Add support for special characters in \href

Resolves issue #9.

Author: xemlock

library/PhpLatex/Node.php

@@ -7,6 +7,8 @@ class PhpLatex_Node
 {
     protected $_type;
     protected $_props;
+
+    /** @var PhpLatex_Node[] */
     protected $_children = array();
 
     /**
@@ -65,7 +67,7 @@ public function getChild($index)
     }
 
     /**
-     * @return array
+     * @return PhpLatex_Node[]
      */
     public function getChildren()
     {

library/PhpLatex/Parser.php

@@ -581,6 +581,11 @@ protected function _parseControl($token, $mode, $environ = null) // {{{
             // parse arguments
             $numArgs = isset($spec['numArgs']) ? intval($spec['numArgs']) : 0;
 
+            // handle special cases for \url and \href commands
+            if ($value === '\\href' || $value === '\\url') {
+                $nodeArgs[] = $this->_parseUrl();
+            }
+
             while (count($nodeArgs) < $numArgs) {
                 if (false === ($arg = $this->_parseArg($nodeMode, $environ, $parseArgs))) {
                     // no argument found, create an artificial one
@@ -1047,6 +1052,64 @@ protected function _parseLeftRight($token, $mode, $environs)
         return $node;
     }
 
+    /**
+     * Parse URL agrument of \href and \url commands
+     *
+     * @return PhpLatex_Node
+     */
+    protected function _parseUrl()
+    {
+        $content = '';
+        // skip comments and spaces before '{'
+        $this->_skipSpacesAndComments();
+
+        $arg = $this->_createNode(self::TYPE_GROUP, self::MODE_TEXT);
+
+        // expect an argument to be present
+        $next = $this->_peek();
+        if ($next['type'] !== PhpLatex_Lexer::TYPE_SPECIAL || $next['value'] !== '{') {
+            return $arg;
+        } else {
+            $this->_next();
+        }
+
+        // keep track of nested curly brackets
+        $nesting = 0;
+        while ($next = $this->_next()) {
+            if ($next['type'] === PhpLatex_Lexer::TYPE_SPECIAL) {
+                if ($next['value'] === '{') {
+                    ++$nesting;
+                } elseif ($next['value'] === '}') {
+                    // check if it's the end of the nested group, or the end of url argument
+                    if ($nesting === 0) {
+                        // end of url argument
+                        break;
+                    }
+                    --$nesting;
+                }
+            }
+
+            // unescape symbols
+            // in the urls backslash is escaped simply as '\\'
+            if ($next['type'] === PhpLatex_Lexer::TYPE_CSYMBOL) {
+                $content .= substr($next['value'], 1);
+            } else {
+                $content .= isset($next['raw']) ? $next['raw'] : $next['value'];
+            }
+        }
+
+        $content = trim($content);
+
+        // LF in urls are ignored
+        $content = str_replace("\n", '', $content);
+
+        $url = $this->_createNode(self::TYPE_TEXT, self::MODE_TEXT);
+        $url->value = $content;
+        $arg->addChild($url);
+
+        return $arg;
+    }
+
     /**
      * @param array $token
      * @return bool

library/PhpLatex/Renderer/Abstract.php

@@ -55,6 +55,32 @@ public static function toLatex($node) // {{{
                         }
                         return $value;
                     }
+
+                    if ($node->value === '\\href' || $node->value === '\\url') {
+                        foreach ($node->getChildren() as $index => $child) {
+                            if ($index === 0) {
+                                $url = $child->getChild(0);
+                                if ($url) {
+                                    // The special characters # and % do not need to be escaped in any way (unless
+                                    // the command is used in the argument of another command).
+                                    // https://texdoc.org/serve/hyperref/0
+
+                                    // Only control words (i.e. starting with a backslash) cause miscellaneous errors, among others:
+                                    // - You've closed more groups than you opened.
+                                    // - TeX capacity exceeded, sorry
+                                    // - Paragraph ended before \hyper@n@rmalise was complete.
+                                    // all other characters, normally treated as special, lose their meanings inside \href and \url
+                                    $value .= '{' . str_replace('\\', '\\\\', (string) $url->value) . '}';
+                                } else {
+                                    $value .= '{}';
+                                }
+                            } else {
+                                $value .= self::toLatex($child);
+                            }
+                        }
+                        return $value;
+                    }
+
                     if ($node->symbol || $node->hasChildren()) {
                         return $value . self::toLatex($node->getChildren());
                     }

tests/PhpLatex/Parser/HrefTest.php

@@ -0,0 +1,38 @@
+<?php
+
+class PhpLatex_Parser_HrefTest extends PHPUnit_Framework_TestCase
+{
+    /**
+     * @var PhpLatex_Parser
+     */
+    protected $parser;
+
+    protected function setUp()
+    {
+        parent::setUp();
+        $this->parser = new PhpLatex_Parser();
+    }
+
+    /**
+     * @see https://github.com/xemlock/php-latex/issues/9
+     */
+    function testParse()
+    {
+        $input = '\href{https://ja.wikipedia.org/wiki/%E9%9B%BB%E5%AD%90?utm_source=test&utm_medium=email}{electron}';
+        $tree = $this->parser->parse($input);
+        $this->assertSame($input, PhpLatex_Renderer_Abstract::toLatex($tree));
+    }
+
+    function testSpecialSymbols()
+    {
+        $input = '\url{https://test.com/~user{a}$2%20x#test}';
+        $tree = $this->parser->parse($input);
+        $this->assertSame($input, PhpLatex_Renderer_Abstract::toLatex($tree));
+    }
+
+    function testBackslash() {
+        $input = '\href{run:C:\path\to\script.bat}{File}';
+        $tree = $this->parser->parse($input);
+        $this->assertSame('\href{run:C:\\\\path\\\\to\\\\script.bat}{File}', PhpLatex_Renderer_Abstract::toLatex($tree));
+    }
+}