diff --git a/serendipity_event_geshi/ChangeLog b/serendipity_event_geshi/ChangeLog index 0356ab3e..a6303635 100644 --- a/serendipity_event_geshi/ChangeLog +++ b/serendipity_event_geshi/ChangeLog @@ -1,3 +1,8 @@ +1.1: +----- + * Upgrade GeSHi to 1.0.9.0; minimal PHP version is now 7.0. + + 1.0: Another fix for PHP7.2 compatibility. 0.9.9: Fix constructor for compatibility with PHP7.2 diff --git a/serendipity_event_geshi/geshi.php b/serendipity_event_geshi/geshi.php index df3452eb..4effa488 100644 --- a/serendipity_event_geshi/geshi.php +++ b/serendipity_event_geshi/geshi.php @@ -28,9 +28,8 @@ * @package geshi * @subpackage core * @author Nigel McNie , Benny Baumann - * @copyright (C) 2004 - 2007 Nigel McNie, (C) 2007 - 2008 Benny Baumann + * @copyright (C) 2004 - 2007 Nigel McNie, (C) 2007 - 2014 Benny Baumann * @license http://gnu.org/copyleft/gpl.html GNU GPL - * */ // @@ -41,7 +40,7 @@ // /** The version of this GeSHi file */ -define('GESHI_VERSION', '1.0.8.1'); +define('GESHI_VERSION', '1.0.9.0'); // Define the root directory for the GeSHi code tree if (!defined('GESHI_ROOT')) { @@ -149,39 +148,6 @@ define('GESHI_CLASS', 5); /** Used in language files to mark comments */ define('GESHI_COMMENTS', 0); -/** Used to work around missing PHP features **/ -define('GESHI_PHP_PRE_433', !(version_compare(PHP_VERSION, '4.3.3') === 1)); - -/** make sure we can call stripos **/ -if (!function_exists('stripos')) { - // the offset param of preg_match is not supported below PHP 4.3.3 - if (GESHI_PHP_PRE_433) { - /** - * @ignore - */ - function stripos($haystack, $needle, $offset = null) { - if (!is_null($offset)) { - $haystack = substr($haystack, $offset); - } - if (preg_match('/'. preg_quote($needle, '/') . '/', $haystack, $match, PREG_OFFSET_CAPTURE)) { - return $match[0][1]; - } - return false; - } - } - else { - /** - * @ignore - */ - function stripos($haystack, $needle, $offset = null) { - if (preg_match('/'. preg_quote($needle, '/') . '/', $haystack, $match, PREG_OFFSET_CAPTURE, $offset)) { - return $match[0][1]; - } - return false; - } - } -} - /** some old PHP / PCRE subpatterns only support up to xxx subpatterns in regular expressions. Set this to false if your PCRE lib is up to date @see GeSHi->optimize_regexp_list() @@ -207,12 +173,18 @@ define('GESHI_NUMBER_BIN_PREFIX_PERCENT', 32); //%[01]+ define('GESHI_NUMBER_BIN_PREFIX_0B', 64); //0b[01]+ /** Number format to highlight octal numbers with a leading zero */ 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', 512); //[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 */ @@ -247,49 +219,48 @@ define('GESHI_ERROR_INVALID_LINE_NUMBER_TYPE', 5); * about how to use this class. * * @package geshi - * @author Nigel McNie , Benny Baumann - * @copyright (C) 2004 - 2007 Nigel McNie, (C) 2007 - 2008 Benny Baumann + * @author Nigel McNie + * @author Benny Baumann + * @copyright (C) 2004 - 2007 Nigel McNie, (C) 2007 - 2014 Benny Baumann */ class GeSHi { - /**#@+ - * @access private - */ + /** * The source code to highlight * @var string */ - var $source = ''; + protected $source = ''; /** * The language to use when highlighting * @var string */ - var $language = ''; + protected $language = ''; /** * The data for the language used * @var array */ - var $language_data = array(); + protected $language_data = array(); /** * The path to the language files * @var string */ - var $language_path = GESHI_LANG_ROOT; + protected $language_path = GESHI_LANG_ROOT; /** * The error message associated with an error * @var string * @todo check err reporting works */ - var $error = false; + protected $error = false; /** * Possible error messages * @var array */ - var $error_messages = array( + protected $error_messages = array( GESHI_ERROR_NO_SUCH_LANG => 'GeSHi could not find the language {LANGUAGE} (using path {PATH})', GESHI_ERROR_FILE_NOT_READABLE => 'The file specified for load_from_file was not readable', GESHI_ERROR_INVALID_HEADER_TYPE => 'The header type specified is invalid', @@ -300,13 +271,13 @@ class GeSHi { * Whether highlighting is strict or not * @var boolean */ - var $strict_mode = false; + protected $strict_mode = false; /** * Whether to use CSS classes in output * @var boolean */ - var $use_classes = false; + protected $use_classes = false; /** * The type of header to use. Can be one of the following @@ -318,13 +289,13 @@ class GeSHi { * * @var int */ - var $header_type = GESHI_HEADER_PRE; + protected $header_type = GESHI_HEADER_PRE; /** * Array of permissions for which lexics should be highlighted * @var array */ - var $lexic_permissions = array( + protected $lexic_permissions = array( 'KEYWORDS' => array(), 'COMMENTS' => array('MULTI' => true), 'REGEXPS' => array(), @@ -341,44 +312,44 @@ class GeSHi { * The time it took to parse the code * @var double */ - var $time = 0; + protected $time = 0; /** * The content of the header block * @var string */ - var $header_content = ''; + protected $header_content = ''; /** * The content of the footer block * @var string */ - var $footer_content = ''; + protected $footer_content = ''; /** * The style of the header block * @var string */ - var $header_content_style = ''; + protected $header_content_style = ''; /** * The style of the footer block * @var string */ - var $footer_content_style = ''; + protected $footer_content_style = ''; /** * Tells if a block around the highlighted source should be forced * if not using line numbering * @var boolean */ - var $force_code_block = false; + protected $force_code_block = false; /** * The styles for hyperlinks in the code * @var array */ - var $link_styles = array(); + protected $link_styles = array(); /** * Whether important blocks should be recognised or not @@ -386,7 +357,7 @@ class GeSHi { * @deprecated * @todo REMOVE THIS FUNCTIONALITY! */ - var $enable_important_blocks = false; + protected $enable_important_blocks = false; /** * Styles for important parts of the code @@ -395,31 +366,31 @@ class GeSHi { * @todo As above - rethink the whole idea of important blocks as it is buggy and * will be hard to implement in 1.2 */ - var $important_styles = 'font-weight: bold; color: red;'; // Styles for important parts of the code + protected $important_styles = 'font-weight: bold; color: red;'; // Styles for important parts of the code /** * Whether CSS IDs should be added to the code * @var boolean */ - var $add_ids = false; + protected $add_ids = false; /** * Lines that should be highlighted extra * @var array */ - var $highlight_extra_lines = array(); + protected $highlight_extra_lines = array(); /** * Styles of lines that should be highlighted extra * @var array */ - var $highlight_extra_lines_styles = array(); + protected $highlight_extra_lines_styles = array(); /** * Styles of extra-highlighted lines * @var string */ - var $highlight_extra_lines_style = 'background-color: #ffc;'; + protected $highlight_extra_lines_style = 'background-color: #ffc;'; /** * The line ending @@ -427,120 +398,120 @@ class GeSHi { * Otherwise, all instances of \n will be replaced with $line_ending * @var string */ - var $line_ending = null; + protected $line_ending = null; /** * Number at which line numbers should start at * @var int */ - var $line_numbers_start = 1; + protected $line_numbers_start = 1; /** * The overall style for this code block * @var string */ - var $overall_style = 'font-family:monospace;'; + protected $overall_style = 'font-family:monospace;'; /** * The style for the actual code * @var string */ - var $code_style = 'font: normal normal 1em/1.2em monospace; margin:0; padding:0; background:none; vertical-align:top;'; + protected $code_style = 'font: normal normal 1em/1.2em monospace; margin:0; padding:0; background:none; vertical-align:top;'; /** * The overall class for this code block * @var string */ - var $overall_class = ''; + protected $overall_class = ''; /** * The overall ID for this code block * @var string */ - var $overall_id = ''; + protected $overall_id = ''; /** * Line number styles * @var string */ - var $line_style1 = 'font-weight: normal; vertical-align:top;'; + protected $line_style1 = 'font-weight: normal; vertical-align:top;'; /** * Line number styles for fancy lines * @var string */ - var $line_style2 = 'font-weight: bold; vertical-align:top;'; + protected $line_style2 = 'font-weight: bold; vertical-align:top;'; /** * Style for line numbers when GESHI_HEADER_PRE_TABLE is chosen * @var string */ - var $table_linenumber_style = 'width:1px;text-align:right;margin:0;padding:0 2px;vertical-align:top;'; + protected $table_linenumber_style = 'width:1px;text-align:right;margin:0;padding:0 2px;vertical-align:top;'; /** * Flag for how line numbers are displayed * @var boolean */ - var $line_numbers = GESHI_NO_LINE_NUMBERS; + protected $line_numbers = GESHI_NO_LINE_NUMBERS; /** * Flag to decide if multi line spans are allowed. Set it to false to make sure * each tag is closed before and reopened after each linefeed. * @var boolean */ - var $allow_multiline_span = true; + protected $allow_multiline_span = true; /** * The "nth" value for fancy line highlighting * @var int */ - var $line_nth_row = 0; + protected $line_nth_row = 0; /** * The size of tab stops * @var int */ - var $tab_width = 8; + protected $tab_width = 8; /** * Should we use language-defined tab stop widths? * @var int */ - var $use_language_tab_width = false; + protected $use_language_tab_width = false; /** * Default target for keyword links * @var string */ - var $link_target = ''; + protected $link_target = ''; /** * The encoding to use for entity encoding * NOTE: Used with Escape Char Sequences to fix UTF-8 handling (cf. SF#2037598) * @var string */ - var $encoding = 'utf-8'; + protected $encoding = 'utf-8'; /** * Should keywords be linked? * @var boolean */ - var $keyword_links = true; + protected $keyword_links = true; /** * Currently loaded language file - * @var string + * @var string * @since 1.0.7.22 */ - var $loaded_language = ''; + protected $loaded_language = ''; /** * Wether the caches needed for parsing are built or not * - * @var bool + * @var bool * @since 1.0.8 */ - var $parse_cache_built = false; + protected $parse_cache_built = false; /** * Work around for Suhosin Patch with disabled /e modifier @@ -554,32 +525,30 @@ class GeSHi { * use preg_replace_callback() instead * * - * @var array + * @var array * @since 1.0.8 */ - var $_kw_replace_group = 0; - var $_rx_key = 0; + protected $_kw_replace_group = 0; + protected $_rx_key = 0; /** * some "callback parameters" for handle_multiline_regexps * - * @since 1.0.8 + * @since 1.0.8 * @access private - * @var string + * @var string */ - var $_hmr_before = ''; - var $_hmr_replace = ''; - var $_hmr_after = ''; - var $_hmr_key = 0; - - /**#@-*/ + protected $_hmr_before = ''; + protected $_hmr_replace = ''; + protected $_hmr_after = ''; + protected $_hmr_key = 0; /** * Creates a new GeSHi object, with source and language * - * @param string The source code to highlight - * @param string The language to highlight the source with - * @param string The path to the language file directory. This + * @param string $source The source code to highlight + * @param string $language The language to highlight the source with + * @param string $path The path to the language file directory. This * is deprecated! I've backported the auto path * detection from the 1.1.X dev branch, so now it * should be automatically set correctly. If you have @@ -588,32 +557,35 @@ class GeSHi { * {@link GeSHi->set_language_path()} * @since 1.0.0 */ - - function __construct($source = '', $language = '', $path = '') { - - if (!empty($source)) { + public function __construct($source = '', $language = '', $path = '') { + if ( is_string($source) && ($source !