Re: [SyncEvolution] automated testing of ActiveSync

On Fr, 2011-07-15 at 13:12 +0100, Andy Gould wrote: > On 15/07/11 12:58, Patrick Ohly wrote: >> On Fr, 2011-07-15 at 12:44 +0100, Andy Gould wrote: >>> On 15/07/11 11:42, Patrick Ohly wrote: >>>> On Fr, 2011-07-15 at 11:21 +0100, Andy Gould wrote: >>> Hi Patrick, >>> I may have misunderstood the point of the 2 activsyncd accounts ( Foo, >>> and Foo_B in gconf) - can you clarify for me.... should they point to >>> different Exchange accounts ( andy(a)cstylianou.com and >>> andy_B(a)cstylianou.com) - or the same one? >> The same one, but with different device IDs. >> >> So if you configure target-config@client-test with username=Foo, then >> you need accounts called "Foo" and "Foo_B" in gconf. Both must have the >> same username/password/URL. But because they are different local >> accounts, my expectation is that activesyncd treats them completely >> separate and thus assigns different device IDs. To the Exchange servers >> this should look like two devices who happen to contact it with the same >> IP address. >> > yeah, thats what I thought unfortunately, your assumption about the way > that the accounts are handled was not correct. At the moment, the > device id is exactly that - fixed for the device, so I needed to > allocate account level ids instead, I have done that - but they are not > being used, because the daemon is too clever for its own good - and > recognises that the 2nd account is actually pointing at the same > Exchange account, and grabs the existing connection for it. One more > modification, and hopefully that will sort it.... getting there slowly. How has this been resolved? Is it necessary to manually request real unique device IDs in gconf for the second Foo_B account?