[SyncEvolution] Moving the dbus server to GIO's GDBus
by Chris Kühl
Hi all,
One of the things I was to look into was whether, and how, I could
port the syncevo-dbus-server to use GIO's GDBus. Currently
syncevo-dbus-server uses an in-tree copy of a dbus library that also
goes by the name gdbus but, I think, originated in Bluez.
We also currently wrap the Bluez gdbus library in C++ classes and the
intention is to do the same for GIO's GDBus. This has the advantage
that the interface the server uses would remain unchanged.
One of the requirements that Patrick gave me was that we need to
remain compatible with Debian stable and Ubuntu LTS. However, both of
these have GLib 2.24 packages which is one version before GLib put
GDBus into GIO. We then decided to look into possibly using an in-tree
copy of GIO's GDBus until we able to make GLib >= 2.26 a requirement.
This, however, won't really work since the GDBus that landed in GLib
2.26 makes use of some new features of 2.26.
It seems to me as if we've got 2 choices right now...
1) Postpone this and wait till we can bump the requirement to GLib >=
2.26 to do this work.
2) Go forward with this and use it when ./configure detects an
adequate GLib version.
I'm thinking the second option sounds like the way to go. Let me know
your opinions on this.
Cheers,
Chris
10 years, 9 months
[SyncEvolution] testing ActiveSync SyncEvolution backend
by Patrick Ohly
Hello!
Beware if you are running Client::Source::eas_contact/event, I made a
change which will require config changes when compiling from source or
updating to the upcoming SyncEvolution 1.1.99.6.
>From the activesyncd README update:
commit 6b1cf2b557e6eb3c9df9c454d0388932d10b631a
Author: Patrick Ohly <patrick.ohly(a)intel.com>
Date: Mon Aug 15 17:35:53 2011 +0200
SyncEvolution: use target-config@client-test-exchange for Client::Source
SyncEvolution has changed how it finds config in the commit quoted
below. Adapted README accordingly.
commit 24efbe11dbae9394d4befb2b91a25cd403907ec2
Author: Patrick Ohly <patrick.ohly(a)intel.com>
Date: Mon Aug 15 17:32:01 2011 +0200
testing: use target-config@client-test-$CLIENT_TEST_SERVER for Client::Source
For testing sources instantiated via SyncSource.cpp (for example, EDS)
and client-test-app.cpp (ActiveSync) append the name of the server set
via the CLIENT_TEST_SERVER env variable to target-config@client-test,
separated by a hyphen.
This allows configuring testing differently for different servers.
Was already done for CalDAV sources, inside WebDAVSourceRegister.cpp.
--
Best Regards, Patrick Ohly
The content of this message is my personal opinion only and although
I am an employee of Intel, the statements I make here in no way
represent Intel's position on the issue, nor am I authorized to speak
on behalf of Intel on this matter.
10 years, 9 months
[SyncEvolution] syncing only possible in LAN
by Roman Morawek
Hello,
I am using syncevolution on Maemo 5 (N900) and Harmattan (N950, in
development). I have a Funambol server in my local LAN, which is also
accessible via a dyndns domain name (the server has no static IP).
Syncing inside my local LAN works well. However, syncing via Internet
aborts with failures:
CurlTransport Failure: couldn't connect to host
But the server/port is reachable. A wget
http://epia.selfip.net:8080/funambol/ds from the N900 works fine.
In the log files I recognized the following, which I expect as the
source of the problem:
Source (Remote party): 'http://epia.selfip.net:8080/funambol/ds'
Response to be sent to URI='http://epia:8080/funambol/ds;jsessionid=...
Thus, the response is not sent to the full qualified host name but only
to the server name "epia". This works fine within the local LAN but does
not work via Internet connection.
Where does this RespURI come from?
Is the syncevolution client configured wrongly?
Or is the server somehow setup wrong (same problem on N900 and N950...)?
Thanks a lot for your ideas,
regards,
Roman
10 years, 9 months
[SyncEvolution] unsupported transport type?
by Klaus Weidenbach
Hello,
I finally tried to synchronize my Horde server with KDE PIM and found
SyncEvolution as an interesting solution.
First I have followed the instructions in HACKING and used
git://git.moblin.org/syncevolution.git to compile from source. I started
with addressbook only and I got it synchronize to a folder with files
successfully. But I had read actually that it should also work with Akonadi
directly and found then the GIT repository at git://gitorious.org/meego-
middleware/syncevolution.git that seems to support it.
I have started with a new configuration, but I could not get the
synchronisation to work anymore. I always get the error message:
"unsupported transport type is specified in the configuration".