== '') ) { $this->set_source($source); } - if (!empty($language)) { + if ( is_string($language) && ($language !== '') ) { $this->set_language($language); } $this->set_language_path($path); } - - function GeSHi($source = '', $language = '', $path = '') { - self::__construct($source, $language, $path); + /** + * Returns the version of GeSHi + * + * @return string + * @since 1.0.8.11 + */ + public function get_version() + { + return GESHI_VERSION; } - /** * Returns an error message associated with the last GeSHi operation, - * or false if no error has occured + * or false if no error has occurred * * @return string|false An error message if there has been an error, else false * @since 1.0.0 */ - function error() { + public function error() { if ($this->error) { //Put some template variables for debugging here ... $debug_tpl_vars = array( @@ -637,7 +609,7 @@ class GeSHi { * @return string The name for the current language * @since 1.0.2 */ - function get_language_name() { + public function get_language_name() { if (GESHI_ERROR_NO_SUCH_LANG == $this->error) { return $this->language_data['LANG_NAME'] . ' (Unknown Language)'; } @@ -647,10 +619,10 @@ class GeSHi { /** * Sets the source code for this object * - * @param string The source code to highlight + * @param string $source The source code to highlight * @since 1.0.0 */ - function set_source($source) { + public function set_source($source) { $this->source = $source; $this->highlight_extra_lines = array(); } @@ -661,10 +633,14 @@ class GeSHi { * @note since 1.0.8 this function won't reset language-settings by default anymore! * if you need this set $force_reset = true * - * @param string The name of the language to use + * @param string $language The name of the language to use + * @param bool $force_reset * @since 1.0.0 */ - function set_language($language, $force_reset = false) { + public function set_language($language, $force_reset = false) { + $this->error = false; + $this->strict_mode = GESHI_NEVER; + if ($force_reset) { $this->loaded_language = false; } @@ -683,9 +659,6 @@ class GeSHi { $this->language = $language; - $this->error = false; - $this->strict_mode = GESHI_NEVER; - //Check if we can read the desired file if (!is_readable($file_name)) { $this->error = GESHI_ERROR_NO_SUCH_LANG; @@ -701,14 +674,14 @@ class GeSHi { * that this path is relative to the directory of the script that included * geshi.php, NOT geshi.php itself. * - * @param string The path to the language directory + * @param string $path The path to the language directory * @since 1.0.0 * @deprecated The path to the language files should now be automatically * detected, so this method should no longer be needed. The * 1.1.X branch handles manual setting of the path differently * so this method will disappear in 1.2.0. */ - function set_language_path($path) { + public function set_language_path($path) { if(strpos($path,':')) { //Security Fix to prevent external directories using fopen wrappers. if(DIRECTORY_SEPARATOR == "\\") { @@ -737,6 +710,88 @@ class GeSHi { } } + /** + * Get supported langs or an associative array lang=>full_name. + * @param boolean $full_names + * @return array + */ + public function get_supported_languages($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 $language short langname (html4strict for example) + * @return mixed + */ + public function get_language_fullname($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. * @@ -748,10 +803,10 @@ class GeSHi { * From 1.0.7.2, you can use GESHI_HEADER_NONE to specify that no header code * should be outputted. * - * @param int The type of header to be used + * @param int $type The type of header to be used * @since 1.0.0 */ - function set_header_type($type) { + public function set_header_type($type) { //Check if we got a valid header type if (!in_array($type, array(GESHI_HEADER_NONE, GESHI_HEADER_DIV, GESHI_HEADER_PRE, GESHI_HEADER_PRE_VALID, GESHI_HEADER_PRE_TABLE))) { @@ -768,11 +823,11 @@ class GeSHi { * when this object is parsed. The style should be a * string of valid stylesheet declarations * - * @param string The overall style for the outputted code block - * @param boolean Whether to merge the styles with the current styles or not + * @param string $style The overall style for the outputted code block + * @param boolean $preserve_defaults Whether to merge the styles with the current styles or not * @since 1.0.0 */ - function set_overall_style($style, $preserve_defaults = false) { + public function set_overall_style($style, $preserve_defaults = false) { if (!$preserve_defaults) { $this->overall_style = $style; } else { @@ -785,10 +840,10 @@ class GeSHi { * class can then be used in a stylesheet to style this object's * output * - * @param string The class name to use for this block of code + * @param string $class The class name to use for this block of code * @since 1.0.0 */ - function set_overall_class($class) { + public function set_overall_class($class) { $this->overall_class = $class; } @@ -796,10 +851,10 @@ class GeSHi { * Sets the overall id for this block of code. This id can then * be used in a stylesheet to style this object's output * - * @param string The ID to use for this block of code + * @param string $id The ID to use for this block of code * @since 1.0.0 */ - function set_overall_id($id) { + public function set_overall_id($id) { $this->overall_id = $id; } @@ -807,10 +862,10 @@ class GeSHi { * Sets whether CSS classes should be used to highlight the source. Default * is off, calling this method with no arguments will turn it on * - * @param boolean Whether to turn classes on or not + * @param boolean $flag Whether to turn classes on or not * @since 1.0.0 */ - function enable_classes($flag = true) { + public function enable_classes($flag = true) { $this->use_classes = ($flag) ? true : false; } @@ -825,11 +880,11 @@ class GeSHi { * code will have the same style as the line number! Consult the * GeSHi documentation for more information about this. * - * @param string The style to use for actual code - * @param boolean Whether to merge the current styles with the new styles + * @param string $style The style to use for actual code + * @param boolean $preserve_defaults Whether to merge the current styles with the new styles * @since 1.0.2 */ - function set_code_style($style, $preserve_defaults = false) { + public function set_code_style($style, $preserve_defaults = false) { if (!$preserve_defaults) { $this->code_style = $style; } else { @@ -840,16 +895,16 @@ class GeSHi { /** * Sets the styles for the line numbers. * - * @param string The style for the line numbers that are "normal" - * @param string|boolean If a string, this is the style of the line + * @param string $style1 The style for the line numbers that are "normal" + * @param string|boolean $style2 If a string, this is the style of the line * numbers that are "fancy", otherwise if boolean then this * defines whether the normal styles should be merged with the * new normal styles or not - * @param boolean If set, is the flag for whether to merge the "fancy" + * @param boolean $preserve_defaults If set, is the flag for whether to merge the "fancy" * styles with the current styles or not * @since 1.0.2 */ - function set_line_style($style1, $style2 = '', $preserve_defaults = false) { + public function set_line_style($style1, $style2 = '', $preserve_defaults = false) { //Check if we got 2 or three parameters if (is_bool($style2)) { $preserve_defaults = $style2; @@ -879,11 +934,11 @@ class GeSHi { * are to be fancy. For example, if the value of this parameter is 5 then every * 5th line will be fancy. * - * @param int How line numbers should be displayed - * @param int Defines which lines are fancy + * @param int $flag How line numbers should be displayed + * @param int $nth_row Defines which lines are fancy * @since 1.0.0 */ - function enable_line_numbers($flag, $nth_row = 5) { + public function enable_line_numbers($flag, $nth_row = 5) { if (GESHI_NO_LINE_NUMBERS != $flag && GESHI_NORMAL_LINE_NUMBERS != $flag && GESHI_FANCY_LINE_NUMBERS != $flag) { $this->error = GESHI_ERROR_INVALID_LINE_NUMBER_TYPE; @@ -898,10 +953,10 @@ class GeSHi { * Set it to false if you want to manipulate the output or manually display * the code in an ordered list. * - * @param boolean Wether multiline spans are allowed or not + * @param boolean $flag Wether multiline spans are allowed or not * @since 1.0.7.22 */ - function enable_multiline_span($flag) { + public function enable_multiline_span($flag) { $this->allow_multiline_span = (bool) $flag; } @@ -911,7 +966,7 @@ class GeSHi { * @see enable_multiline_span * @return bool */ - function get_multiline_span() { + public function get_multiline_span() { return $this->allow_multiline_span; } @@ -920,18 +975,28 @@ class GeSHi { * true, then styles are merged with the default styles, with the * user defined styles having priority * - * @param int The key of the keyword group to change the styles of - * @param string The style to make the keywords - * @param boolean Whether to merge the new styles with the old or just + * @param int $key The key of the keyword group to change the styles of + * @param string $style The style to make the keywords + * @param boolean $preserve_defaults Whether to merge the new styles with the old or just * to overwrite them * @since 1.0.0 */ - function set_keyword_group_style($key, $style, $preserve_defaults = false) { + public function set_keyword_group_style($key, $style, $preserve_defaults = false) { //Set the style for this keyword group - if (!