<?xml version="1.0" encoding="UTF-8"?> 
<?xml-stylesheet href="https://dev.horde.org/themes/horde//default/feed-rss.xsl" type="text/xsl"?> 
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"> 
 <channel> 
  <title>ical webdav and realms</title> 
  <pubDate>Tue, 07 Apr 2026 17:34:21 +0000</pubDate> 
  <link>https://bugs.horde.org/ticket/6746</link> 
  <atom:link rel="self" type="application/rss+xml" title="ical webdav and realms" href="https://bugs.horde.org/ticket/6746/rss" /> 
  <description>ical webdav and realms</description> 
 
   
   
  <item> 
   <title>When using imp as authentication application and realms are </title> 
   <description>When using imp as authentication application and realms are set in servers.php then it is not possible to fetch calendars via webdav. See also: &lt;http://marc.info/?l=horde-dev&amp;m=121136921813785&amp;w=2&gt;</description> 
   <pubDate>Fri, 23 May 2008 14:45:27 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t45485</link> 
  </item> 
   
  <item> 
   <title>Not sure if this is an IMP or Kronolith problem, but it defi</title> 
   <description>Not sure if this is an IMP or Kronolith problem, but it definitely doesn&#039;t have anything to do with SyncML.</description> 
   <pubDate>Sun, 25 May 2008 14:25:34 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t45573</link> 
  </item> 
   
  <item> 
   <title>&gt; Not sure if this is an IMP or Kronolith problem, but it de</title> 
   <description>&gt; Not sure if this is an IMP or Kronolith problem, but it definitely 

&gt; doesn&#039;t have anything to do with SyncML.



ok, sorry for that

</description> 
   <pubDate>Mon, 26 May 2008 13:20:56 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t45606</link> 
  </item> 
   
  <item> 
   <title>I tried to track the problem. So far I just see that the use</title> 
   <description>I tried to track the problem. So far I just see that the userId in the session gets overwritten by a the userID without @realm sometime after the session is created by imp/lib/Session.php

Therefore Auth::getAuth returns the plain userID without the realm and no calendars of shares are found for it.</description> 
   <pubDate>Thu, 05 Jun 2008 09:55:35 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t45992</link> 
  </item> 
   
  <item> 
   <title>I think I found the problem. But it is complicated for me to</title> 
   <description>I think I found the problem. But it is complicated for me to explain, I&#039;ll try anyway and hope that someone can follow my potentially confusing words, sorry for that:



When RPC webdav does &quot;check_auth&quot;, first a Horde Auth instance of Auth_application is created and the Auth::authenticate method stores the credentials with the plain $userId (no realm) then the imp api method &quot;authenticate&quot; is called by (Auth_application) &quot;_authenticate&quot;. The imp api method &quot;authenticate&quot; calls createSession from imp/lib/Session.php where the realm gets added to the userId.  $_SESSION[&#039;imp&#039;][&#039;user&#039;] and $_SESSION[&#039;imp&#039;][&#039;uniquser&#039;] are stored.

Now a second Auth instance of type Auth_imp is created and Auth_imp::authenticate is called which then calls the parent::authenticate (Auth::authenticate) again which now stores the credentials with the realmed $userId. The Auth_imp::_authenticate method then authenticates the user at the imap server and the Auth::setAuth sets the realmed userId after that the Auth_imp is done.

Now the first Auth instance goes on with the authenticate method and sets the plain userId via Auth::setAuth which over writes the realmed userId that was set by the Auth_imp instance.



To avoid this behavior I was thinking of changing webdav.php and call ind the authenticate method with the &quot;login = false&quot; parameter: 



$auth-&gt;authenticate($username, array(&#039;password&#039; =&gt; $password), false); 



when using imp as authentication driver (the Auth_imp instance set it to true anyway by itself).

By doing this the Auth::setAuth from the initial Auth instance is not called and therefore it is not overwriting the credentials which where set by the Auth_imp instance.



Any comments on that?



