Re: [SyncEvolution] Debugging a connection to Oracle Calendar Server
Actually in a parallel thread Julien writes that Oracle is a SyncML 1.1 server (and if we can track my problem down, we might be able to help him too). Andrzej On Mon, 12 Oct 2009, Patrick Ohly wrote: On Mon, 2009-10-12 at 08:26 +0100, Andrzej Wąsowski wrote: On Mon, 12 Oct 2009, Patrick Ohly wrote: On Sun, 2009-10-11 at 21:30 +0100, Andrzej Wąsowski wrote: When I examine my configs the only thing that is different between what I see in the guides for mobile users and my own setup is that they *always* specify syncML version 1.1. However I can see no way to specify this in in syncevolution files. I can see in the dumped log that SyncML 1.2 is consistently being send. Is there any way to configure this? You could place a syncclient_sample_config.xml configuration in the current directory where you run SyncEvolution and modify it: http://git.moblin.org/cgit.cgi/syncevolution/plain/src/syncclient_sample_config.xml In the client part, add defaultsyncmlversion1.1/defaultsyncmlversion Does libsynthesis automatically reads the file called syncclient_sample_config.xml in current directory? It does not seem to have any effect for me. The message is consistently send out in 1.2 version. The .html log contains a dump of the XML config that was used. It indeed doesn't contain a defaultsyncmlversion1.1/defaultsyncmlversion. Are you sure that the current directory had a modified syncclient_sample_config.xml? The file is loaded by libsyncevolution.so: $ strings -a /usr/lib/syncevolution/libsyncevolution.so | grep sample $ ls *sample*xml syncclient_sample_config.xml $ strace -e trace=open syncevolution scheduleworld_1 vcard30 21 | grep sample open(syncclient_sample_config.xml, O_RDONLY) = 6 open(syncclient_sample_config.xml, O_RDONLY) = 25 I have attached an example - I am not sure if this mailing list accepts attachments. I will repost in some other way if it does not work. It does up to 40KB, but in this case I would have to okay it because the size is exceeded. Let me reply directly instead, quoting your whole email. I have also attached a log of the same run, at the debug level 4. The relevant (and unfortunately not very informative part) is quoted below. But as Congwu said, there should be a fallback to that version if 1.2 fails, so it might be easier to check the logs first. Yes. I suspected so, but I somehow just do not trust this Oracle implementation. Since everybody on the web using mobile fonts sets to 1.1 there must be somethign about it (perhaps mobile phones only implement 1.1?). I would expect that at least some are more modern than that. I look forward to any suggestions from you guys. This thing has been driving me mad for weeks now. Wow, you've been really patient. On Mon, 2009-10-12 at 08:26 +0100, Andrzej Wąsowski wrote: * [2009-10-12 09:20:36.661] Calling smlProcessData(NEXT_COMMAND) * [2009-10-12 09:20:36.661] === smlProcessData failed, returned 0x10 Whatever the server sent back, it couldn't be parsed at all. There isn't even a dump of the incoming data (unless you forgot to attach it). Can you try with --run --sync-property enableWBXML=0? If you don't get the incoming data dumped by SyncEvolution/libsynthesis, then you can at least grab it from a normal network dump in a readable XML format -- Andrzej Wąsowski, http://www.wasowski.org___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
Re: [SyncEvolution] syncevolution and oracle = transport failure
Hello, Le lundi 12 octobre 2009 à 11:28 +0200, Andrzej Wąsowski a écrit : Dear Julien, On Mon, 12 Oct 2009, Julien Garet wrote: I get an error when trying to sync : [INFO] SoupTransport Failure: https://corp.agenda.com/ocst-bin/ocas.fcgi via libsoup: Connection terminated unexpectedly In my setup I was able to advance a bit by switching the requirement for certificates off (I also was not able to get the certificates succeed, with the same error as you do). Do you mean setting SSLVerifyServer = 0 and SSLVerifyHost = 0 ? It does not work better for me. I am also using an Oracle server. Unfortunately I got stuck a bit later. I feel less alone ;-) Here is my config.ini (only what isn't commented in the file), did I miss something ? I did not make any change to the .synthesis files, only changed the config.ini files in sources syncURL = https://agenda.inria.fr:/ocst-bin/ocas.fcgi username = garet password = my_password logdir = /tmp/syncevo loglevel = 5 maxlogdirs = 1 RetryInterval = 5 deviceId = sc-pim-290f49b4-ab78-4c7b-9c82-6bed35815f2c SSLVerifyServer = 0 SSLVerifyHost = 0 WebURL = IconURI = ConsumerReady = 1 and here are my sources files : addressbook/config.ini:sync = none addressbook/config.ini:type = addressbook:text/vcard addressbook/config.ini:uri = ./Contacts calendar/config.ini:sync = two-way calendar/config.ini:type = calendar calendar/config.ini:uri = ./Calendar/Events memo/config.ini:sync = none memo/config.ini:type = memo memo/config.ini:uri = note todo/config.ini:sync = none todo/config.ini:type = todo todo/config.ini:uri = ./Calendar/Tasks Regards, Julien ___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