The horde server is accessed trough https:// and I use CURL in
syncevolution. It seems as if it never gets into "if (!gmainloop) {" inside
the function SyncContext::createTransportAgent() in
src/syncevo/SyncContext.cpp. I tried to change that and I got much more
debug output, but failed with other error messages later.
Do you have an idea what the problem could be or where I could start
debugging?
Regards,
Klaus
10 years, 9 months
Re: [SyncEvolution] Syncevolution cant find evolution data
by Patrick Ohly
On Di, 2011-08-02 at 14:39 +0200, Tobias Hägele wrote:
> Hello,
>
> Am 2011-08-02 13:21, schrieb Patrick Ohly:
> > According to the output there your configuration doesn't have any active
> > sources. How did you create it?
>
> I've created with the sync-ui.
> I've done it now on another machine with the same Bluetooth-dongle&Phone.
> There i have seen, that i can select wich parts of my phones should be
> Synchronised.
> On My main PC, this checkboxes wasnt shown.
[...]
> After i do this, following message appesars:
>
> #####
> [INFO] memo: inactive
> [INFO] todo: inactive
> [INFO] calendar: inactive
> [INFO] Server sending SAN
> [ERROR] Error allocating address book
That explains why the checkboxes can't be selected: unusable sources
aren't offered.
The reason for this is probably a mismatch between your libecal/ebook
and the running Evolution Data Server that it talks to.
Your earlier emails says that you have libebook-1.2.so.9 from EDS 2.30
installed on Ubuntu 11.04. That package is obsolete, it is no longer
part of Ubuntu 11.04.
>From where did you install SyncEvolution? What do the following commands
say?
apt-cache policy syncevolution
apt-cache policy libebook1.2-9
The way to get a working system again is to uninstall libebook1.2-9 and
libecal1.2-7. If your copy of SyncEvolution depends on them, then you
need to find a different .deb for it, for example via
http://syncevolution.org/documentation/installation
--
Best Regards, Patrick Ohly
The content of this message is my personal opinion only and although
I am an employee of Intel, the statements I make here in no way
represent Intel's position on the issue, nor am I authorized to speak
on behalf of Intel on this matter.
10 years, 9 months
Re: [SyncEvolution] Syncevolution cant find evolution data
by Tobias Hägele
Am 2011-08-10 11:12, schrieb Patrick Ohly:
> On Mi, 2011-08-10 at 10:39 +0200, Tobias Hägele wrote:
>> is this right:
>> $ syncevolution --daemon=no loglevel=4 blubber calendar
>> then
>> $ syncevolution blubber
>
> No, this needs to be:
> $ syncevolution --daemon=no --run loglevel=4 blubber calendar
>
> The first command should have complained like this:
> [usage information]
> ERROR: Properties specified, but neither '--configure' nor '--run' - what did you want?
>
> Didn't it do that? It does when I try it.
Sorry, typo...
But it ends with:
Speicherzugriffsfehler
>> (syncevo-dbus-server:3054): GLib-CRITICAL **: g_variant_new_string:
>> assertion `g_utf8_validate (string, -1, NULL)' failed
>
> That last line in the log file confirms that there is a character
> encoding problem.
>
> Thanks, I can reproduce that now. No need for further testing at the
> moment.
Ok. And what can i Do now?
Delete all entries with an Umlaut?
10 years, 9 months
[SyncEvolution] DTD version 1.1
by fabien schmidt
Hello,
Im using syncevolution 1.1.1. I just want to force use of syncml 1.1 and
basic mode in authentication .
How can i make it?
thanx
10 years, 9 months
[SyncEvolution] EDS backend: Support for the new ECalClient API in EDS 3.2
by Dumez, Christophe
Hi,
Please find attached a patch to port the EDS calendar backend from the
ECal API to ECalClient when using EDS >= v3.2.
A view is used to list the components but no optimization is made yet
because e_cal_client_view_set_fields_of_interest() is not ready for
it.
***
[PATCH] Support for the new ECalClient API in EDS 3.2
SyncEvolution EDS backend currently uses the ECal API which is
deprecated as of EDS v3.2 and replaced by the new ECalClient API.
This patch brings support for this new ECalClient API (when EDS
3.2 is detected at compile time).
The new code based on the ECalClient API will enable a few
optimizations, especially the partial fetching of components to
avoid useless DBus trafic. This results in better component change
tracking by fetching only the UID/RID/REV attributes for each
component.
Note: The new code uses the new GLibSupport header instead of the
deprecated SmartPointer (eptr).