Didi</description> 
   <pubDate>Fri, 06 Jun 2008 13:07:02 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t46051</link> 
  </item> 
   
  <item> 
   <title>&gt; To avoid this behavior I was thinking of changing webdav.p</title> 
   <description>&gt; To avoid this behavior I was thinking of changing webdav.php and call 

&gt; the authenticate method with the &quot;login = false&quot; parameter:

&gt;

&gt; $auth-&gt;authenticate($username, array(&#039;password&#039; =&gt; $password), false);



I tested this for a while now and it seems to work fine.</description> 
   <pubDate>Tue, 10 Jun 2008 13:57:19 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t46190</link> 
  </item> 
   
  <item> 
   <title>I don&#039;t see how that works; without actually logging the use</title> 
   <description>I don&#039;t see how that works; without actually logging the user in, subsequent checks (inside the various api methods) should return empty for Auth::getAuth().</description> 
   <pubDate>Tue, 10 Jun 2008 18:34:01 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t46209</link> 
  </item> 
   
  <item> 
   <title>&gt; I don&#039;t see how that works; without actually logging the u</title> 
   <description>&gt; I don&#039;t see how that works; without actually logging the user in, 

&gt; subsequent checks (inside the various api methods) should return 

&gt; empty for Auth::getAuth().



if you carefully follow the calls that are made (it took me some time :-) you will be faster) when the authdriver is set to application - imp then you will see that the imp api authenticate method will call IMP_Session::createSession and there $auth_imp-&gt;authenticate($_SESSION[&#039;imp&#039;][&#039;uniquser&#039;], array(&#039;password&#039; =&gt; $password), true) will be called where the &quot;login&quot; parameter is set to &quot;true&quot; so the user gets logged in.

But when you set the &quot;login&quot; parameter &quot;true&quot; also in the authenticate call from webdav, this instance of Auth will then overwrite the auth credentials set by $auth_imp.



ok, the fix might work only with the auth driver set to application + imp but it works. 

</description> 
   <pubDate>Tue, 10 Jun 2008 20:09:53 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t46215</link> 
  </item> 
   
  <item> 
   <title>&gt; ok, the fix might work only with the auth driver set to ap</title> 
   <description>&gt; ok, the fix might work only with the auth driver set to application + 

&gt; imp but it works.



... and breaks webdav for everyone else. nice. Not going to be committed.</description> 
   <pubDate>Tue, 10 Jun 2008 20:11:24 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t46216</link> 
  </item> 
   
  <item> 
   <title>&gt;&gt; ok, the fix might work only with the auth driver set to a</title> 
   <description>&gt;&gt; ok, the fix might work only with the auth driver set to application +

&gt;&gt; imp but it works.

&gt;

&gt; ... and breaks webdav for everyone else. nice. Not going to be committed.

 

:-)

What about doing it in a conditional way? I mean just if the driver is set to this combination? </description> 
   <pubDate>Tue, 10 Jun 2008 20:33:24 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t46221</link> 
  </item> 
   
  <item> 
   <title>&gt; What about doing it in a conditional way? I mean just if t</title> 
   <description>&gt; What about doing it in a conditional way? I mean just if the driver 

&gt; is set to this combination?



Then if we fix the _actual_ bug it&#039;ll break again. Still no.</description> 
   <pubDate>Tue, 10 Jun 2008 20:34:39 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t46222</link> 
  </item> 
   
  <item> 
   <title>&gt;&gt; What about doing it in a conditional way? I mean just if </title> 
   <description>&gt;&gt; What about doing it in a conditional way? I mean just if the driver

&gt;&gt; is set to this combination?

&gt;

&gt; Then if we fix the _actual_ bug it&#039;ll break again. Still no.



I understand.

I hope I could at least help tracking the problem. I&#039;m looking forward to the fix of the _actual_ bug and will use the quick hack meanwhile.



</description> 
   <pubDate>Tue, 10 Jun 2008 21:30:58 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t46223</link> 
  </item> 
   
  <item> 
   <title>I see that also syncml is affected by the same problem. Shou</title> 
   <description>I see that also syncml is affected by the same problem. Should I open a separate ticket for it?</description> 
   <pubDate>Mon, 16 Jun 2008 09:54:40 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t46549</link> 
  </item> 
   
  <item> 
   <title>No, it&#039;s exactly the same issue.</title> 
   <description>No, it&#039;s exactly the same issue.</description> 
   <pubDate>Mon, 16 Jun 2008 17:51:33 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t46577</link> 
  </item> 
   
  <item> 
   <title>Ping</title> 
   <description>Ping</description> 
   <pubDate>Thu, 28 Aug 2008 07:53:37 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t48508</link> 
  </item> 
   
  <item> 
   <title>Marking as duplicate of Ticket #6749.</title> 
   <description>Marking as duplicate of Ticket #6749.</description> 
   <pubDate>Thu, 25 Sep 2008 03:19:03 +0000</pubDate> 
   <link>https://bugs.horde.org/ticket/6746#t49221</link> 
  </item> 
   
   
 
 </channel> 
</rss> 