Re: [SyncEvolution] Debugging a connection to Oracle Calendar Server
On Mon, 12 Oct 2009, Patrick Ohly wrote: Can you try with --run --sync-property enableWBXML=0? If you don't get the incoming data dumped by SyncEvolution/libsynthesis, then you can at least grab it from a normal network dump in a readable XML format Can I actually do this? Now that I remembered that this is running over ssl, I believe? Andrzej ___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
Re: [SyncEvolution] syncevolution and oracle = transport failure
Julien, On Mon, 12 Oct 2009, Julien Garet wrote: In my setup I was able to advance a bit by switching the requirement for certificates off (I also was not able to get the certificates succeed, with the same error as you do). Do you mean setting SSLVerifyServer = 0 and SSLVerifyHost = 0 ? Yes, this is what I did (I tried to add the certificate by exporting it from FireFox on the calendar website, but this did not work. Libsoup was complaining). It does not work better for me. Were you still getting the same error message? I start to wonder if I do not suffer from the ssl problem myself, while I try to find the solution elsewhere. syncURL = https://agenda.inria.fr:/ocst-bin/ocas.fcgi username = garet password = my_password logdir = /tmp/syncevo loglevel = 5 maxlogdirs = 1 RetryInterval = 5 deviceId = sc-pim-290f49b4-ab78-4c7b-9c82-6bed35815f2c SSLVerifyServer = 0 SSLVerifyHost = 0 WebURL = IconURI = ConsumerReady = 1 and here are my sources files : addressbook/config.ini:sync = none addressbook/config.ini:type = addressbook:text/vcard addressbook/config.ini:uri = ./Contacts calendar/config.ini:sync = two-way calendar/config.ini:type = calendar calendar/config.ini:uri = ./Calendar/Events I only use the calendar at the moment - I want to make it work first. But my set up is essentially the same as yours. Other users of my server also add a query limitiing the scope of synchronization. I heard rumours that unlimited does not work, but I do not know if this is a client or a server limitation in their cases. calendar/config.ini:uri = ./Calendar/Events?/dr(-60,180) Gives a scope of 60 days back and 180 in future. Andrzej ___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
Re: [SyncEvolution] syncevolution and oracle = transport failure
Hi Andrzej, On Oct 12, 2009, at 13:51 , Andrzej Wąsowski wrote: I only use the calendar at the moment - I want to make it work first. But my set up is essentially the same as yours. Other users of my server also add a query limitiing the scope of synchronization. I heard rumours that unlimited does not work, but I do not know if this is a client or a server limitation in their cases. AFAIK Oracle servers just use a default limit (often quite tight, such as only 7 days into the future) when none is specified using the /dr(- x,y) syntax. Note also that the Oracle server can be configured to reject any sync attempt with clients it does not know (i.e. for which no so-called profile exists on the server). This might be the problem - altough according to Oracle by default the server does not reject unknown devices. If someone has admin access to the server, it might be worth trying to create a profile for SyncEvolution, probably easiest way would be duplicating the Synthesis iPhone client profile (as it is a close relative to SyncEvolution) and adapt the model matching strings. See http://download.oracle.com/docs/cd/B25553_01/calendar.1012/b25485/ocas.htm#BABCEGID for documentation - I have no experience myself as I never had admin access to an OCS installation. But I have done a lot of IOT with Oracle. Best Regards, Lukas Zeller (l...@synthesis.ch) - Synthesis AG, SyncML Solutions Sustainable Software Concepts i...@synthesis.ch, http://www.synthesis.ch ___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
Re: [SyncEvolution] syncevolution and oracle = transport failure
