6.0.0-alpha14
6/24/25

[#9438] authentication fails via syncml
Summary authentication fails via syncml
Queue Synchronization
Queue Version Git master
Type Bug
State Resolved
Priority 2. Medium
Owners jan (at) horde (dot) org
Requester d-fens (at) 00010111 (dot) de
Created 12/09/2010 (5311 days ago)
Due
Updated 11/13/2012 (4606 days ago)
Assigned 03/06/2011 (5224 days ago)
Resolved 03/07/2011 (5223 days ago)
Github Issue Link
Github Pull Request
Milestone 4
Patch Yes

History
11/13/2012 01:24:56 AM luichanfe (at) gmail (dot) com Comment #35 Reply to this comment
I'm experiencing the same problem but with the SyncMl Backend 
standalone implementation. What are the needed changes for this file 
/lib/Horde/SyncMl/Backend.php ?
03/07/2011 09:16:40 PM Jan Schneider State ⇒ Resolved
 
03/07/2011 08:12:03 PM logan (dot) owen (at) gmail (dot) com Comment #34 Reply to this comment
Seems to be working without any problems.  Thanks for committing these 
changes!

-- Logan

03/07/2011 04:38:28 PM Jan Schneider Comment #33 Reply to this comment
Understood. Try now?
03/07/2011 04:38:16 PM Git Commit Comment #32 Reply to this comment
Changes have been made in Git for this ticket:

Bug #9438: restore the original session id after authentication.

  1 files changed, 20 insertions(+), 4 deletions(-)
http://git.horde.org/horde-git/-/commit/dc3188a1e60519652b431f2f2352f86630c8b885
03/07/2011 04:09:52 PM logan (dot) owen (at) gmail (dot) com Comment #31 Reply to this comment
Jan,

Sorry, I did not do a very good job of explaining.

Basically, the problem occurs when the client makes multiple requests 
to the server.

The first request starts the session with the md5(sessionId . 
deviceId) [KEY A].  Then it authenticates successfully, which causes 
the server to regenerate the session key [KEY B].  Then,  at close(), 
the server stores that the client is auth'd under the new session key 
[KEY B].

The client makes a second request, which does not include auth data, 
because it auth'd successfully on the first request.  However, it is 
not aware of the new session id,  as there is no facility in SyncML 
for the server to regenerate the session key.  So, the server starts a 
session with md5(sessionId . deviceId) [KEY A] and looks to see if the 
client is auth'd, but it isn't, because the auth data was stored under 
a different session key [KEY B].  So this time the client request 
fails because the server does not recognize it as being authenticated.

Sorry if I am still not being clear.

-- Logan
03/07/2011 03:59:36 PM Jan Schneider Comment #30 Reply to this comment
No, AFAICS whether the user is authenticated or not is stored in the 
state object, and only if he is *not* authenticated, we're trying to 
authenticate at the backend. If this happens during a sync session, 
then the state object is *not* reflecting the authentication state.
03/07/2011 03:56:54 PM Git Commit Comment #29 Reply to this comment
Changes have been made in Git for this ticket:

Bug #9438: add missing php tag.

  1 files changed, 1 insertions(+), 0 deletions(-)
http://git.horde.org/horde-git/-/commit/2c45ce168d2bc71788a6f80bd3aacbc2d4f74d28
03/07/2011 03:35:21 PM logan (dot) owen (at) gmail (dot) com Comment #28 Reply to this comment
Jan,

We still need the change to _checkAuthentication().  The reason for it 
is in the patch I sent, but basically the Horde authentication backend 
creates a new session id upon successful authentication.  This is 
standard good practice to avoid "session fixation", but unfortunately 
SyncML relies on session fixation, because the client specifies the 
session id, and the server has no way to request a new id.

So, you have to store the session id before authentication and reset 
it after, unless you want to update the authentication backend to add 
a parameter to suppress session regeneration.

-- Logan
03/07/2011 03:31:44 PM logan (dot) owen (at) gmail (dot) com Comment #27 Reply to this comment
Jan,

I did, but the error was still there.  The problem is because 
LifoQueue.php doesn't have "<?php" at the top of the file, so the code 
isn't being interpreted as PHP.

