla.moblin.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching someone on the CC list of the bug.