This is committed and will be in Turba 2.1. Thanks!

updated patch

Only reason I was checking just the last token was the _isString 
function did that as well.. I agree though that It should check the 
whole UID. (both functions should).

And as for the _getSchema, I do check if the _schema is a a 
Net_LDAP_Schema object and re-fetch, if it's not.  It does, however, 
make more sense to be a static member.

I'll update the patch with these two suggestions.

Looks good. Only two minor suggestions/questions: Is there a reason 
that you don't check the schema against the full UID for the LDAP 
postal address syntax, but only the last token? And you may want to 
store the schema in _getSchema in a static variable to cache it and 
don't create a new schema object for each request.

Here is an updated patch that adds (optional) checks on the schema if 
schema checking is enabled.. otherwise it will fail back to relying on 
the definitions in attributes.php

This should not be closed as it is a real bug that does exist and 
should be fixed.   I just haven't had the time to work on it recently, 
and as I'm not using the horde-CVS it hasn't been that high of a 
priority as I already had my own patch in the stable tree I'm using.   
I have attached an updated patch off of CVS-HEAD.

This does not use the schema checking that was added, As that is an 
"optional" feature  (checkRequired) which has to be called first to 
initialize the _schema object..   The solution attached back 
references the $GLOBALS['attributes'] configuration array after 
querying $this->map to map attribute.

Looks like noone is really interested in this feature or at least in 
putting any work into it.

This is rather an enhancement request than a bug.

This patch has recently been applied to the LDAP driver that checks 
for schemes and field syntaxes: 
http://cvs.horde.org/diff.php/turba/lib/Driver/ldap.php?r1=1.49&r2=1.50&ty=u

Maybe part of this code could be used/adapted to determine the fields 
where we want line breaks encoded as "$"?

Suggested priority: 3. High

don't be silly

Here is an updated patch off of the CVS turba..
I'm the same person as urkle@d.
A few notes though, right now I am explicitly checking for 
homepostaladdress, postaladdress, otherpostaladdress, and 
registeredaddress with that lovely regular expression. As those are 
the standard core scheme that use this syntax.
It would be better if there were a way to access what type of field is 
being processed from within the driver,, or a sheme query on the LDAP 
to determine what attributes are of the syntax 
1.3.6.1.4.1.1466.115.121.1.41.
And this breaks things with evolution, as I am waiting for them to 
accept my patch to fix the bug there as well. Neeed to update it for 
the 1.5CVS release.

Here is an updated patch off of the CVS turba..
I'm the same person as urkle@d.
A few notes though, right now I am explicitly checking for 
homepostaladdress, postaladdress, otherpostaladdress, and 
registeredaddress with that lovely regular expression. As those are 
the standard core scheme that use this syntax.
It would be better if there were a way to access what type of field is 
being processed from within the driver,, or a sheme query on the LDAP 
to determine what attributes are of the syntax 
1.3.6.1.4.1.1466.115.121.1.41.
And this breaks things with evolution, as I am waiting for them to 
accept my patch to fix the bug there as well. Neeed to update it for 
the 1.5CVS release.

This patch fixes an issue with Turba 1.2 where it would insert a Carriage
Return, Line Feed instead of just Line Feed between the lines of a
multiline entry (ie. home address)  This causes a problem with Evolution
where evolution will not display the entry in the address book if there is
a carriage return in the entry..  But the entries should only have line
feeds..

Patch for Turba 1.2 AND CVS HEAD


------- Additional Comments From urkle@drip.ws  06/17/03 16:45 -------
CVS HEAD Patch
--- ldap.php.orig        Fri Jun 13 10:21:15 2003
+++ ldap.php        Tue Jun 17 14:41:25 2003
@@ -302,10 +302,10 @@
          // Don't add empty attributes.
          $attributes = array_filter($attributes, array($this,
'cleanEmptyAttributes'));

-        // Encode entries.
+        // Encode entries & do CRLF to LF conversion
          foreach ($attributes as $key => $val) {
              if (!is_array($val)) {
-                $attributes[$key] = String::convertCharset($val,
NLS::getCharset(), $this->charset);
+                $attributes[$key] =
String::convertCharset(str_replace("\r\n",\n",$val), NLS::getCharset(),
$this->charset);
              }
          }

@@ -377,7 +377,7 @@
          // Encode entries.
          foreach ($attributes as $key => $val) {
              if (!is_array($val)) {
-                $attributes[$key] = String::convertCharset($val,
NLS::getCharset(), $this->charset);
+                $attributes[$key] =
String::convertCharset(str_replace("\r\n","\n",$val), NLS::getCharset(),
$this->charset);
              }
          }



