From: Stef E. <st...@ep...> - 2006-03-05 19:53:19
|
Hello, I just noticed the front page of the opensync.org website advises everyone to test the SyncML plugin. After fighting a bit with makefiles in libsyncml in order to not get any HTTP support, I managed to compile everything. I am really very exited about this release and thank all of you very much for it, I can't wait to have sync support for my phone! But in the end I'm stuck: it looks as if only obex over USB or HTTP is supported, but I need bluetooth in order to test. I've tried opening an rfcomm connection to the SyncML Client channel on my phone and bind it to /dev/rfcomm0, then attempting to use the BLUETOOTH transport of libsyncml by giving it the path to the device, but that did not work. It looks as if making the syncml plugin support the bluetooth libsyncml transport is easy enough, but there doesn't seem to be the proper support for bluetooth in libsyncml yet, or am I wrong ? Do you have any release plan for bluetooth support so that I can test this great plugin ? Thanks. --=20 St=C3=A9phane Epardaud |
From: Armin B. <arm...@de...> - 2006-03-06 01:37:45
Attachments:
signature.asc
|
Stef Epardaud wrote: > Hello, > > I just noticed the front page of the opensync.org website advises > everyone to test the SyncML plugin. After fighting a bit with makefiles > in libsyncml in order to not get any HTTP support, I managed to compile > everything. I am really very exited about this release and thank all of > you very much for it, I can't wait to have sync support for my phone! > > But in the end I'm stuck: it looks as if only obex over USB or HTTP is > supported, but I need bluetooth in order to test. I've tried opening an > rfcomm connection to the SyncML Client channel on my phone and bind it > to /dev/rfcomm0, then attempting to use the BLUETOOTH transport of > libsyncml by giving it the path to the device, but that did not work. > It looks as if making the syncml plugin support the bluetooth libsyncml > transport is easy enough, but there doesn't seem to be the proper > support for bluetooth in libsyncml yet, or am I wrong ? > Do you have any release plan for bluetooth support so that I can test > this great plugin ? > Thanks. Hi Stef, adding bluetooth support to libsyncml is fairly trivial. The only thing that has to be done is to expose the configuration through the config if the obex transport and add the appropiate connection code (which can be copied from obexftp). After that, libsyncml can handle bluetooth. Adding support for irda would be the same procedure. Ill see if i get around to adding this the next days. I will drop a mail to the mailing list once it is finished. Armin |
From: Anders H. <an...@gi...> - 2006-03-06 13:54:22
Attachments:
patch-libsyncml-bt.diff
|
Armin Bauer wrote: > adding bluetooth support to libsyncml is fairly trivial. The only thing > that has to be done is to expose the configuration through the config if > the obex transport and add the appropiate connection code (which can be > copied from obexftp). After that, libsyncml can handle bluetooth. Adding > support for irda would be the same procedure. Attached is a patch which does this. I tested it with the syncml-obex-client and my 6680. And it works :) I have not tried the syncml-plugin yet. But it probably needs some easy configuration patching too. To find your bluetooth address run: hcitool scan To find your bluetooth channel run: sdptool browse <bt-addr> Look for SyncMLClient, on my 6680 it was channel 11. Instead of "-u <usb-id>" use "-b <bt-addr> <bt-channel>". Armin: I just reused some of the available config variables. Tell me if you want me to change that so the names are more logic. Armin2: I still get hangups when doing a slow-sync. Can I send you some more dumps for debugging? Anders |
From: Armin B. <arm...@de...> - 2006-03-06 14:01:25
Attachments:
signature.asc
|
