Factor out auto-paragraph to injector class.

git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1210 48356398-32a2-884e-a903-53898d9a118a
2025-07-30 19:00:10 +02:00 · 2007-06-23 18:50:41 +00:00
parent 3066ca357a
commit 9191877740
3 changed files with 161 additions and 132 deletions
--- a/library/HTMLPurifier/Injector/AutoParagraph.php
+++ b/library/HTMLPurifier/Injector/AutoParagraph.php
@@ -0,0 +1,132 @@
+<?php
+
+require_once 'HTMLPurifier/Injector.php';
+
+/**
+ * Injector that auto paragraphs text in the root node based on
+ * double-spacing.
+ */
+class HTMLPurifier_Injector_AutoParagraph extends HTMLPurifier_Injector
+{
+    
+    function handleText(&$token, $config, &$context) {
+        $dnl = PHP_EOL . PHP_EOL; // double-newline
+        $current_nesting =& $context->get('CurrentNesting');
+        // paragraphing is on
+        if (empty($current_nesting)) {
+            // we're in root node, great time to start a paragraph
+            // since we're also dealing with a text node
+            $result =& $context->get('OutputTokens');
+            $result[] = new HTMLPurifier_Token_Start('p');
+            $current_nesting[] = new HTMLPurifier_Token_Start('p');
+            $this->_splitText($token, $config, $context);
+        } else {
+            // we're not in root node, so let's see whether or not
+            // we're in a paragraph
+            
+            // losslessly access the parent element
+            $parent = array_pop($current_nesting);
+            $current_nesting[] = $parent;
+            
+            if ($parent->name === 'p') {
+                $this->_splitText($token, $config, $context);
+            }
+        }
+    }
+    
+    /**
+     * Sub-function for auto-paragraphing that takes a token and splits it 
+     * up into paragraphs unconditionally. Requires that a paragraph was
+     * already started
+     */
+    function _splitText(&$token, $config, &$context) {
+        $dnl = PHP_EOL . PHP_EOL; // double-newline
+        $definition = $config->getHTMLDefinition();
+        $current_nesting =& $context->get('CurrentNesting');
+        
+        $raw_paragraphs = explode($dnl, $token->data);
+        
+        $token = false; // token has been completely dismantled
+        
+        // remove empty paragraphs
+        $paragraphs = array();
+        foreach ($raw_paragraphs as $par) {
+            if (trim($par) !== '') $paragraphs[] = $par;
+        }
+        
+        $result =& $context->get('OutputTokens');
+        
+        if (empty($paragraphs) && count($raw_paragraphs) > 1) {
+            $result[] = new HTMLPurifier_Token_End('p');
+            array_pop($current_nesting);
+            return;
+        }
+        
+        foreach ($paragraphs as $data) {
+            $result[] = new HTMLPurifier_Token_Text($data);
+            $result[] = new HTMLPurifier_Token_End('p');
+            $result[] = new HTMLPurifier_Token_Start('p');
+        }
+        array_pop($result); // remove trailing start token
+        
+        // check the outside to determine whether or not end
+        // paragraph tag is needed (it's already there)
+        $end_paragraph = $this->_needsEndTag(
+            $context->get('InputTokens'),
+            $context->get('InputIndex'),
+            $definition
+        );
+        
+        if ($end_paragraph) {
+            // things are good as they stand, remove top-level parent
+            // that we deferred
+            array_pop($current_nesting);
+        } else {
+            // remove the ending tag, no nesting modifications necessary
+            array_pop($result);
+        }
+        
+    }
+    
+    /**
+     * Determines if up-coming code requires an end-paragraph tag,
+     * otherwise, keep the paragraph open (don't make another one)
+     * @protected
+     */
+    function _needsEndTag($tokens, $k, $definition) {
+        $end_paragraph = false;
+        for ($j = $k + 1; isset($tokens[$j]); $j++) {
+            if ($tokens[$j]->type == 'start' || $tokens[$j]->type == 'empty') {
+                if ($tokens[$j]->name == 'p') {
+                    $end_paragraph = true;
+                } else {
+                    $end_paragraph = isset($definition->info['p']->auto_close[$tokens[$j]->name]);
+                }
+                break;
+            } elseif ($tokens[$j]->type == 'text') {
+                if (!$tokens[$j]->is_whitespace) {
+                    $end_paragraph = false;
+                    break;
+                }
+            } elseif ($tokens[$j]->type == 'end') {
+                // nonsensical case
+                $end_paragraph = false;
+                break;
+            }
+        }
+        return $end_paragraph;
+    }
+    
+    function handleStart(&$token, $config, &$context) {
+        $current_nesting =& $context->get('CurrentNesting');
+        if (!empty($current_nesting)) return;
+        $definition = $config->getHTMLDefinition();
+        // to be replaced with new auto-auto_close algorithm
+        if (isset($definition->info['p']->auto_close[$token->name])) return;
+        $result =& $context->get('OutputTokens');
+        $result[] = $current_nesting[] = new HTMLPurifier_Token_Start('p');
+    }
+    
+}
+
+?>