$preserve_defaults) { - $this->language_data['STYLES']['KEYWORDS'][$key] = $style; + if('*' == $key) { + foreach($this->language_data['STYLES']['KEYWORDS'] as $_key => $_value) { + if (!$preserve_defaults) { + $this->language_data['STYLES']['KEYWORDS'][$_key] = $style; + } else { + $this->language_data['STYLES']['KEYWORDS'][$_key] .= $style; + } + } } else { - $this->language_data['STYLES']['KEYWORDS'][$key] .= $style; + if (!$preserve_defaults) { + $this->language_data['STYLES']['KEYWORDS'][$key] = $style; + } else { + $this->language_data['STYLES']['KEYWORDS'][$key] .= $style; + } } //Update the lexic permissions @@ -943,11 +1008,11 @@ class GeSHi { /** * Turns highlighting on/off for a keyword group * - * @param int The key of the keyword group to turn on or off - * @param boolean Whether to turn highlighting for that group on or off + * @param int $key The key of the keyword group to turn on or off + * @param boolean $flag Whether to turn highlighting for that group on or off * @since 1.0.0 */ - function set_keyword_group_highlighting($key, $flag = true) { + public function set_keyword_group_highlighting($key, $flag = true) { $this->lexic_permissions['KEYWORDS'][$key] = ($flag) ? true : false; } @@ -956,28 +1021,38 @@ class GeSHi { * true, then styles are merged with the default styles, with the * user defined styles having priority * - * @param int The key of the comment group to change the styles of - * @param string The style to make the comments - * @param boolean Whether to merge the new styles with the old or just + * @param int $key The key of the comment group to change the styles of + * @param string $style The style to make the comments + * @param boolean $preserve_defaults Whether to merge the new styles with the old or just * to overwrite them * @since 1.0.0 */ - function set_comments_style($key, $style, $preserve_defaults = false) { - if (!$preserve_defaults) { - $this->language_data['STYLES']['COMMENTS'][$key] = $style; + public function set_comments_style($key, $style, $preserve_defaults = false) { + if('*' == $key) { + foreach($this->language_data['STYLES']['COMMENTS'] as $_key => $_value) { + if (!$preserve_defaults) { + $this->language_data['STYLES']['COMMENTS'][$_key] = $style; + } else { + $this->language_data['STYLES']['COMMENTS'][$_key] .= $style; + } + } } else { - $this->language_data['STYLES']['COMMENTS'][$key] .= $style; + if (!$preserve_defaults) { + $this->language_data['STYLES']['COMMENTS'][$key] = $style; + } else { + $this->language_data['STYLES']['COMMENTS'][$key] .= $style; + } } } /** * Turns highlighting on/off for comment groups * - * @param int The key of the comment group to turn on or off - * @param boolean Whether to turn highlighting for that group on or off + * @param int $key The key of the comment group to turn on or off + * @param boolean $flag Whether to turn highlighting for that group on or off * @since 1.0.0 */ - function set_comments_highlighting($key, $flag = true) { + public function set_comments_highlighting($key, $flag = true) { $this->lexic_permissions['COMMENTS'][$key] = ($flag) ? true : false; } @@ -986,26 +1061,27 @@ class GeSHi { * true, then styles are merged with the default styles, with the * user defined styles having priority * - * @param string The style to make the escape characters - * @param boolean Whether to merge the new styles with the old or just - * to overwrite them + * @param string $style The style to make the escape characters + * @param boolean $preserve_defaults Whether to merge the new styles with the old or just + * to overwrite them + * @param int $group Tells the group of symbols for which style should be set. * @since 1.0.0 */ - function set_escape_characters_style($style, $preserve_defaults = false) { + public function set_escape_characters_style($style, $preserve_defaults = false, $group = 0) { if (!$preserve_defaults) { - $this->language_data['STYLES']['ESCAPE_CHAR'][0] = $style; + $this->language_data['STYLES']['ESCAPE_CHAR'][$group] = $style; } else { - $this->language_data['STYLES']['ESCAPE_CHAR'][0] .= $style; + $this->language_data['STYLES']['ESCAPE_CHAR'][$group] .= $style; } } /** * Turns highlighting on/off for escaped characters * - * @param boolean Whether to turn highlighting for escape characters on or off + * @param boolean $flag Whether to turn highlighting for escape characters on or off * @since 1.0.0 */ - function set_escape_characters_highlighting($flag = true) { + public function set_escape_characters_highlighting($flag = true) { $this->lexic_permissions['ESCAPE_CHAR'] = ($flag) ? true : false; } @@ -1017,13 +1093,13 @@ class GeSHi { * This method is DEPRECATED: use set_symbols_style instead. * This method will be removed in 1.2.X * - * @param string The style to make the brackets - * @param boolean Whether to merge the new styles with the old or just + * @param string $style The style to make the brackets + * @param boolean $preserve_defaults Whether to merge the new styles with the old or just * to overwrite them * @since 1.0.0 * @deprecated In favour of set_symbols_style */ - function set_brackets_style($style, $preserve_defaults = false) { + public function set_brackets_style($style, $preserve_defaults = false) { if (!$preserve_defaults) { $this->language_data['STYLES']['BRACKETS'][0] = $style; } else { @@ -1037,11 +1113,11 @@ class GeSHi { * This method is DEPRECATED: use set_symbols_highlighting instead. * This method will be remove in 1.2.X * - * @param boolean Whether to turn highlighting for brackets on or off + * @param boolean $flag Whether to turn highlighting for brackets on or off * @since 1.0.0 * @deprecated In favour of set_symbols_highlighting */ - function set_brackets_highlighting($flag) { + public function set_brackets_highlighting($flag) { $this->lexic_permissions['BRACKETS'] = ($flag) ? true : false; } @@ -1050,13 +1126,13 @@ class GeSHi { * true, then styles are merged with the default styles, with the * user defined styles having priority * - * @param string The style to make the symbols - * @param boolean Whether to merge the new styles with the old or just - * to overwrite them - * @param int Tells the group of symbols for which style should be set. + * @param string $style The style to make the symbols + * @param boolean $preserve_defaults Whether to merge the new styles with the old or just + * to overwrite them + * @param int $group Tells the group of symbols for which style should be set. * @since 1.0.1 */ - function set_symbols_style($style, $preserve_defaults = false, $group = 0) { + public function set_symbols_style($style, $preserve_defaults = false, $group = 0) { // Update the style of symbols if (!$preserve_defaults) { $this->language_data['STYLES']['SYMBOLS'][$group] = $style; @@ -1073,10 +1149,10 @@ class GeSHi { /** * Turns highlighting on/off for symbols * - * @param boolean Whether to turn highlighting for symbols on or off + * @param boolean $flag Whether to turn highlighting for symbols on or off * @since 1.0.0 */ - function set_symbols_highlighting($flag) { + public function set_symbols_highlighting($flag) { // Update lexic permissions for this symbol group $this->lexic_permissions['SYMBOLS'] = ($flag) ? true : false; @@ -1089,54 +1165,76 @@ class GeSHi { * true, then styles are merged with the default styles, with the * user defined styles having priority * - * @param string The style to make the escape characters - * @param boolean Whether to merge the new styles with the old or just - * to overwrite them + * @param string $style The style to make the escape characters + * @param boolean $preserve_defaults Whether to merge the new styles with the old or just + * to overwrite them + * @param int $group Tells the group of strings for which style should be set. * @since 1.0.0 */ - function set_strings_style($style, $preserve_defaults = false) { + public function set_strings_style($style, $preserve_defaults = false, $group = 0) { if (!$preserve_defaults) { - $this->language_data['STYLES']['STRINGS'][0] = $style; + $this->language_data['STYLES']['STRINGS'][$group] = $style; } else { - $this->language_data['STYLES']['STRINGS'][0] .= $style; + $this->language_data['STYLES']['STRINGS'][$group] .= $style; } } /** * Turns highlighting on/off for strings * - * @param boolean Whether to turn highlighting for strings on or off + * @param boolean $flag Whether to turn highlighting for strings on or off * @since 1.0.0 */ - function set_strings_highlighting($flag) { + public function set_strings_highlighting($flag) { $this->lexic_permissions['STRINGS'] = ($flag) ? true : false; } + /** + * Sets the styles for strict code blocks. If $preserve_defaults is + * true, then styles are merged with the default styles, with the + * user defined styles having priority + * + * @param string $style The style to make the script blocks + * @param boolean $preserve_defaults Whether to merge the new styles with the old or just + * to overwrite them + * @param int $group Tells the group of script blocks for which style should be set. + * @since 1.0.8.4 + */ + public function set_script_style($style, $preserve_defaults = false, $group = 0) { + // Update the style of symbols + if (!$preserve_defaults) { + $this->language_data['STYLES']['SCRIPT'][$group] = $style; + } else { + $this->language_data['STYLES']['SCRIPT'][$group] .= $style; + } + } + /** * Sets the styles for numbers. If $preserve_defaults is * true, then styles are merged with the default styles, with the * user defined styles having priority * - * @param string The style to make the numbers - * @param boolean Whether to merge the new styles with the old or just - * to overwrite them + * @param string $style The style to make the numbers + * @param boolean $preserve_defaults Whether to merge the new styles with the old or just + * to overwrite them + * @param int $group Tells the group of numbers for which style should be set. * @since 1.0.0 */ - function set_numbers_style($style, $preserve_defaults = false) { + public function set_numbers_style($style, $preserve_defaults = false, $group = 0) { if (!$preserve_defaults) { - $this->language_data['STYLES']['NUMBERS'][0] = $style; + $this->language_data['STYLES']['NUMBERS'][$group] = $style; } else { - $this->language_data['STYLES']['NUMBERS'][0] .= $style; + $this->language_data['STYLES']['NUMBERS'][$group] .