Anders Hermansen wrote: > Armin Bauer wrote: > >>adding bluetooth support to libsyncml is fairly trivial. The only thing >>that has to be done is to expose the configuration through the config if >>the obex transport and add the appropiate connection code (which can be >>copied from obexftp). After that, libsyncml can handle bluetooth. Adding >>support for irda would be the same procedure. > > > Attached is a patch which does this. I tested it with the > syncml-obex-client and my 6680. And it works :) > cool :) i will try your patch right away and then commit it to the repository. thanks! > I have not tried the syncml-plugin yet. But it probably needs some easy > configuration patching too. > > To find your bluetooth address run: > hcitool scan > > To find your bluetooth channel run: > sdptool browse <bt-addr> > > Look for SyncMLClient, on my 6680 it was channel 11. > > Instead of "-u <usb-id>" use "-b <bt-addr> <bt-channel>". > > Armin: I just reused some of the available config variables. Tell me if > you want me to change that so the names are more logic. i think it is ok to reuse. I think the developers will understand what they mean. > > Armin2: I still get hangups when doing a slow-sync. Can I send you some > more dumps for debugging? Sure! traces are always welcome. Armin |
From: Norm D. <no...@dr...> - 2006-03-08 01:02:44
|
On Mon, 2006-03-06 at 14:54 +0100, Anders Hermansen wrote: > Armin Bauer wrote: > > adding bluetooth support to libsyncml is fairly trivial. The only thing > > that has to be done is to expose the configuration through the config if > > the obex transport and add the appropiate connection code (which can be > > copied from obexftp). After that, libsyncml can handle bluetooth. Adding > > support for irda would be the same procedure. > > Attached is a patch which does this. I tested it with the > syncml-obex-client and my 6680. And it works :) > > I have not tried the syncml-plugin yet. But it probably needs some easy > configuration patching too. > I tried the patch with my Nokia 6620: syncml-obex-client -b 00:02:EE:6D:62:FC 11 --sync vcard /datasync/Contacts --wbxml --identifier "PC Suite" connection with device succeeded The problem is this locks up my phone -- it says sync incomplete on the phone and is stuck there. Question -- Armin, is this the same bug you mentioned before that you haven't hunted down? Question -- Is there a way to make the output more verbose or trace? Question -- is the file location in my command line (/datasync/Contacts) suppose to match what the phone uses? I only 'touched' the file in that directory and doesn't contain anything -- should it? Thanks! Norm |
From: Stef E. <st...@ep...> - 2006-03-10 19:03:45
Attachments:
logs-opensync.tgz
|
Yeah! We're getting somewhere, thanks to your patch :) But unfortunately I get errors :( I've created a sync pair of file and multisync and I'm using this config: <config><interface>11</interface><path>00:13:70:D6:4E:39</path><identifie= r>PC Suite</identifier><version>1</version><wbxml>1</wbxml><username></usernam= e><password></password><type>5</type><usestringtable>1</usestringtable><o= nlyreplace>0</onlyreplace><recvLimit>0</recvLimit><contact_db>Contacts</c= ontact_db><calendar_db>Agenda</calendar_db><note_db>Notes</note_db></conf= ig> for my Nokia 6230 and it fails with error 67 (is it forbidden?) I'm attaching the logs I get, not sure where to start looking.... Note that I'm not sure about the names of databases on the device or how I could know, but when I tried other names than Agenda for the calendar_db it failed with error 68 (not found?) nad my phone told me it could not sync Agenda, so I used that name. Then the error became 67 :( What can I do ? Thanks for all the help I'm really getting exited about this :) --=20 St=C3=A9phane Epardaud |
From: Ademar de S. R. Jr. <ad...@co...> - 2006-03-06 14:41:39
|
Armin Bauer wrote: > > Stef Epardaud wrote: > > > Hi Stef, > > adding bluetooth support to libsyncml is fairly trivial. The only thing > that has to be done is to expose the configuration through the config if > the obex transport and add the appropiate connection code (which can be > copied from obexftp). After that, libsyncml can handle bluetooth. Adding > support for irda would be the same procedure. > > Ill see if i get around to adding this the next days. I will drop a mail > to the mailing list once it is finished. Hi there. Please remember that BlueZ is GPL, so libsyncml must also be released under this license. Making bluetooth support optional is also a possibility. What do you think? []'s - Ademar -- Ademar de Souza Reis Jr. Mandriva Conectiva - http://www.mandriva.com.br ^[:wq! |
From: Armin B. <arm...@de...> - 2006-03-06 15:07:30
Attachments:
signature.asc
|