-- Logan
03/07/2011 03:15:16 PM Jan Schneider Comment #26 Reply to this comment
PHP Fatal error:  Class 'Horde_Xml_Wbxml_LifoQueue' not found in
You didn't run install_dev.
03/07/2011 03:01:51 PM logan (dot) owen (at) gmail (dot) com Comment #25 Reply to this comment
Jan,

That error has been fixed, but now in my Apache error log I get the 
following error:

PHP Fatal error:  Class 'Horde_Xml_Wbxml_LifoQueue' not found in 
/usr/share/horde/framework/Xml_Wbxml/lib/Horde/Xml/Wbxml/ContentHandler.php on 
line 34

Thanks,
Logan
03/07/2011 11:30:05 AM Git Commit Comment #24 Reply to this comment
Changes have been made in Git for this ticket:

Bug #9438: Those classes are autoloadable now.

  1 files changed, 0 insertions(+), 5 deletions(-)
http://git.horde.org/horde-git/-/commit/fa107ceddec4f25b73e6441534d5f6d02afb4743
03/07/2011 03:49:33 AM logan (dot) owen (at) gmail (dot) com Comment #23 Reply to this comment
A recent commit seems to have introduced a bug.  I'm now getting the 
following message in my debug log:

DEBUG: HORDE [horde] PHP ERROR: 
require_once(XML/WBXML/ContentHandler.php) [<a 
href='function.require-once'>function.require-once</a>]: failed to 
open stream: No such file or directory [pid 1666 on line 119 of 
"/usr/share/horde/framework/SyncMl/lib/Horde/SyncMl/ContentHandler.php"]

Is that path supposed to be 'Horde/Xml/Wbxml/ContentHandler.php'?

-- Logan
03/06/2011 09:47:53 PM Jan Schneider Comment #22
Taken from Gunnar Wrobel
State ⇒ Feedback
Reply to this comment
Did this fix things? I don't see why the patch to 
_checkAuthentication() would be necessary.
03/06/2011 09:46:52 PM Git Commit Comment #21 Reply to this comment
Changes have been made in Git for this ticket:

Bug 9438: Use the client's session id and store the state on message 
end, not on session end.

  2 files changed, 9 insertions(+), 4 deletions(-)
http://git.horde.org/horde-git/-/commit/ad914b9601cb42e1702fd71e1450666cfcac6dfb
03/02/2011 01:51:13 PM d-fens (at) 00010111 (dot) de Comment #20 Reply to this comment
ok found another issue with funambol client:

when resetting contacts from client to server the session is started 
and continued ok but closed one packet too early. All contacts were 
replaces(added) and a final is sent to the client, which in turn 
generates another packet but as the session got closed this last 
packet fails, leading to a "sync error" on the client.
when commenting out the

$GLOBALS['backend']->sessionClose();
in horde/framework/SyncML/SyncML.php on line 529 the sync completes 
fine, so somehow the handleEnd() Method still has a bug, i add the 
last server/client packet that is kind of unexpected by the server

last server response:

<?xml version="1.0" encoding="UTF-8"?><SyncML 
xmlns="syncml:syncml1.2"><SyncHdr><VerDTD>1.2</VerDTD><VerProto>SyncML/1.2</VerProto><SessionID>1299072838078</SessionID><MsgID>4</MsgID><Target><LocURI>fac-357841037330855</LocURI><LocName>simon@aktionspotenzial.de</LocName></Target><Source><LocURI>http://192.168.1.99/horde/rpc.php</LocURI></Source><RespURI>http://192.168.1.99/horde/rpc.php</RespURI><Meta><MaxMsgSize xmlns="syncml:metinf">1000000000</MaxMsgSize><MaxObjSize 
xmlns="syncml:metinf">1000000000</MaxObjSize></Meta></SyncHdr><SyncBody><Status><CmdID>1</CmdID><MsgRef>4</MsgRef><CmdRef>0</CmdRef><Cmd>SyncHdr</Cmd><TargetRef>http://192.168.1.99/horde/rpc.php</TargetRef><SourceRef>fac-357841037330855</SourceRef><Data>200</Data></Status><Status><CmdID>2</CmdID><MsgRef>4</MsgRef><CmdRef>2</CmdRef><Cmd>Sync</Cmd><TargetRef>contacts</TargetRef><SourceRef>contacts</SourceRef><Data>200</Data></Status><Status><CmdID>3</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>741</SourceRef><Data>201</Data></Status><Status><CmdID>4</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>742</SourceRef><Data>201</Data></Status><Status><CmdID>5</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>743</SourceRef><Data>201</Data></Status><Status><CmdID>6</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>744</SourceRef><Data>201</Data></Status><Status><CmdID>7</CmdID><MsgRef>4</MsgRef><CmdR
ef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>745</SourceRef><Data>201</Data></Status><Status><CmdID>8</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>746</SourceRef><Data>201</Data></Status><Status><CmdID>9</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>747</SourceRef><Data>201</Data></Status><Status><CmdID>10</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>748</SourceRef><Data>201</Data></Status><Status><CmdID>11</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>749</SourceRef><Data>201</Data></Status><Status><CmdID>12</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>750</SourceRef><Data>201</Data></Status><Status><CmdID>13</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>751</SourceRef><Data>201</Data></Status><Status><CmdID>14</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>752</SourceRef><Data>201</Data></Status><Status>
<CmdID>15</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>753</SourceRef><Data>201</Data></Status><Status><CmdID>16</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>754</SourceRef><Data>201</Data></Status><Status><CmdID>17</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>755</SourceRef><Data>201</Data></Status><Status><CmdID>18</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>756</SourceRef><Data>201</Data></Status><Status><CmdID>19</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>757</SourceRef><Data>201</Data></Status><Status><CmdID>20</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>758</SourceRef><Data>201</Data></Status><Status><CmdID>21</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>759</SourceRef><Data>201</Data></Status><Status><CmdID>22</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>760</S
ourceRef><Data>201</Data></Status><Status><CmdID>23</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>761</SourceRef><Data>201</Data></Status><Status><CmdID>24</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>762</SourceRef><Data>201</Data></Status><Status><CmdID>25</CmdID><MsgRef>4</MsgRef><CmdRef>3</CmdRef><Cmd>Replace</Cmd><SourceRef>763</SourceRef><Data>201</Data></Status><Final></Final></SyncBody></SyncML>