= $style; } } /** * Turns highlighting on/off for numbers * - * @param boolean Whether to turn highlighting for numbers on or off + * @param boolean $flag Whether to turn highlighting for numbers on or off * @since 1.0.0 */ - function set_numbers_highlighting($flag) { + public function set_numbers_highlighting($flag) { $this->lexic_permissions['NUMBERS'] = ($flag) ? true : false; } @@ -1147,13 +1245,13 @@ class GeSHi { * true, then styles are merged with the default styles, with the * user defined styles having priority * - * @param int The key of the object splitter to change the styles of - * @param string The style to make the methods - * @param boolean Whether to merge the new styles with the old or just - * to overwrite them + * @param int $key The key of the object splitter to change the styles of + * @param string $style The style to make the methods + * @param boolean $preserve_defaults Whether to merge the new styles with the old or just + * to overwrite them * @since 1.0.0 */ - function set_methods_style($key, $style, $preserve_defaults = false) { + public function set_methods_style($key, $style, $preserve_defaults = false) { if (!$preserve_defaults) { $this->language_data['STYLES']['METHODS'][$key] = $style; } else { @@ -1164,10 +1262,10 @@ class GeSHi { /** * Turns highlighting on/off for methods * - * @param boolean Whether to turn highlighting for methods on or off + * @param boolean $flag Whether to turn highlighting for methods on or off * @since 1.0.0 */ - function set_methods_highlighting($flag) { + public function set_methods_highlighting($flag) { $this->lexic_permissions['METHODS'] = ($flag) ? true : false; } @@ -1176,12 +1274,14 @@ class GeSHi { * true, then styles are merged with the default styles, with the * user defined styles having priority * - * @param string The style to make the regular expression matches - * @param boolean Whether to merge the new styles with the old or just - * to overwrite them + * @param string $key The style to make the regular expression matches + * @param boolean $style Whether to merge the new styles with the old or just + * to overwrite them + * @param bool $preserve_defaults Whether to merge the new styles with the old or just + * to overwrite them * @since 1.0.0 */ - function set_regexps_style($key, $style, $preserve_defaults = false) { + public function set_regexps_style($key, $style, $preserve_defaults = false) { if (!$preserve_defaults) { $this->language_data['STYLES']['REGEXPS'][$key] = $style; } else { @@ -1192,22 +1292,22 @@ class GeSHi { /** * Turns highlighting on/off for regexps * - * @param int The key of the regular expression group to turn on or off - * @param boolean Whether to turn highlighting for the regular expression group on or off + * @param int $key The key of the regular expression group to turn on or off + * @param boolean $flag Whether to turn highlighting for the regular expression group on or off * @since 1.0.0 */ - function set_regexps_highlighting($key, $flag) { + public function set_regexps_highlighting($key, $flag) { $this->lexic_permissions['REGEXPS'][$key] = ($flag) ? true : false; } /** * Sets whether a set of keywords are checked for in a case sensitive manner * - * @param int The key of the keyword group to change the case sensitivity of - * @param boolean Whether to check in a case sensitive manner or not + * @param int $key The key of the keyword group to change the case sensitivity of + * @param boolean $case Whether to check in a case sensitive manner or not * @since 1.0.0 */ - function set_case_sensitivity($key, $case) { + public function set_case_sensitivity($key, $case) { $this->language_data['CASE_SENSITIVE'][$key] = ($case) ? true : false; } @@ -1218,10 +1318,10 @@ class GeSHi { * - GESHI_CAPS_UPPER: convert all keywords to uppercase where found * - GESHI_CAPS_LOWER: convert all keywords to lowercase where found * - * @param int A constant specifying what to do with matched keywords + * @param int $case A constant specifying what to do with matched keywords * @since 1.0.1 */ - function set_case_keywords($case) { + public function set_case_keywords($case) { if (in_array($case, array( GESHI_CAPS_NO_CHANGE, GESHI_CAPS_UPPER, GESHI_CAPS_LOWER))) { $this->language_data['CASE_KEYWORDS'] = $case; @@ -1233,10 +1333,10 @@ class GeSHi { * * Widths below zero are ignored * - * @param int The tab width + * @param int $width The tab width * @since 1.0.0 */ - function set_tab_width($width) { + public function set_tab_width($width) { $this->tab_width = intval($width); //Check if it fit's the constraints: @@ -1249,10 +1349,10 @@ class GeSHi { /** * Sets whether or not to use tab-stop width specifed by language * - * @param boolean Whether to use language-specific tab-stop widths + * @param boolean $use Whether to use language-specific tab-stop widths * @since 1.0.7.20 */ - function set_use_language_tab_width($use) { + public function set_use_language_tab_width($use) { $this->use_language_tab_width = (bool) $use; } @@ -1263,7 +1363,7 @@ class GeSHi { * @return int Tab width * @since 1.0.7.20 */ - function get_real_tab_width() { + public function get_real_tab_width() { if (!$this->use_language_tab_width || !isset($this->language_data['TAB_WIDTH'])) { return $this->tab_width; @@ -1277,10 +1377,10 @@ class GeSHi { * method without parameters will turn it on. See documentation * for more details on strict mode and where to use it. * - * @param boolean Whether to enable strict mode or not + * @param boolean $mode Whether to enable strict mode or not * @since 1.0.0 */ - function enable_strict_mode($mode = true) { + public function enable_strict_mode($mode = true) { if (GESHI_MAYBE == $this->language_data['STRICT_MODE_APPLIES']) { $this->strict_mode = ($mode) ? GESHI_ALWAYS : GESHI_NEVER; } @@ -1293,7 +1393,7 @@ class GeSHi { * @todo Rewrite with array traversal * @deprecated In favour of enable_highlighting */ - function disable_highlighting() { + public function disable_highlighting() { $this->enable_highlighting(false); } @@ -1304,10 +1404,10 @@ class GeSHi { * to enable (true) or disable (false) all highlighting. * * @since 1.0.0 - * @param boolean A flag specifying whether to enable or disable all highlighting + * @param boolean $flag A flag specifying whether to enable or disable all highlighting * @todo Rewrite with array traversal */ - function enable_highlighting($flag = true) { + public function enable_highlighting($flag = true) { $flag = $flag ? true : false; foreach ($this->lexic_permissions as $key => $value) { if (is_array($value)) { @@ -1327,60 +1427,91 @@ class GeSHi { * Given a file extension, this method returns either a valid geshi language * name, or the empty string if it couldn't be found * - * @param string The extension to get a language name for - * @param array A lookup array to use instead of the default one + * @param string $extension The extension to get a language name for + * @param array $lookup A lookup array to use instead of the default one * @since 1.0.5 * @todo Re-think about how this method works (maybe make it private and/or make it * a extension->lang lookup?) - * @todo static? + * @return int|string */ - function get_language_name_from_extension( $extension, $lookup = array() ) { + public static function get_language_name_from_extension( $extension, $lookup = array() ) { + $extension = strtolower($extension); + 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'), 'apache' => array('conf'), - 'asm' => array('ash', 'asm'), + 'asm' => array('ash', 'asm', 'inc'), 'asp' => array('asp'), 'bash' => array('sh'), + 'bf' => array('bf'), 'c' => array('c', 'h'), 'c_mac' => array('c', 'h'), 'caddcl' => array(), 'cadlisp' => array(), 'cdfg' => array('cdfg'), 'cobol' => array('cbl'), - 'cpp' => array('cpp', 'h', 'hpp'), - 'csharp' => array(), + 'cpp' => array('cpp', 'hpp', 'C', 'H', 'CPP', 'HPP'), + 'csharp' => array('cs'), 'css' => array('css'), + 'd' => array('d'), 'delphi' => array('dpk', 'dpr', 'pp', 'pas'), + 'diff' => array('diff', 'patch'), 'dos' => array('bat', 'cmd'), + 'gdb' => array('kcrash', 'crash', 'bt'), 'gettext' => array('po', 'pot'), + 'gml' => array('gml'), + 'gnuplot' => array('plt'), + 'groovy' => array('groovy'), + 'haskell' => array('hs'), + 'haxe' => array('hx'), 'html4strict' => array('html', 'htm'), 'ini' => array('ini', 'desktop'), 'java' => array('java'), 'javascript' => array('js'), 'klonec' => array('kl1'), 'klonecpp' => array('klx'), + 'latex' => array('tex'), 'lisp' => array('lisp'), 'lua' => array('lua'), 'matlab' => array('m'), 'mpasm' => array(), + 'mysql' => array('sql'), 'nsis' => array(), 'objc' => array(), 'oobas' => array(), 'oracle8' => array(), - 'pascal' => array(), + 'oracle10' => array(), + 'pascal' => array('pas'), 'perl' => array('pl', 'pm'), 'php' => array('php', 'php5', 'phtml', 'phps'), + 'povray' => array('pov'), + 'providex' => array('pvc', 'pvx'), + 'prolog' => array('pl'), 'python' => array('py'), 'qbasic' => array('bi'), + 'reg' => array('reg'), + 'ruby' => array('rb'), 'sas' => array('sas'), + 'scala' => array('scala'), + 'scheme' => array('scm'), 'scilab' => array('sci'), + 'smalltalk' => array('st'), 'smarty' => array(), + 'tcl' => array('tcl'), + 'text' => array('txt'), 'vb' => array('bas'), 'vbnet' => array(), 'visualfoxpro' => array(), - 'xml' => array('xml') + 'whitespace' => array('ws'), + 'xml' => array('xml', 'svg', 'xrc'), + 'z80' => array('z80', 'asm', 'inc') ); } @@ -1389,7 +1520,8 @@ class GeSHi { return $lang; } } - return ''; + + return 'text'; } /** @@ -1404,15 +1536,15 @@ class GeSHi { * 'lang_name' ... * ); * - * @param string The filename to load the source from - * @param array A lookup array to use instead of the default one + * @param string $file_name The filename to load the source from + * @param array $lookup A lookup array to use instead of the default one * @todo Complete rethink of this and above method * @since 1.0.5 */ - function load_from_file($file_name, $lookup = array()) { + public function load_from_file($file_name, $lookup = array()) { if (is_readable($file_name)) { $this->set_source(file_get_contents($file_name)); - $this->set_language($this->get_language_name_from_extension(substr(strrchr($file_name, '.'), 1), $lookup)); + $this->set_language(self::get_language_name_from_extension(substr(strrchr($file_name, '.'), 1), $lookup)); } else { $this->error = GESHI_ERROR_FILE_NOT_READABLE; } @@ -1421,11 +1553,14 @@ class GeSHi { /** * Adds a keyword to a keyword group for highlighting * - * @param int The key of the keyword group to add the keyword to - * @param string The word to add to the keyword group + * @param int $key The key of the keyword group to add the keyword to + * @param string $word The word to add to the keyword group * @since 1.0.0 */ - function add_keyword($key, $word) { + public function add_keyword($key, $word) { + if (!is_array($this->language_data['KEYWORDS'][$key])) { + $this->language_data['KEYWORDS'][$key] = array(); + } if (!in_array($word, $this->language_data['KEYWORDS'][$key])) { $this->language_data['KEYWORDS'][$key][] = $word; @@ -1440,9 +1575,9 @@ class GeSHi { /** * Removes a keyword from a keyword group * - * @param int The key of the keyword group to remove the keyword from - * @param string The word to remove from the keyword group - * @param bool Wether to automatically recompile the optimized regexp list or not. + * @param int $key The key of the keyword group to remove the keyword from + * @param string $word The word to remove from the keyword group + * @param bool $recompile Wether to automatically recompile the optimized regexp list or not. * Note: if you set this to false and @see GeSHi->parse_code() was already called once, * for the current language, you have to manually call @see GeSHi->optimize_keyword_group() * or the removed keyword will stay in cache and still be highlighted! On the other hand @@ -1450,7 +1585,7 @@ class GeSHi { * remove a lot of keywords. * @since 1.0.0 */ - function remove_keyword($key, $word, $recompile = true) { + public function remove_keyword($key, $word, $recompile = true) { $key_to_remove = array_search($word, $this->language_data['KEYWORDS'][$key]); if ($key_to_remove !