Re: [Opensync-users] Partial success on Nokia 6630

[Armin B. <arm...@de...>]

Conrad Wolf wrote:
> On Friday, 2. June 2006 21:39, Armin Bauer wrote:
>> Conrad Wolf wrote:
>>>>> One other problem occured when I imported my real contact list (all=

>>>>> umlauts substituted) to evolution to sync it to my phone. The sync
>>>>> started okey and the phone display was showing a progress bar and t=
he
>>>>> number of transferred contacts, but suddenly after about 30 of 150
>>>>> contacts the sync stoped. Sometimes this occured at entry 27 someti=
mes
>>>>> at 30 and so on. I had the impression that this coincides with the
>>>>> following line appearing on the CL:
>>>>>
>>>>> I/O error : Attempt to load network entity
>>>>> http://www.syncml.org/docs/syncml_represent_v11_20020213.dtd
>>>>>
>>>>> After a while tons of error messages of this type appear:
>>>>>
>>>>> Error writing entry pas-id-445FBF6C00000060 to member 2: Timeout wh=
ile
>>>>> waiting for a reply to message "COMMIT_CHANGE"
>>>>> Mapping Write Error: Timeout while waiting for a reply to
>>>>> message "COMMIT_CHANGE"
>>>>>
>>>>> The syncml-obex-client trace ends with the following lines:
>>>>>
>>>>> [1147126085.317415]	<<<<<<<  client_message_handler
>>>>> [1147126085.755768]	>>>>>>>  _manager_event(0x80831e0, 6, 0x80c9768=
,
>>>>> (nil), 0x8082500)
>>>>> [1147126085.755848]	<<<<<<<  _manager_event
>>>>> [1147126085.755888]	>>>>>>>  _smlManagerEventFree(0x858dc98)
>>>>> [1147126085.755927]		>>>>>>>  smlSessionUnref(0x80c9768)
>>>>> [1147126085.755965]		<<<<<<<  smlSessionUnref
>>>>> [1147126085.756003]	<<<<<<<  _smlManagerEventFree
>>>>> [1147126091.76960]	>>>>>>>  _manager_event(0x80831e0, 7, 0x80c9768,=

>>>>> 0x8667ef8, 0x8082500)
>>>>> [1147126091.77014]		There was an error in the session 219: Received=
 a
>>>>> status after commands
>>>>> [1147126091.77071]	<--- ERROR --- _manager_event: Received a status=

>>>>> after commands
>>>>> [1147126091.77110]	>>>>>>>  _smlManagerEventFree(0x820e7e0)
>>>>> [1147126091.77149]		>>>>>>>  smlSessionUnref(0x80c9768)
>>>>> [1147126091.77187]		<<<<<<<  smlSessionUnref
>>>>> [1147126091.77227]	<<<<<<<  _smlManagerEventFree
>>>>>
>>>>> No idea what happened here, but it's not nice. It never happened wh=
en I
>>>>> did the testing with only a few entries.
>>>>>
>>>>> I've taken traces of all the above syncs, but I decided not to atta=
ch
>>>>> them to this mail. I'll send em off-list.
>>>> the syncml protocols clearly states that no status has to be sent af=
ter
>>>> a command has been sent. but some devices simply ignore the standard=
 and
>>>> do this anyways.
>>>>
>>>> there is a option in the config AllowLateStatus which you can set to=
 1
>>>> to allow such behaviour.
>>>>
>>>> Best Regards,
>>>> Armin
>>> Hi Armin,
>>>
>>> I put
>>>
>>> <AllowLateStatus>1</AllowLateStatus>
>>>
>>> into my syncml-obex-client.conf, but that didn't help. I also split m=
y
>>> contact list into 2 parts and tried to sync these parts separately to=

>>> check whether I have a corrupt entry, but the sync freezes with both
>>> parts at something between entry 25 to 30. The freezing of the progre=
ss
>>> bar and the counting of transferred entries on the phone always occur=
s
>>> when
>>>
>>> I/O error : Attempt to load network entity
>>> http://www.syncml.org/docs/syncml_represent_v11_20020213.dtd
>>>
>>> appears on the command line. I read somewhere on the list that one co=
uld
>>> ignore this error message. Is that definite or is there a chance that=
 it
>>> is correlated to the freezing of the sync?
>> this is definitely not related to the freeze (actually i finally manag=
ed
>> to squelch the error message).
>>
>> allowLateStatus is not needed anymore (it is now enabled by default).
>>
>> i uploaded some fixes to the svn lately. so maybe it works for you now=
=2E
>>
>> Armin
>=20
>=20
> I checked out the new version from svn a few days ago and the problem w=
ent=20
> away :-). Thanks a lot!!!
>=20
> I still have the problem though that sometimes some random entry gives =
me=20
> a 'Error 500' message during the sync, while the same entry works at ot=
her=20
> times. I'll do some more testing on that issue and then report again.

This is a known "bug". A lot of phones dont understand the events that
are generated by applications like evolution. the solution to this will
be, that opensync will convert such events internally to something that
is understood by the phones.

Writing this converter is a lot of work, thats why this isnt fixed yet.

Armin

>=20
> Aside from that everything seems to work fine now. Good work!
>=20
> Conrad.
>=20