then closing session in handleEnd()

and the unexpected response from client:

<SyncML>
<SyncHdr>
<VerDTD><![CDATA[1.2]]></VerDTD>
<VerProto><![CDATA[SyncML/1.2]]></VerProto>
<SessionID><![CDATA[1299072838078]]></SessionID>
<MsgID>5</MsgID>
<Target>
<LocURI><![CDATA[http://192.168.1.99/horde/rpc.php]]></LocURI>
</Target>
<Source>
<LocURI><![CDATA[fac-357841037330855]]></LocURI>
</Source>
</SyncHdr>
<SyncBody>
<Status>
<CmdID><![CDATA[1]]></CmdID>
<MsgRef><![CDATA[4]]></MsgRef>
<CmdRef><![CDATA[0]]></CmdRef>
<Cmd><![CDATA[SyncHdr]]></Cmd>
<TargetRef><![CDATA[fac-357841037330855]]></TargetRef>
<SourceRef><![CDATA[http://192.168.1.99/horde/rpc.php]]></SourceRef>
<Data><![CDATA[200]]></Data>
</Status>
<Final></Final>
</SyncBody>
</SyncML>






03/01/2011 12:27:24 PM Jan Schneider Assigned to Jan Schneider
Summary ⇒ authentication fails via syncml
Milestone ⇒ 4
Patch ⇒ Yes
 
03/01/2011 12:21:38 PM d-fens (at) 00010111 (dot) de Comment #19 Reply to this comment
kudos!
works great, thanks a lot!
02/28/2011 03:54:28 PM logan (dot) owen (at) gmail (dot) com Comment #18
New Attachment: 0001-Update-SyncML-to-enable-syncing-with-at-least-the-An.patch Download
Reply to this comment
I believe I have fixed this particular problem.  I am attaching my 
patch against the latest git HEAD.

-- Logan
02/18/2011 03:47:39 PM Michael Rubinsky Comment #17
Summary ⇒ [Regression] authentication fails via syncmlSe
Reply to this comment
See also Bug: 9492
02/17/2011 04:00:33 PM hordetest (at) trash-mail (dot) com Comment #16 Reply to this comment
fair enough, point taken :)
Horde Sync is a great feature :) Is there anything I can do to support 
this issue?
02/01/2011 12:58:19 PM d-fens (at) 00010111 (dot) de Comment #15 Reply to this comment
fair enough, point taken :)
02/01/2011 09:08:45 AM Jan Schneider Comment #14 Reply to this comment
If it is so important for you, at this point, in a development 
version, you can hire our consulting services to get this bug fixed.
01/31/2011 11:12:33 PM d-fens (at) 00010111 (dot) de Comment #13 Reply to this comment
does this bug still have a priority? in the times of a increasing 
demand for smartphone sync solutions horde should sport a reliable 
syncml support, please ...!
01/27/2011 12:44:47 PM hordetest (at) trash-mail (dot) com Comment #12 Reply to this comment
i tried again and still the first login attempt is successfull but 
the second packet in the conversation fails with "authentication 
failed", can i do anything about this?
I can confirm this issue.
01/20/2011 12:06:14 PM d-fens (at) 00010111 (dot) de Comment #11 Reply to this comment
i tried again and still the first login attempt is successfull but the 
second packet in the conversation fails with "authentication failed", 
can i do anything about this?
12/10/2010 02:14:21 PM Michael Rubinsky Comment #10 Reply to this comment
FWIW, this commit fixed the ActiveSync issue...
12/09/2010 06:49:01 PM d-fens (at) 00010111 (dot) de Comment #9 Reply to this comment
99,99% sure.