== false) { unset($this->language_data['KEYWORDS'][$key][$key_to_remove]); @@ -1465,13 +1600,14 @@ class GeSHi { /** * Creates a new keyword group * - * @param int The key of the keyword group to create - * @param string The styles for the keyword group - * @param boolean Whether the keyword group is case sensitive ornot - * @param array The words to use for the keyword group + * @param int $key The key of the keyword group to create + * @param string $styles The styles for the keyword group + * @param boolean $case_sensitive Whether the keyword group is case sensitive ornot + * @param array $words The words to use for the keyword group * @since 1.0.0 + * @return bool */ - function add_keyword_group($key, $styles, $case_sensitive = true, $words = array()) { + public function add_keyword_group($key, $styles, $case_sensitive = true, $words = array()) { $words = (array) $words; if (empty($words)) { // empty word lists mess up highlighting @@ -1488,15 +1624,16 @@ class GeSHi { if ($this->parse_cache_built) { $this->optimize_keyword_group($key); } + return true; } /** * Removes a keyword group * - * @param int The key of the keyword group to remove + * @param int $key The key of the keyword group to remove * @since 1.0.0 */ - function remove_keyword_group ($key) { + public function remove_keyword_group ($key) { //Remove the keyword group internally unset($this->language_data['KEYWORDS'][$key]); unset($this->lexic_permissions['KEYWORDS'][$key]); @@ -1510,51 +1647,70 @@ class GeSHi { /** * compile optimized regexp list for keyword group * - * @param int The key of the keyword group to compile & optimize + * @param int $key The key of the keyword group to compile & optimize * @since 1.0.8 */ - function optimize_keyword_group($key) { + public function optimize_keyword_group($key) { $this->language_data['CACHED_KEYWORD_LISTS'][$key] = $this->optimize_regexp_list($this->language_data['KEYWORDS'][$key]); + $space_as_whitespace = false; + if(isset($this->language_data['PARSER_CONTROL'])) { + if(isset($this->language_data['PARSER_CONTROL']['KEYWORDS'])) { + if(isset($this->language_data['PARSER_CONTROL']['KEYWORDS']['SPACE_AS_WHITESPACE'])) { + $space_as_whitespace = $this->language_data['PARSER_CONTROL']['KEYWORDS']['SPACE_AS_WHITESPACE']; + } + if(isset($this->language_data['PARSER_CONTROL']['KEYWORDS'][$key]['SPACE_AS_WHITESPACE'])) { + if(isset($this->language_data['PARSER_CONTROL']['KEYWORDS'][$key]['SPACE_AS_WHITESPACE'])) { + $space_as_whitespace = $this->language_data['PARSER_CONTROL']['KEYWORDS'][$key]['SPACE_AS_WHITESPACE']; + } + } + } + } + if($space_as_whitespace) { + foreach($this->language_data['CACHED_KEYWORD_LISTS'][$key] as $rxk => $rxv) { + $this->language_data['CACHED_KEYWORD_LISTS'][$key][$rxk] = + str_replace(" ", "\\s+", $rxv); + } + } } /** * Sets the content of the header block * - * @param string The content of the header block + * @param string $content The content of the header block * @since 1.0.2 */ - function set_header_content($content) { + public function set_header_content($content) { $this->header_content = $content; } /** * Sets the content of the footer block * - * @param string The content of the footer block + * @param string $content The content of the footer block * @since 1.0.2 */ - function set_footer_content($content) { + public function set_footer_content($content) { $this->footer_content = $content; } /** * Sets the style for the header content * - * @param string The style for the header content + * @param string $style The style for the header content * @since 1.0.2 */ - function set_header_content_style($style) { + public function set_header_content_style($style) { $this->header_content_style = $style; } /** * Sets the style for the footer content * - * @param string The style for the footer content + * @param string $style The style for the footer content * @since 1.0.2 */ - function set_footer_content_style($style) { + public function set_footer_content_style($style) { $this->footer_content_style = $style; } @@ -1562,81 +1718,81 @@ class GeSHi { * Sets whether to force a surrounding block around * the highlighted code or not * - * @param boolean Tells whether to enable or disable this feature + * @param boolean $flag Tells whether to enable or disable this feature * @since 1.0.7.20 */ - function enable_inner_code_block($flag) { + public function enable_inner_code_block($flag) { $this->force_code_block = (bool)$flag; } /** * Sets the base URL to be used for keywords * - * @param int The key of the keyword group to set the URL for - * @param string The URL to set for the group. If {FNAME} is in - * the url somewhere, it is replaced by the keyword - * that the URL is being made for + * @param int $group The key of the keyword group to set the URL for + * @param string $url The URL to set for the group. If {FNAME} is in + * the url somewhere, it is replaced by the keyword + * that the URL is being made for * @since 1.0.2 */ - function set_url_for_keyword_group($group, $url) { + public function set_url_for_keyword_group($group, $url) { $this->language_data['URLS'][$group] = $url; } /** * Sets styles for links in code * - * @param int A constant that specifies what state the style is being - * set for - e.g. :hover or :visited - * @param string The styles to use for that state + * @param int $type A constant that specifies what state the style is being + * set for - e.g. :hover or :visited + * @param string $styles The styles to use for that state * @since 1.0.2 */ - function set_link_styles($type, $styles) { + public function set_link_styles($type, $styles) { $this->link_styles[$type] = $styles; } /** * Sets the target for links in code * - * @param string The target for links in the code, e.g. _blank + * @param string $target The target for links in the code, e.g. _blank * @since 1.0.3 */ - function set_link_target($target) { + public function set_link_target($target) { if (!$target) { $this->link_target = ''; } else { - $this->link_target = ' target="' . $target . '" '; + $this->link_target = ' target="' . $target . '"'; } } /** * Sets styles for important parts of the code * - * @param string The styles to use on important parts of the code + * @param string $styles The styles to use on important parts of the code * @since 1.0.2 */ - function set_important_styles($styles) { + public function set_important_styles($styles) { $this->important_styles = $styles; } /** * Sets whether context-important blocks are highlighted * - * @param boolean Tells whether to enable or disable highlighting of important blocks + * @param boolean $flag Tells whether to enable or disable highlighting of important blocks * @todo REMOVE THIS SHIZ FROM GESHI! * @deprecated * @since 1.0.2 */ - function enable_important_blocks($flag) { + public function enable_important_blocks($flag) { $this->enable_important_blocks = ( $flag ) ? true : false; } /** * Whether CSS IDs should be added to each line * - * @param boolean If true, IDs will be added to each line. + * @param boolean $flag If true, IDs will be added to each line. * @since 1.0.2 */ - function enable_ids($flag = true) { + public function enable_ids($flag = true) { $this->add_ids = ($flag) ? true : false; } @@ -1645,16 +1801,16 @@ class GeSHi { * * The extra style parameter was added in 1.0.7.21. * - * @param mixed An array of line numbers to highlight, or just a line - * number on its own. - * @param string A string specifying the style to use for this line. - * If null is specified, the default style is used. - * If false is specified, the line will be removed from - * special highlighting + * @param mixed $lines An array of line numbers to highlight, or just a line + * number on its own. + * @param string $style A string specifying the style to use for this line. + * If null is specified, the default style is used. + * If false is specified, the line will be removed from + * special highlighting * @since 1.0.2 * @todo Some data replication here that could be cut down on */ - function highlight_lines_extra($lines, $style = null) { + public function highlight_lines_extra($lines, $style = null) { if (is_array($lines)) { //Split up the job using single lines at a time foreach ($lines as $line) { @@ -1668,7 +1824,7 @@ class GeSHi { //Decide on which style to use if ($style === null) { //Check if we should use default style unset($this->highlight_extra_lines_styles[$lines]); - } else if ($style === false) { //Check if to remove this line + } elseif ($style === false) { //Check if to remove this line unset($this->highlight_extra_lines[$lines]); unset($this->highlight_extra_lines_styles[$lines]); } else { @@ -1680,20 +1836,20 @@ class GeSHi { /** * Sets the style for extra-highlighted lines * - * @param string The style for extra-highlighted lines + * @param string $styles The style for extra-highlighted lines * @since 1.0.2 */ - function set_highlight_lines_extra_style($styles) { + public function set_highlight_lines_extra_style($styles) { $this->highlight_extra_lines_style = $styles; } /** * Sets the line-ending * - * @param string The new line-ending + * @param string $line_ending The new line-ending * @since 1.0.2 */ - function set_line_ending($line_ending) { + public function set_line_ending($line_ending) { $this->line_ending = (string)$line_ending; } @@ -1709,10 +1865,10 @@ class GeSHi { * has support for the CSS method, but (of course) IE doesn't * so it's not worth doing it the CSS way yet. * - * @param int The number to start line numbers at + * @param int $number