GeSHi updated to 1.0.8.10

1 files changed, 107 insertions, 8 deletions

diff --git a/inc/geshi.php b/inc/geshi.php
index 31d2da49f..aedc64f84 100644
--- a/inc/geshi.php
+++ b/inc/geshi.php
@@ -41,7 +41,7 @@
 //
 
 /** The version of this GeSHi file */
-define('GESHI_VERSION', '1.0.8.8');
+define('GESHI_VERSION', '1.0.8.10');
 
 // Define the root directory for the GeSHi code tree
 if (!defined('GESHI_ROOT')) {
@@ -209,12 +209,16 @@ define('GESHI_NUMBER_BIN_PREFIX_0B', 64);        //0b[01]+
 define('GESHI_NUMBER_OCT_PREFIX', 256);           //0[0-7]+
 /** Number format to highlight octal numbers with a prefix 0o (logtalk) */
 define('GESHI_NUMBER_OCT_PREFIX_0O', 512);           //0[0-7]+
+/** Number format to highlight octal numbers with a leading @ (Used in HiSofts Devpac series). */
+define('GESHI_NUMBER_OCT_PREFIX_AT', 1024);           //@[0-7]+
 /** Number format to highlight octal numbers with a suffix of o */
-define('GESHI_NUMBER_OCT_SUFFIX', 1024);           //[0-7]+[oO]
+define('GESHI_NUMBER_OCT_SUFFIX', 2048);           //[0-7]+[oO]
 /** Number format to highlight hex numbers with a prefix 0x */
 define('GESHI_NUMBER_HEX_PREFIX', 4096);           //0x[0-9a-fA-F]+
+/** Number format to highlight hex numbers with a prefix $ */
+define('GESHI_NUMBER_HEX_PREFIX_DOLLAR', 8192);           //$[0-9a-fA-F]+
 /** Number format to highlight hex numbers with a suffix of h */
-define('GESHI_NUMBER_HEX_SUFFIX', 8192);           //[0-9][0-9a-fA-F]*h
+define('GESHI_NUMBER_HEX_SUFFIX', 16384);           //[0-9][0-9a-fA-F]*h
 /** Number format to highlight floating-point numbers without support for scientific notation */
 define('GESHI_NUMBER_FLT_NONSCI', 65536);          //\d+\.\d+
 /** Number format to highlight floating-point numbers without support for scientific notation */
@@ -732,6 +736,88 @@ class GeSHi {
     }
 