Dear Lukas, Thank you for a hint. On Mon, 12 Oct 2009, Lukas Zeller wrote: Note also that the Oracle server can be configured to reject any sync attempt with clients it does not know (i.e. for which no so-called profile exists on the server). This might be the problem - altough according to Oracle by default the server does not reject unknown devices. If someone has admin access to the server, it might be worth trying to create a profile for SyncEvolution. This may be a real pain, as I am officially not entitled to use our IT people time to support Linux - I am doing this in my free time, and I should also be careful how do I use the time of admin people. Nevertheless, I asked to see whether this could be the reason. If this is the reason, perhaps I could hack syncevolution to pretend it is an iPhone? Andrzej ___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
Re: [SyncEvolution] Debugging a connection to Oracle Calendar Server
On Mon, 2009-10-12 at 13:58 +0100, Andrzej Wąsowski wrote: I can now confirm that (when it appears) I get a message of length zero saved with log (synclib_linux_msg001_003_incoming_ERR_0x0010.xml). I'm pretty confident that this is what is sent by the server. Is there a chance to debug this from the server side? With nothing sent back to us it is kind of hard to diagnose what the problem might be. And there are two outgoing messages stored in the cache too (I attach both). Unfortunately both of them are still in SyncML 1.2 That defaultsyncmlversion doesn't have the intended effect is worth investigating, but I think this whole SyncML 1.1 aspect is a red herring. According to Lukas, Oracle has supported 1.2 for several years now. What was the log level again that you are using? If you use something sufficiently large (try 10), then you should get message dumps of the raw message. This is in addition to the trm*.xml files that are generated while libsynthesis parses the message. Hmm. I cannot see them in the log (I tried with 30) but perhaps I do not know where to search. It is attached. The _ERR_ file is what I was looking for. Seems it is only written when raw message dumping is enabled (= higher log level). It might be worthwhile to dump it also at lower levels. -- 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. ___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
Re: [SyncEvolution] syncevolution and oracle = transport failure
Hello Andrzej, On Oct 12, 2009, at 15:18 , Andrzej Wąsowski wrote: If this is the reason, perhaps I could hack syncevolution to pretend it is an iPhone? Yes, at least to see if it makes any difference. You could change the model and manufacturer settings in the XML config. The iPhone client identifies itself as: Model: SySync Client iPhone Contacts+TodoZ Manufacturer : Synthesis AG You might want to try the much longer available PocketPC client first, as the profile for the iPhone client was added to OCS no so long ago, so it might not yet be known on a older installation: Model: SySync Client PocketPC PRO Manufacturer : Synthesis AG Please note however that using these identifiers is not a solution, only a temporary hack and testing setup. Best Regards, Lukas Zeller (l...@synthesis.ch) - Synthesis AG, SyncML Solutions Sustainable Software Concepts i...@synthesis.ch, http://www.synthesis.ch ___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
Re: [SyncEvolution] syncevolution and oracle = transport failure
On Mon, 2009-10-12 at 14:18 +0100, Andrzej Wąsowski wrote: If this is the reason, perhaps I could hack syncevolution to pretend it is an iPhone? Oracle seems to match against Model and Manufacturer, plus optionally the firmware version. You can override this in your client XML config by changing model/ manufacturer/ into manufacturerSonyEricsson/manufacturer modelK750i/model I picked the K750i as example here because the strings were in the documentation referenced by Lukas whereas I don't know what the strings for the Synthesis iPhone client are. -- 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. ___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
Re: [SyncEvolution] syncevolution and oracle = transport failure
Le lundi 12 octobre 2009 à 13:51 +0200, Andrzej Wąsowski a écrit : Julien, On Mon, 12 Oct 2009, Julien Garet wrote: In my setup I was able to advance a bit by switching the requirement for certificates off (I also was not able to get the certificates succeed, with the same error as you do). Do you mean setting SSLVerifyServer = 0 and SSLVerifyHost = 0 ? Yes, this is what I did (I tried to add the certificate by exporting it from FireFox on the calendar website, but this did not work. Libsoup was complaining). It does not work better for me. Were you still getting the same error message? I start to wonder if I do not suffer from the ssl problem myself, while I try to find the solution elsewhere. Yes, I am getting error messages from libsoups : [INFO] SoupTransport Failure: https://agenda.inria.fr:/ocst-bin/ocas.fcgi via libsoup: Connection terminated unexpectedly telnet agenda.inria.fr: works well... There might be