The number to start line numbers at * @since 1.0.2 */ - function start_line_numbers_at($number) { + public function start_line_numbers_at($number) { $this->line_numbers_start = abs(intval($number)); } @@ -1725,10 +1881,10 @@ class GeSHi { * Maybe in a future version it may make a return for speed reasons, but * I doubt it. * - * @param string The encoding to use for the source + * @param string $encoding The encoding to use for the source * @since 1.0.3 */ - function set_encoding($encoding) { + public function set_encoding($encoding) { if ($encoding) { $this->encoding = strtolower($encoding); } @@ -1737,10 +1893,10 @@ class GeSHi { /** * Turns linking of keywords on or off. * - * @param boolean If true, links will be added to keywords + * @param boolean $enable If true, links will be added to keywords * @since 1.0.2 */ - function enable_keyword_links($enable = true) { + public function enable_keyword_links($enable = true) { $this->keyword_links = (bool) $enable; } @@ -1751,9 +1907,8 @@ class GeSHi { * preprocessed * * @since 1.0.8 - * @access private */ - function build_style_cache() { + protected function build_style_cache() { //Build the style cache needed to highlight numbers appropriate if($this->lexic_permissions['NUMBERS']) { //First check what way highlighting information for numbers are given @@ -1802,9 +1957,8 @@ class GeSHi { * This function makes stylesheet generators much faster as they do not need these caches. * * @since 1.0.8 - * @access private */ - function build_parse_cache() { + protected function build_parse_cache() { // cache symbol regexp //As this is a costy operation, we avoid doing it for multiple groups ... //Instead we perform it for all symbols at once. @@ -1840,7 +1994,7 @@ class GeSHi { $this->language_data['SYMBOL_DATA'][$symbols] = 0; if (isset($symbols[1])) { // multiple chars $symbol_preg_multi[] = preg_quote($symbols, '/'); - } else if ($symbols == '-') { + } elseif ($symbols == '-') { // don't trigger range out of order error $symbol_preg_single[] = '\-'; } else { // single char @@ -1913,31 +2067,37 @@ class GeSHi { //All this formats are matched case-insensitively! static $numbers_format = array( GESHI_NUMBER_INT_BASIC => - '(? - '(? - '(? - '(? - '(? - '(? + '(? + '(? - '(? - '(? + '(? - '(? - '(? - '(? - '(? - '(?language_data['NUMBERS_RXCACHE'][$key] = - "/(?)($regexp)(?!\|>)/i"; + "/(?)($regexp)(?!(?:|(?>[^\<]))+>)(?![^<]*>)(?!\|>)(?!\/>)/i"; + } + + if(!isset($this->language_data['PARSER_CONTROL']['NUMBERS']['PRECHECK_RX'])) { + $this->language_data['PARSER_CONTROL']['NUMBERS']['PRECHECK_RX'] = '#\d#'; } } @@ -1976,10 +2140,14 @@ class GeSHi { * * @since 1.0.0 */ - function parse_code () { + public function parse_code () { // Start the timer $start_time = microtime(); + // Replace all newlines to a common form. + $code = str_replace("\r\n", "\n", $this->source); + $code = str_replace("\r", "\n", $code); + // Firstly, if there is an error, we won't highlight if ($this->error) { //Escape the source for output @@ -2000,13 +2168,6 @@ class GeSHi { $this->build_parse_cache(); } - // Replace all newlines to a common form. - $code = str_replace("\r\n", "\n", $this->source); - $code = str_replace("\r", "\n", $code); - - // Add spaces for regular expression matching and line numbers -// $code = "\n" . $code . "\n"; - // Initialise various stuff $length = strlen($code); $COMMENT_MATCHED = false; @@ -2075,16 +2236,26 @@ class GeSHi { * - Group 1 is the opener * - Group 2 is the closer */ - if(!GESHI_PHP_PRE_433 && //Needs proper rewrite to work with PHP >=4.3.0; 4.3.3 is guaranteed to work. - preg_match($delimiters, $code, $matches_rx, PREG_OFFSET_CAPTURE, $i)) { + if(preg_match($delimiters, $code, $matches_rx, PREG_OFFSET_CAPTURE, $i)) { //We got a match ... - $matches[$dk] = array( - 'next_match' => $matches_rx[1][1], - 'dk' => $dk, + if(isset($matches_rx['start']) && isset($matches_rx['end'])) + { + $matches[$dk] = array( + 'next_match' => $matches_rx['start'][1], + 'dk' => $dk, - 'close_strlen' => strlen($matches_rx[2][0]), - 'close_pos' => $matches_rx[2][1], - ); + 'close_strlen' => strlen($matches_rx['end'][0]), + 'close_pos' => $matches_rx['end'][1], + ); + } else { + $matches[$dk] = array( + 'next_match' => $matches_rx[1][1], + 'dk' => $dk, + + 'close_strlen' => strlen($matches_rx[2][0]), + 'close_pos' => $matches_rx[2][1], + ); + } } else { // no match for this delimiter ever unset($delim_copy[$dk]); @@ -2097,6 +2268,7 @@ class GeSHi { } } } + // non-highlightable text $parts[$k] = array( 1 => substr($code, $i, $next_match_pos - $i) @@ -2225,7 +2397,7 @@ class GeSHi { foreach ($this->language_data['QUOTEMARKS'] as $quotemark) { if (!isset($is_string_starter[$quotemark[0]])) { $is_string_starter[$quotemark[0]] = (string)$quotemark; - } else if (is_string($is_string_starter[$quotemark[0]])) { + } elseif (is_string($is_string_starter[$quotemark[0]])) { $is_string_starter[$quotemark[0]] = array( $is_string_starter[$quotemark[0]], $quotemark); @@ -2297,6 +2469,43 @@ class GeSHi { $char = $part[$i]; $char_len = 1; + // update regexp comment cache if needed + if (isset($this->language_data['COMMENT_REGEXP']) && $next_comment_regexp_pos < $i) { + $next_comment_regexp_pos = $length; + foreach ($this->language_data['COMMENT_REGEXP'] as $comment_key => $regexp) { + $match_i = false; + if (isset($comment_regexp_cache_per_key[$comment_key]) && + ($comment_regexp_cache_per_key[$comment_key]['pos'] >= $i || + $comment_regexp_cache_per_key[$comment_key]['pos'] === false)) { + // we have already matched something + if ($comment_regexp_cache_per_key[$comment_key]['pos'] === false) { + // this comment is never matched + continue; + } + $match_i = $comment_regexp_cache_per_key[$comment_key]['pos']; + } elseif (preg_match($regexp, $part, $match, PREG_OFFSET_CAPTURE, $i)) { + $match_i = $match[0][1]; + + $comment_regexp_cache_per_key[$comment_key] = array( + 'key' => $comment_key, + 'length' => strlen($match[0][0]), + 'pos' => $match_i + ); + } else { + $comment_regexp_cache_per_key[$comment_key]['pos'] = false; + continue; + } + + if ($match_i !== false && $match_i < $next_comment_regexp_pos) { + $next_comment_regexp_pos = $match_i; + $next_comment_regexp_key = $comment_key; + if ($match_i === $i) { + break; + } + } + } + } + $string_started = false; if (isset($is_string_starter[$char])) { @@ -2326,7 +2535,7 @@ class GeSHi { $char_len = strlen($char); } - if ($string_started) { + if ($string_started && ($i != $next_comment_regexp_pos)) { // Hand out the correct style information for this string $string_key = array_search($char, $this->language_data['QUOTEMARKS']); if (!isset($this->language_data['STYLES']['STRINGS'][$string_key]) || @@ -2375,15 +2584,8 @@ class GeSHi { continue; } $match_i = $escape_regexp_cache_per_key[$escape_key]['pos']; - } else if ( - //This is to allow use of the offset parameter in preg_match and stay as compatible with older PHP versions as possible - (GESHI_PHP_PRE_433 && preg_match($regexp, substr($part, $start), $match, PREG_OFFSET_CAPTURE)) || - (!GESHI_PHP_PRE_433 && preg_match($regexp, $part, $match, PREG_OFFSET_CAPTURE, $start)) - ) { + } elseif (preg_match($regexp, $part, $match, PREG_OFFSET_CAPTURE, $start)) { $match_i = $match[0][1]; - if (GESHI_PHP_PRE_433) { - $match_i += $start; - } $escape_regexp_cache_per_key[$escape_key] = array( 'key' => $escape_key, @@ -2445,13 +2647,13 @@ class GeSHi { // don't put a newline around newlines $string .= "\n"; $start = $es_pos + 2; - } else if (ord($es_char) >= 128) { + } elseif (ord($es_char) >= 128) { //This is an non-ASCII char (UTF8 or single byte) //This code tries to work around SF#2037598 ... if(function_exists('mb_substr')) { $es_char_m = mb_substr(substr($part, $es_pos+1, 16), 0, 1, $this->encoding); $string .= $es_char_m . ''; - } else if (!GESHI_PHP_PRE_433 && 'utf-8' == $this->encoding) { + } elseif ('utf-8' == $this->encoding) { if(preg_match("/[\xC2-\xDF][\x80-\xBF]". "|\xE0[\xA0-\xBF][\x80-\xBF]". "|[\xE1-\xEC\xEE\xEF][\x80-\xBF]{2}". @@ -2473,7 +2675,7 @@ class GeSHi { $string .= $this->hsc($es_char) . ''; $start = $es_pos + 2; } - } else if ($next_escape_regexp_pos < $length && + } elseif ($next_escape_regexp_pos < $length && $next_escape_regexp_pos < $close_pos) { $es_pos = $next_escape_regexp_pos; //Add the stuff not in the string yet ... @@ -2517,8 +2719,8 @@ class GeSHi { $string = ''; $i = $start - 1; continue; - } else if ($this->lexic_permissions['STRINGS'] && $hq && $hq[0] == $char && - substr($part, $i, $hq_strlen) == $hq) { + } elseif ($this->lexic_permissions['STRINGS'] && $hq && $hq[0] == $char && + substr($part, $i, $hq_strlen) == $hq && ($i != $next_comment_regexp_pos)) { // The start of a hard quoted string if (!$this->use_classes) { $string_attributes = ' style="' . $this->language_data['STYLES']['STRINGS']['HARD'] . '"'; @@ -2538,14 +2740,15 @@ class GeSHi { $start = $i + $hq_strlen; while ($close_pos = strpos($part, $this->language_data['HARDQUOTE'][1], $start)) { $start = $close_pos + 1; - if ($this->lexic_permissions['ESCAPE_CHAR'] && $part[$close_pos - 1] == $this->language_data['ESCAPE_CHAR']) { + if ($this->lexic_permissions['ESCAPE_CHAR'] && $part[$close_pos - 1] == $this->language_data['HARDCHAR'] && + (($i + $hq_strlen) != ($close_pos))) { //Support empty string for HQ escapes if Starter = Escape // make sure this quote is not escaped foreach ($this->language_data['HARDESCAPE'] as $hardescape) { if (substr($part, $close_pos - 1, strlen($hardescape)) == $hardescape) { // check wether this quote is escaped or if it is something like '\\' $escape_char_pos = $close_pos - 1; while ($escape_char_pos > 0 - && $part[$escape_char_pos - 1] == $this->language_data['ESCAPE_CHAR']) { + && $part[$escape_char_pos - 1] == $this->language_data['HARDCHAR']) { --$escape_char_pos; } if (($close_pos - $escape_char_pos) & 1) { @@ -2625,49 +2828,6 @@ class GeSHi { $string = ''; continue; } else { - // update regexp comment cache if needed - if (isset($this->language_data['COMMENT_REGEXP']) && $next_comment_regexp_pos < $i) { - $next_comment_regexp_pos = $length; - foreach ($this->language_data['COMMENT_REGEXP'] as $comment_key => $regexp) { - $match_i = false; - if (isset($comment_regexp_cache_per_key[$comment_key]) && - ($comment_regexp_cache_per_key[$comment_key]['pos'] >= $i || - $comment_regexp_cache_per_key[$comment_key]['pos'] === false)) { - // we have already matched something - if ($comment_regexp_cache_per_key[$comment_key]['pos'] === false) { - // this comment is never matched - continue; - } - $match_i = $comment_regexp_cache_per_key[$comment_key]['pos']; - } else if ( - //This is to allow use of the offset parameter in preg_match and stay as compatible with older PHP versions as possible - (GESHI_PHP_PRE_433 && preg_match($regexp, substr($part, $i), $match, PREG_OFFSET_CAPTURE)) || - (!GESHI_PHP_PRE_433 && preg_match($regexp, $part, $match, PREG_OFFSET_CAPTURE, $i)) - ) { - $match_i = $match[0][1]; - if (GESHI_PHP_PRE_433) { - $match_i += $i; - } - - $comment_regexp_cache_per_key[$comment_key] = array( - 'key' => $comment_key, - 'length' => strlen($match[0][0]), - 'pos' => $match_i - ); - } else { - $comment_regexp_cache_per_key[$comment_key]['pos'] = false; - continue; - } - - if ($match_i !