------- Additional Comments From urkle@drip.ws  06/17/03 16:46 -------
Turba 1.2 patch
--- ldap.php.orig        Tue Feb 25 20:40:48 2003
+++ ldap.php        Sun Jun 15 23:54:34 2003
@@ -301,14 +301,17 @@
          // Don't add empty attributes.
          $attributes = array_filter($attributes, array($this,
'cleanEmptyAttributes'));

-        // Encode entries in UTF-8 if requested.
-        if ($this->encoding == 'utf8') {
-            foreach ($attributes as $key => $val) {
-                if (!is_array($val)) {
-                    $attributes[$key] = utf8_encode($val);
-                }
-            }
-        }
+        // fix CRLF to LF in multiline entries
+        foreach ($attributes as $key => $val) {
+            if (!is_array($val)) {
+                $val = str_replace("\r\n","\n",$val);
+                    // Encode entries in UTF-8 if requested.
+                if ($this->encoding == 'utf8') {
+                        $val = utf8_encode($val);
+                    }
+                $attributes[$key] = $val;
+            }
+        }

          if (!@ldap_add($this->ds, $dn, $attributes)) {
              return new PEAR_Error('Failed to add an object: [' .
ldap_errno($this->ds) . '] "' . ldap_error($this->ds) . '" (attributes: [' .
serialize($attributes) . ']).');
@@ -375,14 +378,17 @@
              }
          }

-        // Encode entries in UTF-8 if requested.
-        if ($this->encoding == 'utf8') {
-            foreach ($attributes as $key => $val) {
-                if (!is_array($val)) {
-                    $attributes[$key] = utf8_encode($val);
-                }
-            }
-        }
+        // fix CRLF to LF in multiline entries
+        foreach ($attributes as $key => $val) {
+            if (!is_array($val)) {
+                $val = str_replace("\r\n","\n",$val);
+                    // Encode entries in UTF-8 if requested.
+                if ($this->encoding == 'utf8') {
+                        $val = utf8_encode($val);
+                    }
+                $attributes[$key] = $val;
+            }
+        }

          unset($attributes[$object_key]);
          $attributes = array_filter($attributes, array($this,
'cleanEmptyAttributes'));


------- Additional Comments From urkle@drip.ws  03/06/04 23:39 -------
A way needs to be created to let the driver know that the field type is of
postal address..

------- Additional Comments From urkle@drip.ws  03/07/04 00:12 -------
patch to completely fix this bug.
change both ocurrances of
  $val = str_replace("\r\n","\n",$val);
in the second patch posted to this.
if (preg_match("/^((home|other)?postal|registered)address$/",%key)) {
     $val = str_replace(array("\r\n","\n"),array("$","$"),$val);
}

and then in the getResults function (around line 243)
  if ($field == 'dn') {
      $resul[$field] = $entry[$field];
+} elseif (preg_match("/^((home|other)?postal|registered)address$/",$field)) {
+    if (!empty($entry[$field])) {
+        if ($this->encoding == 'utf8') {
+            $result[$field] = utf8_decode($entry[$field][0]);
+        }
+    }
+    $result[$field] = str_replace("$","\r\n",$result[$field]);
  } else {
      $result[$field] = '';
      if (!empty($entry[$field])) {

And

This patch fixes an issue with Turba 1.2 where it would insert a Carriage
Return, Line Feed instead of just Line Feed between the lines of a
multiline entry (ie. home address)  This causes a problem with Evolution
where evolution will not display the entry in the address book if there is
a carriage return in the entry..  But the entries should only have line
feeds..

Patch for Turba 1.2 AND CVS HEAD


------- Additional Comments From urkle@drip.ws  06/17/03 16:45 -------
CVS HEAD Patch
--- ldap.php.orig        Fri Jun 13 10:21:15 2003
+++ ldap.php        Tue Jun 17 14:41:25 2003
@@ -302,10 +302,10 @@
          // Don't add empty attributes.
          $attributes = array_filter($attributes, array($this,
'cleanEmptyAttributes'));

-        // Encode entries.
+        // Encode entries & do CRLF to LF conversion
          foreach ($attributes as $key => $val) {
              if (!is_array($val)) {
-                $attributes[$key] = String::convertCharset($val,
NLS::getCharset(), $this->charset);
+                $attributes[$key] =
String::convertCharset(str_replace("\r\n",\n",$val), NLS::getCharset(),
$this->charset);
              }
          }

@@ -377,7 +377,7 @@
          // Encode entries.
          foreach ($attributes as $key => $val) {
              if (!is_array($val)) {
-                $attributes[$key] = String::convertCharset($val,
NLS::getCharset(), $this->charset);
+                $attributes[$key] =
String::convertCharset(str_replace("\r\n","\n",$val), NLS::getCharset(),
$this->charset);
              }
          }