-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 09/07/12 19:22, Patrick Ohly wrote:
> On Mon, 2012-07-09 at 13:17 +1200, Jane Atkinson wrote:
>> On 09/07/12 07:38, Patrick Ohly wrote:
>>> On Mon, 2012-07-09 at 02:09 +1200, Jane Atkinson wrote:
>>>> On 09/07/12 01:32, Patrick Ohly wrote:
>>>>> On Sat, 2012-07-07 at 16:24 +1200, Jane Atkinson wrote:
>>>>>> Can you quote the output? What kind of changes did you
>>>>>> expect to
>>>>> get synced?
>>>> 
>>>> It's a day or two now since I did this, so it's from memory.
>>>> I had changed the end time of an event - something I've done 
>>>> regularly in the past with no problems. The messages in the 
>>>> terminal were exactly what I'd expect to see if no changes
>>>> had been made.
>>>> 
>>>> Only a calendar.before directory was created - nothing else.
>>>> And a log file, of course.
>>> 
>>> Then the sync process crashed without completing normally. 
>>> Otherwise there would be a calendar.after directory. Can you
>>> send me that syncevolution-log.html file?
>>> 
>>> Can you reproduce the problem? If yes, then try to capture a
>>> stack backtrace of the crash, like this: - killall
>>> syncevo-dbus-server - SYNCEVOLUTION_SYNC_DELAY=60
>>> /usr/libexec/syncevo-dbus-server & - in another shell, run the
>>> command line - in a third shell, run "gdb -p `pidof
>>> syncevo-dbus-helper`" followed by "cont" (when debugger has 
>>> attached) and "thread apply all bt" (once/if it crashed)
>>> 
>> 
>> It's a bit more complicated than I first thought.
>> 
>> When I originally set up on this installation, I simply copied
>> the config files from a working install (same machine, different 
>> partition). That config resulted in the logfile I sent earlier.
> 
> Copying the entire config tree is problematic if you want to
> continue using both old and new config. To the phone it will look
> the same computer. The data storage itself is the same, but
> additional meta data stored on the computer is not, which will
> cause problems during incremental syncs.
> 
Good point. I'll discard the copied config directories.

> For normal configs, this can be fixed by updating the "deviceId" 
> property after copying a config. However, in this case the data is 
> shared between computers (assuming that the WebDAV server is
> external) which would continue to cause problems when switching
> between computers as sync owner (new item on WebDAV will be sent as
> new by computer A, and again by computer B).
> 
>> To do the tests you asked for, I thought I'd set up a fresh
>> config from scratch. No problems while I setup the target-config
>> 
>> Then: ~$ syncevolution --configure \ --template
>> SyncEvolution_Client \ syncURL=local://@webdav \ username= \ 
>> password= \ webdav \ calendar addressbook
>> 
>> [INFO] addressbook: looking for databases... [INFO] addressbook:
>> no backend available [ERROR] error code from SyncEvolution fatal
>> error (local, status 10500): addressbook: no backend available
>> 
>> (If I remove "addressbook" from the last line, I simply get the
>> same error message regarding "calendar".)
> 
> That's normal. The "webdav" config here would be for syncing your 
> default local databases, but because the EDS backend isn't working
> on an installation without Evolution, such a config cannot be
> created.
> 
> KDE users have to explicitly configure the @default 
> addressbook/calendar/... source to use Akonadi.
> 
> You want to configure a phone to sync with WebDAV, so you don't
> need the "webdav" config.

So I can use a "default" config here. Does that mean that sync-ui can
be used for running refresh-from-local and slow-sync if needed? Or is
that only for Evolution-based configurations?
> 
>> ~$ syncevolution --configure \
>>> syncURL=obex-bt://xxxxxxxxxxx \ --template Nokia_N900 \ 
>>> ja_6120c@webdav
>> [INFO] addressbook: looking for databases... [INFO] addressbook:
>> backend failed [INFO] calendar: looking for databases... [INFO]
>> calendar: backend failed [INFO] calendar+todo: looking for
>> databases... [INFO] calendar+todo: backend failed [INFO] memo:
>> looking for databases... [INFO] memo: okay [INFO] todo: looking
>> for databases... [INFO] todo: backend failed
>> 
>> After adding the login and password details to the files, I tried
>> a sync.
>> 
>> Only the memos (on Radicale) synced.
> 
> Did you configure the "backend", "database", "databaseUser", 
> "databasePassword" of calendar/addressbook/todo first?
> 

Yes I configured those before trying to sync anything.