interesting is: it now seems to work one time but the second request 
gets denied (during one sync)

2010-12-09T19:44:03+01:00 DEBUG: HORDE [imp] Your browser is too old 
to display the dynamic mode. Using traditional mode instead. [pid 
17426 on line 27 of 
"/var/www/_horde/horde/framework/Core/lib/Horde/Core/Notification/Hordelog.php"]
2010-12-09T19:44:03+01:00 NOTICE: HORDE [imp] Login success for xxx 
[88.77.140.237] to {localhost:993 [imap]} [pid 17426 on line 199 of 
"/var/www/_horde/horde/imp/lib/Auth.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Authenticated: yes; 
version: 1.2; message ID: 1; source URI: aku-1291724651; target URI: 
http://xxxx/horde-git/rpc.php; user: xxx; charset: UTF-8; wbxml: no 
[pid 17426 on line 295 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] SQL Query by 
SyncML_Backend_Horde::readSyncAnchors(): SELECT syncml_clientanchor, 
syncml_serveranchor FROM horde_syncml_anchors WHERE syncml_syncpartner 
= ? AND syncml_db = ? AND syncml_uid = ?, values: aku-1291724651, 
calendar, xxx [pid 17426 on line 166 of 
"/var/www/_horde/horde/framework/SyncML/SyncML/Command/Alert.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] No info about previous 
syncs found for device aku-1291724651 and database event [pid 17426 on 
line 446 of "/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Creating SyncML_Sync 
object for database event; sync type 201 [pid 17426 on line 446 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Received <Final> from 
client. [pid 17426 on line 446 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Handle <Final> for 
state Init [pid 17426 on line 408 of 
"/var/www/_horde/horde/framework/SyncML/SyncML/State.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Sending <Final> to 
client [pid 17426 on line 310 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Return message 
completed [pid 17426 on line 310 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Finished at 2010-12-09 
19:44:03. Packet logged in /tmp/sync/server_50.xml [pid 17426 on line 
176 of "/var/www/_horde/horde/framework/SyncML/SyncML.php"]

and then:


010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Horde_Rpc::__construct 
complete [pid 17426 on line 97 of 
"/var/www/_horde/horde/framework/Rpc/lib/Horde/Rpc.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Backend of class 
SyncML_Backend_Horde created [pid 17426 on line 31 of 
"/var/www/_horde/horde/framework/SyncML/SyncML/Backend/Horde.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Started at 2010-12-09 
19:44:03. Packet logged in /tmp/sync/client_60.xml [pid 17426 on line 
108 of "/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] New session created:   
[pid 17426 on line 376 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Invalid authentication 
[pid 17426 on line 295 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Authenticated: no; 
version: 1.2; message ID: 2; source URI: aku-1291724651; target URI: 
http://hosting.aktionspotenzial.de/horde-git/rpc.php; user: ; charset: 
UTF-8; wbxml: no [pid 17426 on line 295 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Received <Final> from 
client. [pid 17426 on line 446 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Sending <Final> to 
client [pid 17426 on line 310 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Session completed and 
closed [pid 17426 on line 310 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] PHP ERROR: 
session_destroy() [<a 
href='function.session-destroy'>function.session-destroy</a>]: Trying 
to destroy uninitialized session [pid 17426 on line 432 of 
"/var/www/_horde/horde/framework/SyncML/SyncML/Backend.php"]
2010-12-09T19:44:03+01:00 DEBUG: HORDE [horde] Finished at 2010-12-09 
19:44:03. Packet logged in /tmp/sync/server_60.xml [pid 17426 on line 
176 of "/var/www/_horde/horde/framework/SyncML/SyncML.php"]

so somehow session handling is borked?
12/09/2010 06:29:03 PM Gunnar Wrobel Comment #8 Reply to this comment
Are you really using SyncML from git? If I look at my git checkout the 
lines associated to the log messages are different and some log 
messages even don't seem to exists anymore.
12/09/2010 05:33:52 PM d-fens (at) 00010111 (dot) de Comment #7 Reply to this comment
hi,

thanks for your fix, but it still doesn't work yet:

