Changeset 11bcac5 in github

Timestamp:

Sep 29, 2010 9:32:41 AM (3 years ago)

Author:

alecpl <alec@…>

Branches:

master, HEAD, courier-fix, dev-browser-capabilities, pdo, release-0.6, release-0.7, release-0.8

Children:

c609784

Parents:

b3660bb

Message:

• Fix HTML to plain text conversion doesn't handle citation blocks (#1486921)

Files:

• CHANGELOG (modified) (1 diff)
• program/lib/html2text.php (modified) (9 diffs)
• program/steps/utils/html2text.inc (modified) (1 diff)

CHANGELOG

@@ -5 +5 @@
 - Add link to identities in compose window (#1486729)
 - Add Internationalized Domain Name (IDNA) support (#1483894)
-- Add option to automatically send read notifications for known senders (1485883)
-- Add option to "Return receipt" will be always checked (1486352)
+- Add option to automatically send read notifications for known senders (#1485883)
+- Add option to "Return receipt" will be always checked (#1486352)
+- Fix HTML to plain text conversion doesn't handle citation blocks (#1486921)
 
 RELEASE 0.4.1

program/lib/html2text.php

@@ -202 +202 @@
         "\n\t* ",                               // <li>
         "\n-------------------------\n",        // <hr>
-        "<div>\n",                                   // <div>
+        "<div>\n",                              // <div>
         "\n\n",                                 // <table> and </table>
         "\n",                                   // <tr> and </tr>
@@ -446 +446 @@
 
     /**
+     *  Workhorse function that does actual conversion (calls _converter() method).
+     *
+     *  @access private
+     *  @return void
+     */
+    function _convert()
+    {
+        // Variables used for building the link list
+        $this->_link_count = 0;
+        $this->_link_list = '';
+
+        $text = trim(stripslashes($this->html));
+
+        // Convert HTML to TXT
+        $this->_converter($text);
+
+        // Add link list
+        if ( !empty($this->_link_list) ) {
+            $text .= "\n\nLinks:\n------\n" . $this->_link_list;
+        }
+
+        $this->text = $text;
+
+        $this->_converted = true;
+    }
+
+    /**
      *  Workhorse function that does actual conversion.
      *
@@ -453 +480 @@
      *  $width characters.
      *
-     *  @access private
-     *  @return void
-     */
-    function _convert()
-    {
-        // Variables used for building the link list
-        $this->_link_count = 0;
-        $this->_link_list = '';
-
-        $text = trim(stripslashes($this->html));
+     *  @param string Reference to HTML content string
+     *
+     *  @access private
+     *  @return void
+     */
+    function _converter(&$text)
+    {
+        // Convert <BLOCKQUOTE> (before PRE!)
+        $this->_convert_blockquotes($text);
 
         // Convert <PRE>
@@ -485 +511 @@
         $text = preg_replace("/\n\s+\n/", "\n\n", $text);
         $text = preg_replace("/[\n]{3,}/", "\n\n", $text);
-
-        // Add link list
-        if ( !empty($this->_link_list) ) {
-            $text .= "\n\nLinks:\n------\n" . $this->_link_list;
-        }
 
         // Wrap the text to a readable format
@@ -497 +518 @@
                 $text = wordwrap($text, $this->width);
         }
-
-        $this->text = $text;
-
-        $this->_converted = true;
     }
 
@@ -518 +535 @@
     function _build_link_list( $link, $display )
     {
-        if ( !$this->_do_links ) return $display;
-        
-        if ( substr($link, 0, 7) == 'http://' || substr($link, 0, 8) == 'https://' ||
-             substr($link, 0, 7) == 'mailto:' ) {
+            if ( !$this->_do_links )
+                return $display;
+
+            if ( substr($link, 0, 7) == 'http://' || substr($link, 0, 8) == 'https://' ||
+            substr($link, 0, 7) == 'mailto:'
+        ) {
             $this->_link_count++;
-            $this->_link_list .= "[" . $this->_link_count . "] $link\n";
+            $this->_link_list .= '[' . $this->_link_count . "] $link\n";
             $additional = ' [' . $this->_link_count . ']';
-        } elseif ( substr($link, 0, 11) == 'javascript:' ) {
-                // Don't count the link; ignore it
-                $additional = '';
+            } elseif ( substr($link, 0, 11) == 'javascript:' ) {
+                    // Don't count the link; ignore it
+                    $additional = '';
                 // what about href="#anchor" ?
         } else {
             $this->_link_count++;
-            $this->_link_list .= "[" . $this->_link_count . "] " . $this->url;
+            $this->_link_list .= '[' . $this->_link_count . '] ' . $this->url;
             if ( substr($link, 0, 1) != '/' ) {
                 $this->_link_list .= '/';
@@ -541 +560 @@
         return $display . $additional;
     }
-    
+
     /**
      *  Helper function for PRE body conversion.
@@ -550 +569 @@
     function _convert_pre(&$text)
     {
-        while(preg_match('/<pre[^>]*>(.*)<\/pre>/ismU', $text, $matches)) {
+        while (preg_match('/<pre[^>]*>(.*)<\/pre>/ismU', $text, $matches)) {
             $result = preg_replace($this->pre_search, $this->pre_replace, $matches[1]);
             $text = preg_replace('/<pre[^>]*>.*<\/pre>/ismU', '<div><br>' . $result . '<br></div>', $text, 1);
+        }
+    }
+
+    /**
+     *  Helper function for BLOCKQUOTE body conversion.
+     *
+     *  @param string HTML content
+     *  @access private
+     */
+    function _convert_blockquotes(&$text)
+    {
+        if (preg_match_all('/<\/*blockquote[^>]*>/i', $text, $matches, PREG_OFFSET_CAPTURE)) {
+            $level = 0;
+            $diff = 0;
+            foreach ($matches[0] as $m) {
+                if ($m[0][0] == '<' && $m[0][1] == '/') {
+                    $level--;
+                    if ($level < 0) {
+                        $level = 0; // malformed HTML: go to next blockquote
+                    }
+                    else if ($level > 0) {
+                        // skip inner blockquote
+                    }
+                    else {
+                        $end  = $m[1];
+                        $len  = $end - $taglen - $start;
+                        // Get blockquote content
+                        $body = substr($text, $start + $taglen - $diff, $len);
+
+                        // Set text width
+                        $p_width = $this->width;
+                        if ($this->width > 0) $this->width -= 2;
+                        // Convert blockquote content
+                        $body = trim($body);
+                        $this->_converter($body);
+                        // Add citation markers and create PRE block
+                        $body = preg_replace('/((^|\n)>*)/', '\\1> ', trim($body));
+                        $body = '<pre>' . htmlspecialchars($body) . '</pre>';
+                        // Re-set text width
+                        $this->width = $p_width;
+                        // Replace content
+                        $text = substr($text, 0, $start - $diff)
+                            . $body . substr($text, $end + strlen($m[0]) - $diff);
+
+                        $diff = $len + $taglen + strlen($m[0]) - strlen($body);
+                        unset($body);
+                    }
+                }
+                else {
+                    if ($level == 0) {
+                        $start = $m[1];
+                        $taglen = strlen($m[0]);
+                    }
+                    $level ++;
+                }
+            }
         }
     }
@@ -593 +668 @@
     }
 }
-
-?>

program/steps/utils/html2text.inc

@@ -23 +23 @@
 
 header('Content-Type: text/plain; charset=UTF-8');
-print trim($converter->get_text());
+print rtrim($converter->get_text());
 exit;