Ademar de Souza Reis Jr. wrote: > Armin Bauer wrote: > >>Stef Epardaud wrote: >> >> >>Hi Stef, >> >>adding bluetooth support to libsyncml is fairly trivial. The only thing >>that has to be done is to expose the configuration through the config if >>the obex transport and add the appropiate connection code (which can be >>copied from obexftp). After that, libsyncml can handle bluetooth. Adding >>support for irda would be the same procedure. >> >>Ill see if i get around to adding this the next days. I will drop a mail >>to the mailing list once it is finished. > > > Hi there. > > Please remember that BlueZ is GPL, so libsyncml must also be released > under this license. Making bluetooth support optional is also a > possibility. > Why? libsyncml is LGPL which is compatible to the GPL so the license is fine. The only implication of this is that if you want bluetooth support, you would not be able to use libsyncml with a propietary program (since this would also link against bluez). Armin > What do you think? > > []'s > - Ademar > |
From: Ademar de S. R. Jr. <ad...@co...> - 2006-03-06 16:22:25
|
Armin Bauer wrote: > > Ademar de Souza Reis Jr. wrote: >> >>Please remember that BlueZ is GPL, so libsyncml must also be released >>under this license. Making bluetooth support optional is also a >>possibility. >> > > Why? libsyncml is LGPL which is compatible to the GPL so the license is > fine. The only implication of this is that if you want bluetooth > support, you would not be able to use libsyncml with a propietary > program (since this would also link against bluez). > LGPL is compatible with GPL, but not the other way around. What I mean by "must also be released under this license" is that if you link to BlueZ, you can't use libsyncml as LGPL anymore, so either* you also release it under GPL or make bluez linkage optional at compilation time. Of course, the code can still be LGPL so that someone can recompile it manually without linking to bluez, but since libsyncml will be redistributed (by linux distros, for example) and considered a standard library, I believe it's a good idea to have the licensing terms clearly defined. * possibilities such as releasing BlueZ under LGPL, etc not cosidered for simplicity :-) Cheers, - Ademar -- Ademar de Souza Reis Jr. Mandriva Conectiva - http://www.mandriva.com.br ^[:wq! |
From: Armin B. <arm...@de...> - 2006-03-06 16:36:22
Attachments:
signature.asc
|
Ademar de Souza Reis Jr. wrote: > Armin Bauer wrote: > >>Ademar de Souza Reis Jr. wrote: >> >>>Please remember that BlueZ is GPL, so libsyncml must also be released >>>under this license. Making bluetooth support optional is also a >>>possibility. >>> >> >>Why? libsyncml is LGPL which is compatible to the GPL so the license is >>fine. The only implication of this is that if you want bluetooth >>support, you would not be able to use libsyncml with a propietary >>program (since this would also link against bluez). >> > > > LGPL is compatible with GPL, but not the other way around. What I > mean by "must also be released under this license" is that if > you link to BlueZ, you can't use libsyncml as LGPL anymore, so > either* you also release it under GPL or make bluez linkage > optional at compilation time. ok. so we mean the same thing. yes, bluez linkage is of course going to be optional. > > Of course, the code can still be LGPL so that someone can > recompile it manually without linking to bluez, but since > libsyncml will be redistributed (by linux distros, for > example) and considered a standard library, I believe > it's a good idea to have the licensing terms clearly > defined. > > * possibilities such as releasing BlueZ under LGPL, etc > not cosidered for simplicity :-) > > Cheers, > - Ademar > |
From: Michal <mi...@ci...> - 2006-03-07 10:05:11
|
Hi On Mon, 06 Mar 2006 11:52:21 -0400 "Ademar de Souza Reis Jr." <ad...@co...> wrote: > * possibilities such as releasing BlueZ under LGPL, etc > not cosidered for simplicity :-) See http://sourceforge.net/mailarchive/message.php?msg_id=14866873 -- Michal Čihař | http://cihar.com |
From: Armin B. <arm...@de...> - 2006-03-09 14:11:55
Attachments:
signature.asc
|