2010-12-09T18:22:24+01:00 DEBUG: HORDE [horde] Started at 2010-12-09 
18:22:24. Packet logged in /tmp/sync/client_60.xml [pid 17426 on line 
108 of "/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T18:22:24+01:00 DEBUG: HORDE [horde] New session created:   
[pid 17426 on line 376 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T18:22:24+01:00 DEBUG: HORDE [horde] Invalid authentication 
[pid 17426 on line 295 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T18:22:24+01:00 DEBUG: HORDE [horde] Authenticated: no; 
version: 1.2; message ID: 2; source URI: fac-357841037330855; target 
URI: http://xxxxxxxx/horde-git/rpc.php; user: xxxxxxxxx; charset: 
UTF-8; wbxml: no [pid 17426 on line 295 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T18:22:24+01:00 DEBUG: HORDE [horde] Received <Final> from 
client. [pid 17426 on line 446 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T18:22:24+01:00 DEBUG: HORDE [horde] Sending <Final> to 
client [pid 17426 on line 310 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T18:22:24+01:00 DEBUG: HORDE [horde] Session completed and 
closed [pid 17426 on line 310 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]
2010-12-09T18:22:24+01:00 DEBUG: HORDE [horde] PHP ERROR: 
session_destroy() [<a 
href='function.session-destroy'>function.session-destroy</a>]: Trying 
to destroy uninitialized session [pid 17426 on line 432 of 
"/var/www/_horde/horde/framework/SyncML/SyncML/Backend.php"]
12/09/2010 05:04:43 PM Gunnar Wrobel Comment #6 Reply to this comment
Sorry, was entirely my fault. I incorrectly associated this problem to 
the problems I had with transparent auth. But the problem here was 
solely related to my modification of Horde_Secret. I overlooked that 
the functions previously accepted "null" and "false" as input without 
bailing out.

Can you verify that this commit fixes the problem:

https://github.com/horde/horde/commit/a6267c6747973b74c6b58c2f2a75ad63c03389b6

12/09/2010 02:37:27 PM d-fens (at) 00010111 (dot) de Comment #5 Reply to this comment
tried to comment out in rpc.php but just got another error

Invalid authentication [pid 1380 on line 295 of 
"/var/www/_horde/horde/framework/SyncML/SyncML.php"]

so no real help for me ... hmm

12/09/2010 01:08:44 PM Michael Rubinsky Comment #4
State ⇒ Assigned
Assigned to Gunnar Wrobel
Reply to this comment
Deleting your old cookie before logging in should allow to work 
around the problem for now.

There is data in the old session that fails on relogin under certain 
circumstances. I still need some feedback from Micheal Slusarz to 
understand how this is meant to work.
This is the exact same issue I am seeing with ActiveSync access. At 
least in the case with ActiveSync, it's due to the fact that there 
(correctly) is no session (session_control => 'none') created for 
activesync requests. My guess is syncml also does not need or create a 
session.

As a local workaround, I temporarily removed the 'session_control' => 
none parameter from the appInit call that gets called for activesync 
access, and this works. You could try doing the same for syncml in 
rpc.php temporarily until this gets fixed properly.

12/09/2010 10:53:25 AM d-fens (at) 00010111 (dot) de Comment #3 Reply to this comment
hi,

deleted cookies on my android browser, app data for funambol sync 
client, session files on server -  its still not working
12/09/2010 10:41:13 AM Gunnar Wrobel Comment #2 Reply to this comment
Deleting your old cookie before logging in should allow to work around 
the problem for now.

There is data in the old session that fails on relogin under certain 
circumstances. I still need some feedback from Micheal Slusarz to 
understand how this is meant to work.
12/09/2010 10:22:38 AM d-fens (at) 00010111 (dot) de Comment #1
Priority ⇒ 2. Medium
State ⇒ Unconfirmed
Patch ⇒ No
Milestone ⇒
Queue ⇒ Synchronization
Type ⇒ Bug
Summary ⇒ [Regression] authentication fails via syncml
Reply to this comment
hi,

i can't authenicate to horde anymore via syncml:

EMERG: HORDE [horde] Chiper text must be a string [pid 1380 on line 96 
of "/var/www/_horde/horde/framework/Secret/lib/Horde/Secret.php"]

some recent changes brought this

Saved Queries