== false && $match_i < $next_comment_regexp_pos) { - $next_comment_regexp_pos = $match_i; - $next_comment_regexp_key = $comment_key; - if ($match_i === $i) { - break; - } - } - } - } //Have a look for regexp comments if ($i == $next_comment_regexp_pos) { $COMMENT_MATCHED = true; @@ -2717,7 +2877,7 @@ class GeSHi { continue; } $match_i = $comment_multi_cache_per_key[$open]; - } else if (($match_i = stripos($part, $open, $i)) !== false) { + } elseif (($match_i = stripos($part, $open, $i)) !== false) { $comment_multi_cache_per_key[$open] = $match_i; } else { $comment_multi_cache_per_key[$open] = false; @@ -2814,7 +2974,7 @@ class GeSHi { continue; } $match_i = $comment_single_cache_per_key[$comment_key]; - } else if ( + } elseif ( // case sensitive comments ($this->language_data['CASE_SENSITIVE'][GESHI_COMMENTS] && ($match_i = stripos($part, $comment_mark, $i)) !== false) || @@ -2934,11 +3094,10 @@ class GeSHi { * Swaps out spaces and tabs for HTML indentation. Not needed if * the code is in a pre block... * - * @param string The source to indent (reference!) + * @param string $result The source to indent (reference!) * @since 1.0.0 - * @access private */ - function indent(&$result) { + protected function indent(&$result) { /// Replace tabs with the correct number of spaces if (false !== strpos($result, "\t")) { $lines = explode("\n", $result); @@ -2971,10 +3130,10 @@ class GeSHi { $IN_TAG = false; } $lines[$key] .= $char; - } else if ('<' == $char) { + } elseif ('<' == $char) { $IN_TAG = true; $lines[$key] .= '<'; - } else if ('&' == $char) { + } elseif ('&' == $char) { $substr = substr($line, $i + 3, 5); $posi = strpos($substr, ';'); if (false === $posi) { @@ -2983,7 +3142,7 @@ class GeSHi { $pos -= $posi+2; } $lines[$key] .= $char; - } else if ("\t" == $char) { + } elseif ("\t" == $char) { $str = ''; // OPTIMISE - move $strs out. Make an array: // $tabs = array( @@ -3004,7 +3163,7 @@ class GeSHi { $lines[$key] .= substr($line, $i + 1); break; } - } else if (0 == $pos && ' ' == $char) { + } elseif (0 == $pos && ' ' == $char) { $lines[$key] .= ' '; ++$pos; } else { @@ -3021,7 +3180,7 @@ class GeSHi { $result = preg_replace('/^ /m', ' ', $result); $result = str_replace(' ', '  ', $result); - if ($this->line_numbers == GESHI_NO_LINE_NUMBERS) { + if ($this->line_numbers == GESHI_NO_LINE_NUMBERS && $this->header_type != GESHI_HEADER_PRE_TABLE) { if ($this->line_ending === null) { $result = nl2br($result); } else { @@ -3033,12 +3192,11 @@ class GeSHi { /** * Changes the case of a keyword for those languages where a change is asked for * - * @param string The keyword to change the case of + * @param string $instr The keyword to change the case of * @return string The keyword with its case changed * @since 1.0.0 - * @access private */ - function change_case($instr) { + protected function change_case($instr) { switch ($this->language_data['CASE_KEYWORDS']) { case GESHI_CAPS_UPPER: return strtoupper($instr); @@ -3052,16 +3210,16 @@ class GeSHi { /** * Handles replacements of keywords to include markup and links if requested * - * @param string The keyword to add the Markup to - * @return The HTML for the match found + * @param string $match The keyword to add the Markup to + * @return string The HTML for the match found * @since 1.0.8 - * @access private * * @todo Get rid of ender in keyword links */ - function handle_keyword_replace($match) { + protected function handle_keyword_replace($match) { $k = $this->_kw_replace_group; $keyword = $match[0]; + $keyword_match = $match[1]; $before = ''; $after = ''; @@ -3079,19 +3237,28 @@ class GeSHi { if (!$this->language_data['CASE_SENSITIVE'][$k] && strpos($this->language_data['URLS'][$k], '{FNAME}') !== false) { foreach ($this->language_data['KEYWORDS'][$k] as $word) { - if (strcasecmp($word, $keyword) == 0) { + if (strcasecmp($word, $keyword_match) == 0) { break; } } } else { - $word = $keyword; + $word = $keyword_match; } $before = '<|UR1|"' . str_replace( - array('{FNAME}', '{FNAMEL}', '{FNAMEU}', '.'), - array($this->hsc($word), $this->hsc(strtolower($word)), - $this->hsc(strtoupper($word)), ''), + array( + '{FNAME}', + '{FNAMEL}', + '{FNAMEU}', + '{FNAMEUF}', + '.'), + array( + str_replace('+', '%20', urlencode($this->hsc($word))), + str_replace('+', '%20', urlencode($this->hsc(strtolower($word)))), + str_replace('+', '%20', urlencode($this->hsc(strtoupper($word)))), + str_replace('+', '%20', urlencode($this->hsc(ucfirst($word)))), + ''), $this->language_data['URLS'][$k] ) . '">'; $after = ''; @@ -3106,12 +3273,11 @@ class GeSHi { * * @note this is a callback, don't use it directly * - * @param array the matches array - * @return The highlighted string + * @param array $matches the matches array + * @return string The highlighted string * @since 1.0.8 - * @access private */ - function handle_regexps_callback($matches) { + protected function handle_regexps_callback($matches) { // before: "' style=\"' . call_user_func(\"$func\", '\\1') . '\"\\1|>'", return ' style="' . call_user_func($this->language_data['STYLES']['REGEXPS'][$this->_rx_key], $matches[1]) . '"'. $matches[1] . '|>'; } @@ -3121,12 +3287,11 @@ class GeSHi { * * @note this is a callback, don't use it directly * - * @param array the matches array + * @param array $matches the matches array * @return string * @since 1.0.8 - * @access private */ - function handle_multiline_regexps($matches) { + protected function handle_multiline_regexps($matches) { $before = $this->_hmr_before; $after = $this->_hmr_after; if ($this->_hmr_replace) { @@ -3154,14 +3319,81 @@ class GeSHi { * Takes a string that has no strings or comments in it, and highlights * stuff like keywords, numbers and methods. * - * @param string The string to parse for keyword, numbers etc. + * @param string $stuff_to_parse The string to parse for keyword, numbers etc. * @since 1.0.0 - * @access private * @todo BUGGY! Why? Why not build string and return? + * @return string */ - function parse_non_string_part($stuff_to_parse) { + protected function parse_non_string_part($stuff_to_parse) { $stuff_to_parse = ' ' . $this->hsc($stuff_to_parse); + // Highlight keywords + $disallowed_before = "(?lexic_permissions['STRINGS']) { + $quotemarks = preg_quote(implode($this->language_data['QUOTEMARKS']), '/'); + $disallowed_before .= $quotemarks; + $disallowed_after .= $quotemarks; + } + $disallowed_before .= "])"; + $disallowed_after .= "])"; + + $parser_control_pergroup = false; + if (isset($this->language_data['PARSER_CONTROL'])) { + if (isset($this->language_data['PARSER_CONTROL']['KEYWORDS'])) { + $x = 0; // check wether per-keyword-group parser_control is enabled + if (isset($this->language_data['PARSER_CONTROL']['KEYWORDS']['DISALLOWED_BEFORE'])) { + $disallowed_before = $this->language_data['PARSER_CONTROL']['KEYWORDS']['DISALLOWED_BEFORE']; + ++$x; + } + if (isset($this->language_data['PARSER_CONTROL']['KEYWORDS']['DISALLOWED_AFTER'])) { + $disallowed_after = $this->language_data['PARSER_CONTROL']['KEYWORDS']['DISALLOWED_AFTER']; + ++$x; + } + $parser_control_pergroup = (count($this->language_data['PARSER_CONTROL']['KEYWORDS']) - $x) > 0; + } + } + + foreach (array_keys($this->language_data['KEYWORDS']) as $k) { + if (!isset($this->lexic_permissions['KEYWORDS'][$k]) || + $this->lexic_permissions['KEYWORDS'][$k]) { + + $case_sensitive = $this->language_data['CASE_SENSITIVE'][$k]; + $modifiers = $case_sensitive ? '' : 'i'; + + // NEW in 1.0.8 - per-keyword-group parser control + $disallowed_before_local = $disallowed_before; + $disallowed_after_local = $disallowed_after; + if ($parser_control_pergroup && isset($this->language_data['PARSER_CONTROL']['KEYWORDS'][$k])) { + if (isset($this->language_data['PARSER_CONTROL']['KEYWORDS'][$k]['DISALLOWED_BEFORE'])) { + $disallowed_before_local = + $this->language_data['PARSER_CONTROL']['KEYWORDS'][$k]['DISALLOWED_BEFORE']; + } + + if (isset($this->language_data['PARSER_CONTROL']['KEYWORDS'][$k]['DISALLOWED_AFTER'])) { + $disallowed_after_local = + $this->language_data['PARSER_CONTROL']['KEYWORDS'][$k]['DISALLOWED_AFTER']; + } + } + + $this->_kw_replace_group = $k; + + //NEW in 1.0.8, the cached regexp list + // since we don't want PHP / PCRE to crash due to too large patterns we split them into smaller chunks + for ($set = 0, $set_length = count($this->language_data['CACHED_KEYWORD_LISTS'][$k]); $set < $set_length; ++$set) { + $keywordset =& $this->language_data['CACHED_KEYWORD_LISTS'][$k][$set]; + // Might make a more unique string for putting the number in soon + // 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})(?!\(?:htm|php|aspx?))$disallowed_after_local/$modifiers", + array($this, 'handle_keyword_replace'), + $stuff_to_parse + ); + } + } + } + // Regular expressions foreach ($this->language_data['REGEXPS'] as $key => $regexp) { if ($this->lexic_permissions['REGEXPS'][$key]) { @@ -3199,9 +3431,10 @@ class GeSHi { } } - // Highlight numbers. As of 1.0.8 we support diffent types of numbers + // 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 ... @@ -3211,81 +3444,6 @@ class GeSHi { } } - // Highlight keywords - $disallowed_before = "(?|^&"; - $disallowed_after = "(?![a-zA-Z0-9_\|%\\-&;"; - if ($this->lexic_permissions['STRINGS']) { - $quotemarks = preg_quote(implode($this->language_data['QUOTEMARKS']), '/'); - $disallowed_before .= $quotemarks; - $disallowed_after .= $quotemarks; - } - $disallowed_before .= "])"; - $disallowed_after .