     /**
+     * Get supported langs or an associative array lang=>full_name.
+     * @param boolean $longnames
+     * @return array
+     */
+    function get_supported_languages($full_names=false)
+    {
+        // return array
+        $back = array();
+
+        // we walk the lang root
+        $dir = dir($this->language_path);
+
+        // foreach entry
+        while (false !== ($entry = $dir->read()))
+        {
+            $full_path = $this->language_path.$entry;
+
+            // Skip all dirs
+            if (is_dir($full_path)) {
+                continue;
+            }
+
+            // we only want lang.php files
+            if (!preg_match('/^([^.]+)\.php$/', $entry, $matches)) {
+                continue;
+            }
+
+            // Raw lang name is here
+            $langname = $matches[1];
+
+            // We want the fullname too?
+            if ($full_names === true)
+            {
+                if (false !== ($fullname = $this->get_language_fullname($langname)))
+                {
+                    $back[$langname] = $fullname; // we go associative
+                }
+            }
+            else
+            {
+                // just store raw langname
+                $back[] = $langname;
+            }
+        }
+
+        $dir->close();
+
+        return $back;
+    }
+
+    /**
+     * Get full_name for a lang or false.
+     * @param string $language short langname (html4strict for example)
+     * @return mixed
+     */
+    function get_language_fullname($language)
+    {
+        //Clean up the language name to prevent malicious code injection
+        $language = preg_replace('#[^a-zA-Z0-9\-_]#', '', $language);
+
+        $language = strtolower($language);
+
+        // get fullpath-filename for a langname
+        $fullpath = $this->language_path.$language.'.php';
+
+        // we need to get contents :S
+        if (false === ($data = file_get_contents($fullpath))) {
+            $this->error = sprintf('Geshi::get_lang_fullname() Unknown Language: %s', $language);
+            return false;
+        }
+
+        // match the langname
+        if (!preg_match('/\'LANG_NAME\'\s*=>\s*\'((?:[^\']|\\\')+)\'/', $data, $matches)) {
+            $this->error = sprintf('Geshi::get_lang_fullname(%s): Regex can not detect language', $language);
+            return false;
+        }
+
+        // return fullname for langname
+        return stripcslashes($matches[1]);
+    }
+
+    /**
      * Sets the type of header to be used.
      *
      * If GESHI_HEADER_DIV is used, the code is surrounded in a "div".This
@@ -1353,6 +1439,10 @@ class GeSHi {
     function get_language_name_from_extension( $extension, $lookup = array() ) {
         if ( !is_array($lookup) || empty($lookup)) {
             $lookup = array(
+                '6502acme' => array( 'a', 's', 'asm', 'inc' ),
+                '6502tasm' => array( 'a', 's', 'asm', 'inc' ),
+                '6502kickass' => array( 'a', 's', 'asm', 'inc' ),
+                '68000devpac' => array( 'a', 's', 'asm', 'inc' ),
                 'abap' => array('abap'),
                 'actionscript' => array('as'),
                 'ada' => array('a', 'ada', 'adb', 'ads'),
@@ -1971,7 +2061,7 @@ class GeSHi {
             //All this formats are matched case-insensitively!
             static $numbers_format = array(
                 GESHI_NUMBER_INT_BASIC =>
-                    '(?:(?<![0-9a-z_\.%])|(?<=\.\.))(?<![\d\.]e[+\-])([1-9]\d*?|0)(?![0-9a-z]|\.(?:[eE][+\-]?)?\d)',
+                    '(?:(?<![0-9a-z_\.%$@])|(?<=\.\.))(?<![\d\.]e[+\-])([1-9]\d*?|0)(?![0-9a-z]|\.(?:[eE][+\-]?)?\d)',
                 GESHI_NUMBER_INT_CSTYLE =>
                     '(?<![0-9a-z_\.%])(?<![\d\.]e[+\-])([1-9]\d*?|0)l(?![0-9a-z]|\.(?:[eE][+\-]?)?\d)',
                 GESHI_NUMBER_BIN_SUFFIX =>
@@ -1984,10 +2074,14 @@ class GeSHi {
                     '(?<![0-9a-z_\.])(?<![\d\.]e[+\-])0[0-7]+?(?![0-9a-z]|\.(?:[eE][+\-]?)?\d)',
                 GESHI_NUMBER_OCT_PREFIX_0O =>
                     '(?<![0-9a-z_\.%])(?<![\d\.]e[+\-])0o[0-7]+?(?![0-9a-z]|\.(?:[eE][+\-]?)?\d)',
+                GESHI_NUMBER_OCT_PREFIX_AT =>
+                    '(?<![0-9a-z_\.%])(?<![\d\.]e[+\-])\@[0-7]+?(?![0-9a-z]|\.(?:[eE][+\-]?)?\d)',
                 GESHI_NUMBER_OCT_SUFFIX =>
                     '(?<![0-9a-z_\.])(?<![\d\.]e[+\-])[0-7]+?o(?![0-9a-z]|\.(?:[eE][+\-]?)?\d)',
                 GESHI_NUMBER_HEX_PREFIX =>
                     '(?<![0-9a-z_\.])(?<![\d\.]e[+\-])0x[0-9a-fA-F]+?(?![0-9a-z]|\.(?:[eE][+\-]?)?\d)',
+                GESHI_NUMBER_HEX_PREFIX_DOLLAR =>
+                    '(?<![0-9a-z_\.])(?<![\d\.]e[+\-])\$[0-9a-fA-F]+?(?![0-9a-z]|\.(?:[eE][+\-]?)?\d)',
                 GESHI_NUMBER_HEX_SUFFIX =>
                     '(?<![0-9a-z_\.])(?<![\d\.]e[+\-])\d[0-9a-fA-F]*?[hH](?![0-9a-z]|\.(?:[eE][+\-]?)?\d)',
                 GESHI_NUMBER_FLT_NONSCI =>
@@ -2021,6 +2115,10 @@ class GeSHi {
                 $this->language_data['NUMBERS_RXCACHE'][$key] =
                     "/(?<!<\|\/)(?<!<\|!REG3XP)(?<!<\|\/NUM!)(?<!\d\/>)($regexp)(?!(?:<DOT>|(?>[^\<]))+>)(?![^<]*>)(?!\|>)(?!\/>)/i"; //
             }
+
+            if(!isset($this->language_data['PARSER_CONTROL']['NUMBERS']['PRECHECK_RX'])) {
+                $this->language_data['PARSER_CONTROL']['NUMBERS']['PRECHECK_RX'] = '#\d#';
+            }
         }
 
         $this->parse_cache_built = true;
@@ -3241,7 +3339,7 @@ class GeSHi {
         $stuff_to_parse = ' ' . $this->hsc($stuff_to_parse);
 
         // Highlight keywords
-        $disallowed_before = "(?<![a-zA-Z0-9\$_\|\#;>|^&";
+        $disallowed_before = "(?<![a-zA-Z0-9\$_\|\#|^&";
         $disallowed_after = "(?![a-zA-Z0-9_\|%\\-&;";
         if ($this->lexic_permissions['STRINGS']) {
             $quotemarks = preg_quote(implode($this->language_data['QUOTEMARKS']), '/');
@@ -3299,7 +3397,7 @@ class GeSHi {
                     // Basically, we don't put the styles in yet because then the styles themselves will
                     // get highlighted if the language has a CSS keyword in it (like CSS, for example ;))
                     $stuff_to_parse = preg_replace_callback(
-                        "/$disallowed_before_local({$keywordset})(?!\<DOT\>(?:htm|php))$disallowed_after_local/$modifiers",
+                        "/$disallowed_before_local({$keywordset})(?!\<DOT\>(?:htm|php|aspx?))$disallowed_after_local/$modifiers",
                         array($this, 'handle_keyword_replace'),
                         $stuff_to_parse
                         );
@@ -3346,7 +3444,8 @@ class GeSHi {
 
         // Highlight numbers. As of 1.0.8 we support different types of numbers
         $numbers_found = false;
-        if ($this->lexic_permissions['NUMBERS'] && preg_match('#\d#', $stuff_to_parse )) {
+
+        if ($this->lexic_permissions['NUMBERS'] && preg_match($this->language_data['PARSER_CONTROL']['NUMBERS']['PRECHECK_RX'], $stuff_to_parse )) {
             $numbers_found = true;
 
             //For each of the formats ...
@@ -4465,7 +4564,7 @@ class GeSHi {
     * @access private
     */
     function optimize_regexp_list($list, $regexp_delimiter = '/') {
-        $regex_chars = array('.', '\\', '+', '*', '?', '[', '^', ']', '$',
+        $regex_chars = array('.', '\\', '+', '-', '*', '?', '[', '^', ']', '$',
             '(', ')', '{', '}', '=', '!', '<', '>', '|', ':', $regexp_delimiter);
         sort($list);
         $regexp_list = array('');