something the client says to the server that is unpleasant... I attach the sysynclib* files from the last run of syncevolution. I have from where I run the syncevolution client a syncclient_sample_config.xml containing various tests from the mails of the day. syncURL = https://agenda.inria.fr:/ocst-bin/ocas.fcgi username = garet password = my_password logdir = /tmp/syncevo loglevel = 5 maxlogdirs = 1 RetryInterval = 5 deviceId = sc-pim-290f49b4-ab78-4c7b-9c82-6bed35815f2c SSLVerifyServer = 0 SSLVerifyHost = 0 WebURL = IconURI = ConsumerReady = 1 and here are my sources files : addressbook/config.ini:sync = none addressbook/config.ini:type = addressbook:text/vcard addressbook/config.ini:uri = ./Contacts calendar/config.ini:sync = two-way calendar/config.ini:type = calendar calendar/config.ini:uri = ./Calendar/Events I only use the calendar at the moment - I want to make it work first. But my set up is essentially the same as yours. Other users of my server also add a query limitiing the scope of synchronization. I heard rumours that unlimited does not work, but I do not know if this is a client or a server limitation in their cases. I think our server is set to do something like +60 days in the future, does not know if it's default. calendar/config.ini:uri = ./Calendar/Events?/dr(-60,180) Gives a scope of 60 days back and 180 in future. Andrzej sysynclib_linux.tgz Description: application/compressed-tar ___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
Re: [SyncEvolution] syncevolution and oracle = transport failure
On Mon, 2009-10-12 at 16:46 +0100, Julien Garet wrote: Were you still getting the same error message? I start to wonder if I do not suffer from the ssl problem myself, while I try to find the solution elsewhere. Yes, I am getting error messages from libsoups : [INFO] SoupTransport Failure: https://agenda.inria.fr:/ocst-bin/ocas.fcgi via libsoup: Connection terminated unexpectedly telnet agenda.inria.fr: works well... There might be something the client says to the server that is unpleasant... Your case is different than Andrzej's The HTTP(S) transport in libsoup doesn't get back anything from the server. This is something which has to be debugged at the network level with network dumps, to determine whether the SSL handshake works at all. -- 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. ___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
Re: [SyncEvolution] gdbus-cxx and methods in base classes
On Thu, 2009-10-08 at 17:33 +0100, Patrick Ohly wrote: I couldn't answer this without looking into the problem more thoroughly. The result is in the gdbus-cxx branch. This is still experimental. I need to split it up into smaller commits, want to set the ASYNC flag automatically (should be doable now), const members don't work yet, have to adapt SyncEvolution, ... Okay, done. See the updated gdbus-cxx branch. Yongsheng, can you review this so? It's ready to replace the current code in the dbus-api branch. I didn't get the const methods case to work. The last commit contains an example (hello_const) which is commented out at the moment, because enabling it demonstrates that registering such a const method fails. C++ experts to the rescue... On the other hand, this is not an essential feature as long as one controls the implementation of the methods. An extra const keyword can always be removed, even though it is not nice. -- 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. ___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
[SyncEvolution] PHOTO on server not shown on SyncEvolution client
Hello, Björn We (SyncEvolution) has found a problem integrating with mobical.net[1]: Upload a photo to mobical contact and then sync it to client, the photo can't be shown on client. My first investigation is something like a bug at the server side: From RFC2425: Unfolding is accomplished by regarding CRLF immediately followed by a white space character (namely HTAB ASCII decimal 9 or SPACE ASCII decimal 32) as equivalent to no characters at all (i.e., the CRLF and single white space character are removed). The encoded base 64 value can be split across multiple physical lines in the content type by using the line folding technique described above. While mobical is sending us: PHOTO;TYPE=JPEG;ENCODING=BASE64:/9j/4AAQSkZJRgABAQEAYABgAAD/4QAWRXhpZgAASUkqAAgAAAD/2wBDAAgGBgcG\r\n more data next line... We need a space immediately after CRLF to consider it as line unfolding, otherwise We are thinking this comes end of property value. Could you help verify it? [1] http://bugzilla.moblin.org/show_bug.cgi?id=6668 Thanks, Chen,Congwu ___ SyncEvolution mailing list SyncEvolution@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution
