Improve vcard import: map more fields, support photo urls, better UTF-16 charset detection

program/include/rcube_vcard.php

@@ -40,6 +40,7 @@ class rcube_vcard
     'notes'    => 'NOTE',
     'email'    => 'EMAIL',
     'address'  => 'ADR',
+    'jobtitle' => 'TITLE',
     'gender'      => 'X-GENDER',
     'maidenname'  => 'X-MAIDENNAME',
     'anniversary' => 'X-ANNIVERSARY',
@@ -165,6 +166,10 @@ class rcube_vcard
               }
             }
           }
+
+          // force subtype if none set
+          if (preg_match('/^(email|phone|address|website)/', $key) && !$subtype)
+            $subtype = 'other';
           
           if ($subtype)
             $key .= ':' . $subtype;
@@ -277,8 +282,14 @@ class rcube_vcard
         break;
         
       case 'photo':
-        $encoded = !preg_match('![^a-z0-9/=+-]!i', $value);
-        $this->raw['PHOTO'][0] = array(0 => $encoded ? $value : base64_encode($value), 'BASE64' => true);
+        if (strpos($value, 'http:') === 0) {
+            // TODO: fetch file from URL and save it locally?
+            $this->raw['PHOTO'][0] = array(0 => $value, 'URL' => true);
+        }
+        else {
+            $encoded = !preg_match('![^a-z0-9/=+-]!i', $value);
+            $this->raw['PHOTO'][0] = array(0 => $encoded ? $value : base64_encode($value), 'BASE64' => true);
+        }
         break;
         
       case 'email':
@@ -422,10 +433,16 @@ class rcube_vcard
   {
     // Convert special types (like Skype) to normal type='skype' classes with this simple regex ;)
     $vcard = preg_replace(
-      '/item(\d+)\.(TEL|URL)([^:]*?):(.*?)item\1.X-ABLabel:(?:_\$!<)?([\w-() ]*)(?:>!\$_)?./s',
+      '/item(\d+)\.(TEL|EMAIL|URL)([^:]*?):(.*?)item\1.X-ABLabel:(?:_\$!<)?([\w-() ]*)(?:>!\$_)?./s',
       '\2;type=\5\3:\4',
       $vcard);
 
+    // convert Apple X-ABRELATEDNAMES into X-* fields for better compatibility
+    $vcard = preg_replace_callback(
+      '/item(\d+)\.(X-ABRELATEDNAMES)([^:]*?):(.*?)item\1.X-ABLabel:(?:_\$!<)?([\w-() ]*)(?:>!\$_)?./s',
+      array('self', 'x_abrelatednames_callback'),
+      $vcard);
+
     // Remove cruft like item1.X-AB*, item1.ADR instead of ADR, and empty lines
     $vcard = preg_replace(array('/^item\d*\.X-AB.*$/m', '/^item\d*\./m', "/\n+/"), array('', '', "\n"), $vcard);
     
@@ -440,6 +457,11 @@ class rcube_vcard
 
     return $vcard;
   }
+  
+  private static function x_abrelatednames_callback($matches)
+  {
+    return 'X-' . strtoupper($matches[5]) . $matches[3] . ':'. $matches[4];
+  }
 
   private static function rfc2425_fold_callback($matches)
   {
@@ -631,6 +653,12 @@ class rcube_vcard
     if (substr($string, 0, 2) == "\xFF\xFE")     return 'UTF-16LE';  // Little Endian
     if (substr($string, 0, 3) == "\xEF\xBB\xBF") return 'UTF-8';
 
+    // heuristics
+    if ($string[0] == "\0" && $string[1] == "\0" && $string[2] == "\0" && $string[3] != "\0") return 'UTF-32BE';
+    if ($string[0] != "\0" && $string[1] == "\0" && $string[2] == "\0" && $string[3] == "\0") return 'UTF-32LE';
+    if ($string[0] == "\0" && $string[1] != "\0" && $string[2] == "\0" && $string[3] != "\0") return 'UTF-16BE';
+    if ($string[0] != "\0" && $string[1] == "\0" && $string[2] != "\0" && $string[3] == "\0") return 'UTF-16LE';
+
     // use mb_detect_encoding()
     $encodings = array('UTF-8', 'ISO-8859-1', 'ISO-8859-2', 'ISO-8859-3',
       'ISO-8859-4', 'ISO-8859-5', 'ISO-8859-6', 'ISO-8859-7', 'ISO-8859-8', 'ISO-8859-9',

program/steps/addressbook/func.inc

@@ -365,6 +365,10 @@ function rcmail_contact_form($form, $record, $attrib = null)
                     // skip cols unknown to the backend
                     if (!$coltypes[$col])
                         continue;
+                    
+                    // only string values are expected here
+                    if (is_array($record[$col]))
+                        $record[$col] = join(' ', $record[$col]);
 
                     if ($RCMAIL->action == 'show') {
                         if (!empty($record[$col]))
@@ -563,7 +567,9 @@ function rcmail_contact_photo($attrib)
     if ($CONTACT_COLTYPES['photo']) {
         $RCMAIL->output->set_env('photo_placeholder', $photo_img);
         
-        if ($record['photo'])
+        if (strpos($record['photo'], 'http:') === 0)
+            $photo_img = $record['photo'];
+        else if ($record['photo'])
             $photo_img = $RCMAIL->url(array('_action' => 'photo', '_cid' => $record['ID'], '_source' => $_REQUEST['_source']));
         $img = html::img(array('src' => $photo_img, 'border' => 1, 'alt' => ''));
         $content = html::div($attrib, $img);

program/steps/addressbook/import.inc

@@ -138,7 +138,7 @@ if ($_FILES['_file']['tmp_name'] && is_uploaded_file($_FILES['_file']['tmp_name'
       // We're using UTF8 internally
       $email = rcube_idn_to_utf8($email);
       
-      if (!$replace) {
+      if (!$replace && $email) {
         // compare e-mail address
         $existing = $CONTACTS->search('email', $email, false, false);
         if (!$existing->count) {  // compare display name

tests/vcards.php

@@ -54,4 +54,12 @@ class rcube_test_vcards extends UnitTestCase
     $this->assertEqual("Iksiñski", $vcards2[0]->surname, "Detect charset in encoded values");
   }
   
+  function test_encodings()
+  {
+      $input = file_get_contents($this->_srcpath('utf-16_sample.vcf'));
+      
+      $vcards = rcube_vcard::import($input);
+      $this->assertEqual("Ǽgean ĽdaMonté", $vcards[0]->displayname, "Decoded from UTF-16");
+  }
+  
 }