Re: [SyncEvolution] SyncEvolution and N900 with Evolution - Sporadic addressbook update oddness

Forgot to mention evolution version. Ubuntu 10.04 is using evolution version 2.28.3 On Sat, 2010-05-22 at 11:24 -0400, Jeff wrote: > Update: In short, I'm still seeing this problem - perhaps I have > something configured wrong? > > Since Patrick indicated he has been testing with a newer version of > evolution (slightly) and not seen this issue, I thought I would test > with a newer version. > > New test setup: > > Virtual machine running latest Ubuntu (10.04) > Inside the VM - running > Maemo SDK (March 22) Release including scratchbox etc > > Syncevolution on desktop (VM) version 1.0beta3 compiled > from scratch > with just --disable-shared and --enable-dbus-service > I did NOT use --disable-ebook and --disable-ecal > (I wish to sync directly with evolution) > > Syncevolution 0.9 on Maemo SDK N900 side (from repository) > > > I did some quick smoke tests on this environment and confirmed that > sync in both directions for todo, memos, contact and calendar appears to > be working. > > The testing I'm doing write now is very nary and relates to confirming > that IM contact info is preserved and managed as expected when IM > accounts are created/removed on the N900. So far these tests have only > involved AIM and Yahoo IM accounts (so it's possible issues that come up > are unique to these fields..by chance) > > For the smoke tests I have tested both syncs involving initial person > contact creation on the desktop and on the N900. Seems to work fine.. > > The process that failed most recently was pretty simple and puzzles > me...(which is why despite successful syncs in the past I wonder if > there is still something wrong in my configuration files) > > I did notice I have no evolution source lines in my config - as I > understand it, for a direct sync I don't need them... > > Here is what I did (the entire brief session yesterday) > > Startup VM > Run script to start syncevo-dbus and sync-httpserver > Startup Maemo SDK N900 emulation > N900 side: delete and re-create IM accounts > (seems to be only way to control online/offline state > of IM inside the N900 emulator) > Desktop side: create contact called "Omar LastName" > N900 side: Sync using N900 gui > N900 side: Confirm contact visible on N900, it was > N900 side: Merge Omar contact with an AOL AIM IM contact > N900 side: Sync using N900 gui > Desktop side: Look at Omar contact and notice that the IM info was not > not updated > Desktop side: Restart evolution without using killev to kill EDS, still > still not visible > Desktop side: Restart evolution by shutting it down, killev, > launching again > Notice: AIM contact field is now visble and up-to-date > > While looking at the logs for the two syncs I noticed a few things - not > sure which (if any) are significant in this case. > > + The first sync correctly had detailed information on the new contact > (I only entered first and last name) and about the sync process - looked > fine > > + The second sync contained no sync details (perhaps it didn't think > anything changed?...yet EDS did receive the info ....) > > addressbook.before.ini and addressbook.after.ini > Noted: files appear identical > Noted: the rev field for some items has the form > 20100516T214422Z and for others looks like this > 2010-05-22T01:56:27Z > > Q: Is that ok? Seems to me rev strings should be > consistent format. DO I have a locale issue > perhaps? (the N900 emulator uses UK locale > because it's the only one that seems to work > the desktop in the VM is USA locale) The N900 > applet shows the same as desktop so time is correct > but could perhaps be some other locale issue? > > Q: These probably should have changed... > So I wonder why it didn't think any items changed > > > Looked at the incoming and outgoing files...don't mean much to me > nothing popped out... > > Examined main.html syncevolution log and noted the following lines > > # [2010-05-21 21:57:59.124] addressbook: testState=FALSE - expected > state>='sync_set_ready', found state=='idle' > # [2010-05-21 21:57:59.124] todo: testState=FALSE - expected > state>='sync_set_ready', found state=='idle' > # [2010-05-21 21:57:59.124] memo: testState=FALSE - expected > state>='sync_set_ready', found state=='idle' > # [2010-05-21 21:57:59.124] calendar: testState=FALSE - expected > state>='sync_set_ready', found state=='idle' > > Q: Is this normal, or does it perhaps mean I have not > configured something properly? > > > I have set loglevel to 4 and can send the desktop logs, if needed, but > would rather not as I am testing with my actual contact data. > If needed, perhaps I could send them direct to Patrick rather than > attaching them to a bug. > > This still feels like a configuration issue but most syncs seem fine... > > Any ideas? > > > On Sun, 2010-05-09 at 13:46 +0200, Patrick Ohly wrote: > > On Sun, 2010-05-09 at 03:25 +0100, Jeff wrote: > > > I am currently testing Contact migration. I expect simple > > > add/update/remove and conflict handling will work fine so chose to start > > > with testing the IM fields. There is an existing bug related to > > > some of the non-standard fields used by Maemo to track bindings between > > > IM contacts and people contacts For what it's worth, I don't think that > > > bug is still a bug from what I can tell - I plan to review it and test > > > it more though. > > > > I'm not sure whether this bug is still relevant in the N900. > > > > > Question: > > > > > > The wiki (http://syncevolution.org/wiki/linux-desktop) says > > > syncevolution SHOULD work with Evolution 2.24 and later.... > > > > > > Has anyone tried it with version 2.26.x ? > > > > I'm only using it with 2.28. > > > > > The reason I ask is, the symptoms I'm seeing almost seem like they could > > > be caused by a bug in the evolution ui losing it's link to the evolution > > > data server. > > > > That is mostly likely the reason, given that restarting Evolution fixes > > the problem. > > > > > My current thought is to try upgrading evolution (and of course it's > > > data server) to the latest (version 2.6.30)...this might pose other > > > problems if (as the wiki mentions) the EDS is no longer backward > > > compatible.... > > > > There haven't been relevant changes recently, so migrating should be > > safe. But note that you may not be able to fall back unless you copy > > your .evolution directory first and restore it before falling back. > >