= "])"; - - $parser_control_pergroup = false; - if (isset($this->language_data['PARSER_CONTROL'])) { - if (isset($this->language_data['PARSER_CONTROL']['KEYWORDS'])) { - $x = 0; // check wether per-keyword-group parser_control is enabled - if (isset($this->language_data['PARSER_CONTROL']['KEYWORDS']['DISALLOWED_BEFORE'])) { - $disallowed_before = $this->language_data['PARSER_CONTROL']['KEYWORDS']['DISALLOWED_BEFORE']; - ++$x; - } - if (isset($this->language_data['PARSER_CONTROL']['KEYWORDS']['DISALLOWED_AFTER'])) { - $disallowed_after = $this->language_data['PARSER_CONTROL']['KEYWORDS']['DISALLOWED_AFTER']; - ++$x; - } - $parser_control_pergroup = (count($this->language_data['PARSER_CONTROL']['KEYWORDS']) - $x) > 0; - } - } - - // if this is changed, don't forget to change it below -// if (!empty($disallowed_before)) { -// $disallowed_before = "(?language_data['KEYWORDS']) as $k) { - if (!isset($this->lexic_permissions['KEYWORDS'][$k]) || - $this->lexic_permissions['KEYWORDS'][$k]) { - - $case_sensitive = $this->language_data['CASE_SENSITIVE'][$k]; - $modifiers = $case_sensitive ? '' : 'i'; - - // NEW in 1.0.8 - per-keyword-group parser control - $disallowed_before_local = $disallowed_before; - $disallowed_after_local = $disallowed_after; - if ($parser_control_pergroup && isset($this->language_data['PARSER_CONTROL']['KEYWORDS'][$k])) { - if (isset($this->language_data['PARSER_CONTROL']['KEYWORDS'][$k]['DISALLOWED_BEFORE'])) { - $disallowed_before_local = - $this->language_data['PARSER_CONTROL']['KEYWORDS'][$k]['DISALLOWED_BEFORE']; - } - - if (isset($this->language_data['PARSER_CONTROL']['KEYWORDS'][$k]['DISALLOWED_AFTER'])) { - $disallowed_after_local = - $this->language_data['PARSER_CONTROL']['KEYWORDS'][$k]['DISALLOWED_AFTER']; - } - } - - $this->_kw_replace_group = $k; - - //NEW in 1.0.8, the cached regexp list - // since we don't want PHP / PCRE to crash due to too large patterns we split them into smaller chunks - for ($set = 0, $set_length = count($this->language_data['CACHED_KEYWORD_LISTS'][$k]); $set < $set_length; ++$set) { - $keywordset =& $this->language_data['CACHED_KEYWORD_LISTS'][$k][$set]; - // Might make a more unique string for putting the number in soon - // 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})(?!\(?:htm|php))$disallowed_after_local/$modifiers", - array($this, 'handle_keyword_replace'), - $stuff_to_parse - ); - } - } - } - // // Now that's all done, replace /[number]/ with the correct styles // @@ -3303,19 +3461,19 @@ class GeSHi { if ($numbers_found) { // Put number styles in foreach($this->language_data['NUMBERS_RXCACHE'] as $id => $regexp) { -//Commented out for now, as this needs some review ... -// if ($numbers_permissions & $id) { - //Get the appropriate style ... - //Checking for unset styles is done by the style cache builder ... - if (!$this->use_classes) { - $attributes = ' style="' . $this->language_data['STYLES']['NUMBERS'][$id] . '"'; - } else { - $attributes = ' class="nu'.$id.'"'; - } + //Commented out for now, as this needs some review ... + // if ($numbers_permissions & $id) { + //Get the appropriate style ... + //Checking for unset styles is done by the style cache builder ... + if (!$this->use_classes) { + $attributes = ' style="' . $this->language_data['STYLES']['NUMBERS'][$id] . '"'; + } else { + $attributes = ' class="nu'.$id.'"'; + } - //Set in the correct styles ... - $stuff_to_parse = str_replace("/NUM!$id/", $attributes, $stuff_to_parse); -// } + //Set in the correct styles ... + $stuff_to_parse = str_replace("/NUM!$id/", $attributes, $stuff_to_parse); + // } } } @@ -3365,7 +3523,7 @@ class GeSHi { //FIX for symbol highlighting ... if ($this->lexic_permissions['SYMBOLS'] && !empty($this->language_data['SYMBOLS'])) { //Get all matches and throw away those witin a block that is already highlighted... (i.e. matched by a regexp) - $n_symbols = preg_match_all("/<\|(?:|[^>])+>(?:(?!\|>).*?)\|>|<\/a>|(?:" . $this->language_data['SYMBOL_SEARCH'] . ")+/", $stuff_to_parse, $pot_symbols, PREG_OFFSET_CAPTURE | PREG_SET_ORDER); + $n_symbols = preg_match_all("/<\|(?:|[^>])+>(?:(?!\|>).*?)\|>|<\/a>|(?:" . $this->language_data['SYMBOL_SEARCH'] . ")+(?![^<]+?>)/", $stuff_to_parse, $pot_symbols, PREG_OFFSET_CAPTURE | PREG_SET_ORDER); $global_offset = 0; for ($s_id = 0; $s_id < $n_symbols; ++$s_id) { $symbol_match = $pot_symbols[$s_id][0][0]; @@ -3384,7 +3542,6 @@ class GeSHi { $symbol_length = strlen($symbol_match); $symbol_offset = $pot_symbols[$s_id][0][1]; unset($pot_symbols[$s_id]); - $symbol_end = $symbol_length + $symbol_offset; $symbol_hl = ""; // if we have multiple styles, we have to handle them properly @@ -3483,12 +3640,11 @@ class GeSHi { /** * Sets the time taken to parse the code * - * @param microtime The time when parsing started - * @param microtime The time when parsing ended + * @param string $start_time The time when parsing started as returned by @see microtime() + * @param string $end_time The time when parsing ended as returned by @see microtime() * @since 1.0.2 - * @access private */ - function set_time($start_time, $end_time) { + protected function set_time($start_time, $end_time) { $start = explode(' ', $start_time); $end = explode(' ', $end_time); $this->time = $end[0] + $end[1] - $start[0] - $start[1]; @@ -3500,7 +3656,7 @@ class GeSHi { * @return double The time taken to parse the code * @since 1.0.2 */ - function get_time() { + public function get_time() { return $this->time; } @@ -3508,9 +3664,8 @@ class GeSHi { * Merges arrays recursively, overwriting values of the first array with values of later arrays * * @since 1.0.8 - * @access private */ - function merge_arrays() { + protected function merge_arrays() { $arrays = func_get_args(); $narrays = count($arrays); @@ -3546,12 +3701,11 @@ class GeSHi { /** * Gets language information and stores it for later use * - * @param string The filename of the language file you want to load + * @param string $file_name The filename of the language file you want to load * @since 1.0.0 - * @access private * @todo Needs to load keys for lexic permissions for keywords, regexps etc */ - function load_language($file_name) { + protected function load_language($file_name) { if ($file_name == $this->loaded_language) { // this file is already loaded! return; @@ -3617,6 +3771,12 @@ class GeSHi { unset($this->language_data['PARSER_CONTROL']['ENABLE_FLAGS']); } + //Fix: Problem where hardescapes weren't handled if no ESCAPE_CHAR was given + //You need to set one for HARDESCAPES only in this case. + if(!isset($this->language_data['HARDCHAR'])) { + $this->language_data['HARDCHAR'] = $this->language_data['ESCAPE_CHAR']; + } + //NEW in 1.0.8: Allow styles to be loaded from a separate file to override defaults $style_filename = substr($file_name, 0, -4) . '.style.php'; if (is_readable($style_filename)) { @@ -3640,11 +3800,10 @@ class GeSHi { * Takes the parsed code and various options, and creates the HTML * surrounding it to make it look nice. * - * @param string The code already parsed (reference!) + * @param string $parsed_code The code already parsed (reference!) * @since 1.0.0 - * @access private */ - function finalise(&$parsed_code) { + protected function finalise(&$parsed_code) { // Remove end parts of important declarations // This is BUGGY!! My fault for bad code: fix coming in 1.2 // @todo Remove this crap @@ -3680,9 +3839,6 @@ class GeSHi { // the 
 will line-break them (and the 
  • s already do this for us)
             $ls = ($this->header_type != GESHI_HEADER_PRE && $this->header_type != GESHI_HEADER_PRE_VALID) ? "\n" : '';
 
-            // Set vars to defaults for following loop
-            $i = 0;
-
             // Foreach line...
             for ($i = 0, $n = count($code); $i < $n;) {
                 //Reset the attributes for a new line ...
@@ -3816,7 +3972,7 @@ class GeSHi {
                         $parsed_code .= $this->line_numbers_start + $i;
                         if ($close) {
                             $parsed_code .= str_repeat('', $close);
-                        } else if ($i != $n) {
+                        } elseif ($i != $n) {
                             $parsed_code .= "\n";
                         }
                     }
@@ -3890,24 +4046,23 @@ class GeSHi {
      *
      * @return string The header for the code block
      * @since  1.0.0
-     * @access private
      */
-    function header() {
+    protected function header() {
         // Get attributes needed
         /**
          * @todo   Document behaviour change - class is outputted regardless of whether
          *         we're using classes or not. Same with style
          */
-        $attributes = ' class="' . $this->language;
+        $attributes = ' class="' . $this->_genCSSName($this->language);
         if ($this->overall_class != '') {
-            $attributes .= " ".$this->overall_class;
+            $attributes .= " ".$this->_genCSSName($this->overall_class);
         }
         $attributes .= '"';
 
         if ($this->overall_id != '') {
             $attributes .= " id=\"{$this->overall_id}\"";
         }
-        if ($this->overall_style != '') {
+        if ($this->overall_style != '' && !$this->use_classes) {
             $attributes .= ' style="' . $this->overall_style . '"';
         }
 
@@ -3948,10 +4103,10 @@ class GeSHi {
         if ($this->line_numbers != GESHI_NO_LINE_NUMBERS) {
             if ($this->header_type == GESHI_HEADER_PRE) {
                 return "$header";
-            } else if ($this->header_type == GESHI_HEADER_DIV ||
+            } elseif ($this->header_type == GESHI_HEADER_DIV ||
                 $this->header_type == GESHI_HEADER_PRE_VALID) {
                 return "$header";
-            } else if ($this->header_type == GESHI_HEADER_PRE_TABLE) {
+            } elseif ($this->header_type == GESHI_HEADER_PRE_TABLE) {
                 return "$header";
             }
         } else {
@@ -3970,9 +4125,8 @@ class GeSHi {
      *
      * @return string The footer for the code block
      * @since  1.0.0
-     * @access private
      */
-    function footer() {
+    protected function footer() {
         $footer = $this->footer_content;
         if ($footer) {
             if ($this->header_type == GESHI_HEADER_PRE) {
@@ -3985,7 +4139,7 @@ class GeSHi {
             } else {
                 $attr = " style=\"{$this->footer_content_style}\"";
             }
-            if ($this->header_type == GESHI_HEADER_PRE_TABLE && $this->linenumbers != GESHI_NO_LINE_NUMBERS) {
+            if ($this->header_type == GESHI_HEADER_PRE_TABLE && $this->line_numbers != GESHI_NO_LINE_NUMBERS) {
                 $footer = "$footer";
             } else {
                 $footer = "$footer";
@@ -4023,12 +4177,11 @@ class GeSHi {
      * Replaces certain keywords in the header and footer with
      * certain configuration values
      *
-     * @param  string The header or footer content to do replacement on
+     * @param  string $instr The header or footer content to do replacement on
      * @return string The header or footer with replaced keywords
      * @since  1.0.2
-     * @access private
      */
-    function replace_keywords($instr) {
+    protected function replace_keywords($instr) {
         $keywords = $replacements = array();
 
         $keywords[] = '