Norm Dressler wrote: > On Mon, 2006-03-06 at 14:54 +0100, Anders Hermansen wrote: > >>Armin Bauer wrote: >> >>>adding bluetooth support to libsyncml is fairly trivial. The only thing >>>that has to be done is to expose the configuration through the config if >>>the obex transport and add the appropiate connection code (which can be >>>copied from obexftp). After that, libsyncml can handle bluetooth. Adding >>>support for irda would be the same procedure. >> >>Attached is a patch which does this. I tested it with the >>syncml-obex-client and my 6680. And it works :) >> >>I have not tried the syncml-plugin yet. But it probably needs some easy >>configuration patching too. >> > > I tried the patch with my Nokia 6620: > syncml-obex-client -b 00:02:EE:6D:62:FC 11 --sync > vcard /datasync/Contacts --wbxml --identifier "PC Suite" > connection with device succeeded > > The problem is this locks up my phone -- it says sync incomplete on the > phone and is stuck there. > > Question -- Armin, is this the same bug you mentioned before that you > haven't hunted down? might be. i would have to look at the traces to see whats going on. > > Question -- Is there a way to make the output more verbose or trace? use: export SYNCML_TRACE=/path/to/log/dir > > Question -- is the file location in my command line (/datasync/Contacts) > suppose to match what the phone uses? I only 'touched' the file in that > directory and doesn't contain anything -- should it? it is not important that this directory really exists in the file system. It just the name of a database. You can choose whatever you like there but the names must match for the syncml tools and the phone. For syncml over http, you can normally enter the database name on the phone as well. so just make sure that you use the same for the phone and the tool. for syncml over obex, you normally have to use the exact same name as the phone which is normally fixed depending on phone model and manufacturer. I guess it is possible here to add support to libsyncml to accept any database name, but i have to take a closer look at this. Armin > > Thanks! > Norm > |
From: Norm D. <no...@dr...> - 2006-03-09 14:51:51
Attachments:
trace.tar.gz
|
> > I tried the patch with my Nokia 6620: > > syncml-obex-client -b 00:02:EE:6D:62:FC 11 --sync > > vcard /datasync/Contacts --wbxml --identifier "PC Suite" > > connection with device succeeded > > > > The problem is this locks up my phone -- it says sync incomplete on the > > phone and is stuck there. > > > > Question -- Armin, is this the same bug you mentioned before that you > > haven't hunted down? > > might be. i would have to look at the traces to see whats going on. > > > > > Question -- Is there a way to make the output more verbose or trace? > > use: export SYNCML_TRACE=/path/to/log/dir > > > > > Question -- is the file location in my command line (/datasync/Contacts) > > suppose to match what the phone uses? I only 'touched' the file in that > > directory and doesn't contain anything -- should it? > > it is not important that this directory really exists in the file > system. It just the name of a database. You can choose whatever you like > there but the names must match for the syncml tools and the phone. > > For syncml over http, you can normally enter the database name on the > phone as well. so just make sure that you use the same for the phone and > the tool. > > for syncml over obex, you normally have to use the exact same name as > the phone which is normally fixed depending on phone model and > manufacturer. I guess it is possible here to add support to libsyncml to > accept any database name, but i have to take a closer look at this. > I don't think its needed to add support to accept any database name -- I can modify it on the phone and PC Suite works as the name since it does connect fine. Its just the sync component that it has trouble with. Attached is a trace from an attempt. I also changed the filename to match the phone -- and now get a different error and no lock ups so its getting closer. Can you take a look at this error and trace and let me know what you think? syncml-obex-client -b 00:02:EE:6D:62:FC 11 --sync vcard Contacts --wbxml --identifier "PC Suite" --version 1.1 connection with device succeeded I/O error : Attempt to load network entity http://www.syncml.org/docs/syncml_represent_v11_20020213.dtd Received an Alert for the DS Server at Contacts: Type: 200, Last 20060207T012140Z, Next 20060309T054856Z Just received a new session with ID 22 Session 22 reported final. flushing I/O error : Attempt to load network entity http://www.syncml.org/docs/syncml_represent_v11_20020213.dtd Received an transport error My phone then disconnects and says sync incomplete. Thanks! Norm |