***
The eds_event tests are passing:
Client::Source::eds_event::testOpen okay
Client::Source::eds_event::testIterateTwice okay
Client::Source::eds_event::testSimpleInsert okay
Client::Source::eds_event::testLocalDeleteAll okay
Client::Source::eds_event::testComplexInsert okay
Client::Source::eds_event::testLocalUpdate okay
Client::Source::eds_event::testChanges okay
Client::Source::eds_event::testImport okay
Client::Source::eds_event::testImportDelete okay
Client::Source::eds_event::testManyChanges okay
Client::Source::eds_event::testLinkedItemsParent okay
Client::Source::eds_event::testLinkedItemsChild okay
Client::Source::eds_event::testLinkedItemsParentChild okay
Client::Source::eds_event::testLinkedItemsChildParent okay
Client::Source::eds_event::testLinkedItemsChildChangesParent okay
Client::Source::eds_event::testLinkedItemsRemoveParentFirst okay
Client::Source::eds_event::testLinkedItemsRemoveNormal okay
Client::Source::eds_event::testLinkedItemsInsertParentTwice okay
Client::Source::eds_event::testLinkedItemsInsertChildTwice okay
Client::Source::eds_event::testLinkedItemsParentUpdate okay
Client::Source::eds_event::testLinkedItemsUpdateChild okay
Client::Source::eds_event::testLinkedItemsInsertBothUpdateChild okay
Client::Source::eds_event::testLinkedItemsInsertBothUpdateParent okay
OK (23)
***
Note that the EDS patches attached to
https://bugzilla.gnome.org/show_bug.cgi?id=655986 are required for the
tests to pass (with or without my changes).
Also note that while porting to the ECalClient API, I found the
following bug in EDS (several tests don't pass without this fix):
https://bugzilla.gnome.org/show_bug.cgi?id=656207
Kr,
--
Dr. Christophe Dumez
Linux Software Engineer
Intel Finland Oy - Open Source Technology Center
10 years, 9 months
[SyncEvolution] EDS backend: Support for the new EBookClient API in EDS 3.2
by Dumez, Christophe
Hi,
Please find attached a patch for SyncEvolution bringing support for
the new EBookClient API present in EDS 3.2.
***
Support for the new EBookClient API in EDS 3.2
SyncEvolution EDS backend currently uses the EBook API which is
deprecated as of EDS v3.2 and replaced by the new EBookClient API.
This patch brings support for this new EBookClient API (when EDS
3.2 is detected at compile time).
The new code based on the EBookClient API enables a few
optimizations, especially the partial fetching of contacts to
avoid useless DBus trafic. This results in better contact change
tracking by fetching only the UID/REV attributes for each contact.
Note: The new code uses the new GLibSupport header instead of the
deprecated SmartPointer (eptr).
***
Note that this patch only alters the contacts code (I think the
calendar code should be handled in a separate patch). I also haven't
optimized the EvolutionContactSource::isEmpty() yet and I'm planning
to send another patch for this (the idea would be to use a view, ask
only for the UIDs and stop the view as soon as we get a objectsAdded
signal, this way we don't fetch the whole contacts and hopefully we
don't even fetch all the UIDs, just the first burst).
I'm awaiting review for this patch and for the
EvolutionContactSource::isEmpty() change proposal before I move onto
the next tasks.
Kr,
--
Dr. Christophe Dumez
Linux Software Engineer
Intel Finland Oy - Open Source Technology Center
10 years, 9 months
Re: [SyncEvolution] SyncEvolution on Harmattan
by Roman Morawek
Hello Ove,
On 24.07.2011 00:28, Ove K?ven wrote:
> It looks like I can't compile SyncEvolution on this yet, because there's
> some build dependencies missing, Boost in particular. If you know of a
> libboost package for Harmattan, let me know. (There's also no libneon,
> but that's only needed for CardDAV/CalDAV, and I think you were only
> interested in SyncML.)
Boost seems to be available in the following repository:
http://repo.pub.meego.com/home%3a/ph5%3a/boost/Harmattan/armel/
See also this thread:
http://forum.meego.com/showthread.php?t=4014
> And the simulator seems to have no contacts or calendar application, so
> even if I could build the SyncEvolution package, I would not be able to
> check how well it would work and interoperate with the native PIM
The N950 device program is extended with additional devices available:
http://wiki.meego.com/Community_Office/Community_device_program/Nokia/N95...
I would be glad if you get one for testing syncevolution on it.
Best regards,
Roman
10 years, 9 months
