Tickets :: Comment on [#12175] Add modification sequence support to Horde

* Your Email Address
* Spam protection	Enter the letters below: __.._..___.___.___. (__ \| [__ [__ \| .__)_\|_\| \| \|
Comment	>>> This will miss changes again if the server is too slow (think about a >>> low end server swapping for a few minutes). The "endTS - fudge >>> factor" approach is safe. >> >> True. Sounds like a plan :) > > Actually, thinking about this a bit more, won't either approach be > subject to the same issue from a slow/swapping server? > > Regardless of which code adjusts the timestamp, there is still the > chance that the query will run before the data is available in > history. > > E.g., history writes a bunch of entries at 12345, but due to > swapping, it's not actually available in the history backend until > 12350. In the meantime, at 12347 the activesync code is querying for > changes (so it is asking up until lets say 12346). The changes will > still never be picked up because the next activesync query will have > ts_from of 12346. >
Attachment
Watch this ticket