>> On trying the set of commands you requested I get:
>> 
>> ~$ SYNCEVOLUTION_SYNC_DELAY=60 /usr/libexec/syncevo-dbus-server
>> & [1] 5506 ~$ [WARNING syncevo-dbus-server] libnotify: Failed to
>> connect to proxy [INFO syncevo-dbus-server] ready to run
>> 
>> 
>> Next thing was to remove the new config files and put the old
>> ones back.
>> 
>> On trying to run the set of commands I get:
>> 
>> (first shell) ~$ SYNCEVOLUTION_SYNC_DELAY=60
>> /usr/libexec/syncevo-dbus-server & [1] 1282 ~$ [ERROR
>> syncevo-dbus-server] dbus_get_bus_connection() failed - server
>> already running?
>> 
>> (second shell) ~$ syncevolution ja_6120c@webdav [INFO] Server
>> sending SAN [INFO] calendar: starting slow sync, two-way (peer is
>> client) [INFO] todo: starting slow sync, two-way (peer is
>> client) [INFO] addressbook: starting slow sync, two-way (peer is
>> client) [INFO] memo: starting slow sync, two-way (peer is
>> client) [INFO] creating complete data backup of source calendar
>> before sync (enabled with dumpData and needed for printChanges)
>> 
>> (third shell) ~$ gdb -p `pidof syncevo-dbus-helper` gdb: option
>> '--p' requires an argument Use `gdb --help' for a complete list
>> of options.
> 
> You didn't kill an already syncevo-dbus-server, so the one which
> served the command line's request ran without the artificial 60
> second delay, which prevented catching syncevo-dbus-helper before
> it crashed.
> 
> The log that you sent me confirms that it crashes on the first use
> of libneon for the "calendar" source.
> 
> The following steps might make it easier to get a stack backtrace: 
> $ SYNCEVOLUTION_DEBUG=1 gdb syncevolution gdb> run --daemon=no
> --print-items ja_6120c@webdav calendar ... gdb> thread apply all
> bt
> 

I'll try running these new commands shortly.

>> Incidentally, I'm beginning to wonder if this has something to do
>> with DCS, since the only thing that worked wasn't a DCS
>> connection.
> 
> What is DCS?
> 
Sorry, Darwin CalendarServer. The Radicale backend was OK for syncing
but the CalendarServer backend (addressbook, todo, calendar) isn't
working in this case.

Jane Atkinson


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJP+pJyAAoJEERzSJEx033jnj8H/00rJ6dGP88fTVFAHhbxEpYp
UprkFDdJLeDtIktrrAtcyNk81JcGgNsmZXaVxsJ8U0esTzfdgGhcENh+8IW0wiL4
fIIk7WbW8LFEL5jy59khco8+SlMPFPm5pWVRT+HUiHYx1EknXusut6WCKdJMtkv1
SFyPPZlu+/8AnUuZH8DaI40uzKHjngWUo2uP69P+ER7fyGU28HPZybDIV7gZr3xE
m56xadrDp2QHnflz2opfBdSf5RfiiteR04ejnr2PpiFsTx7W7Ubn9zftPGwxUn9k
p2aWNseAncyoaCyIamsdNb6bgwsyzgtfi2ZdVa9xlDJAI4uaR9iF4rykDTSiIkA=
=a35j
-----END PGP SIGNATURE-----
_______________________________________________
SyncEvolution mailing list
[email protected]
http://lists.syncevolution.org/listinfo/syncevolution

Reply via email to