[Syncevolution-issues] [Bug 6679] All-day events + time zone
http://bugzilla.moblin.org/show_bug.cgi?id=6679 --- Comment #2 from shuangeeer yanshuang.zh...@intel.com 2009-10-12 04:45:34 --- (In reply to comment #1) An all-day event is by definition not bound to a specific time zone. It is something that is bound to the day (like an anniversary), not a specific range of hours during that day. Yeah, from common usage model, we rarely strike a meeting lasting a whole day. All-day events usually are anniversaries. Thanks for clarification. Expected: time zone should also incorporated. Which time zone? If the creator of the event wants to specify an event that is local to his day, he should create an event ranging from 00:00 to 23:59 and specify the time zone. I was thinking that the time of events gain time zone info which is configured when created or defaultly loaded from user account configuration. Shall we close it as won't fix ? -- Configure bugmail: http://bugzilla.moblin.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are watching the assignee of the bug. You are watching someone on the CC list of the bug. ___ Syncevolution-issues mailing list Syncevolution-issues@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution-issues
[Syncevolution-issues] [Bug 6679] All-day events + time zone
http://bugzilla.moblin.org/show_bug.cgi?id=6679 pohly patrick.o...@intel.com changed: What|Removed |Added Status|NEW |RESOLVED Resolution||WONTFIX --- Comment #3 from pohly patrick.o...@intel.com 2009-10-12 07:10:50 --- (In reply to comment #2) Shall we close it as won't fix ? I certainly won't object ;-) -- Configure bugmail: http://bugzilla.moblin.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are watching the assignee of the bug. You are watching someone on the CC list of the bug. ___ Syncevolution-issues mailing list Syncevolution-issues@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution-issues
[Syncevolution-issues] [Bug 6679] All-day events + time zone
http://bugzilla.moblin.org/show_bug.cgi?id=6679 shuangeeer yanshuang.zh...@intel.com changed: What|Removed |Added Status|RESOLVED|CLOSED --- Comment #4 from shuangeeer yanshuang.zh...@intel.com 2009-10-12 18:04:44 --- closing -- Configure bugmail: http://bugzilla.moblin.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are watching the assignee of the bug. You are watching someone on the CC list of the bug. ___ Syncevolution-issues mailing list Syncevolution-issues@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution-issues
[Syncevolution-issues] [Bug 6664] Mobical: priorities of tasks not matched with other applications
http://bugzilla.moblin.org/show_bug.cgi?id=6664 --- Comment #3 from yongsheng zhu yongsheng@intel.com 2009-10-12 22:02:02 --- We should use the iCalendar 2.0 range internally in the Synthesis field list and map to the smaller 1-3 range when talking to peers in vCalendar 1.0. so I'll map iCalendar 2.0 range to vCalendar1.0, not only for Mobical. -- Configure bugmail: http://bugzilla.moblin.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are watching someone on the CC list of the bug. ___ Syncevolution-issues mailing list Syncevolution-issues@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution-issues
[Syncevolution-issues] [Bug 6668] Mobical: PHOTO on server not shown on client
http://bugzilla.moblin.org/show_bug.cgi?id=6668 --- Comment #2 from Chen Congwu congwu.c...@intel.com 2009-10-12 22:38:48 --- (In reply to comment #1) Correct, I consider this is a bug of mobical, I will ask Bjoern for help. From RFC2425: Unfolding is accomplished by regarding CRLF immediately followed by a white space character (namely HTAB ASCII decimal 9 or SPACE ASCII decimal 32) as equivalent to no characters at all (i.e., the CRLF and single white space character are removed). The encoded base 64 value can be split across multiple physical lines in the content type by using the line folding technique described above. While mobical is sending us: PHOTO;TYPE=JPEG;ENCODING=BASE64:/9j/4AAQSkZJRgABAQEAYABgAAD/4QAWRXhpZgAASUkqAAgAAAD/2wBDAAgGBgcG\r\n more data next line... We need a space immediately after CRLF to consider it as line unfolding, otherwise We are thinking this comes end of property value. To me this looks like a line continuation problem. The Mobical encoder might have an issue here, but it could also be the Synthesis parser. -- Configure bugmail: http://bugzilla.moblin.org/userprefs.cgi?tab=email --- You are receiving this mail because: --- You are watching someone on the CC list of the bug. ___ Syncevolution-issues mailing list Syncevolution-issues@syncevolution.org http://lists.syncevolution.org/listinfo/syncevolution-issues
