GeSHi Documentation

The error message you will get will look like this:

GeSHi Error: GeSHi could not find the language thisLangIsNotSupported (using path geshi/) (code 2)

The error outputted will be the last error GeSHi came across, just like how mysql_error() works.

3.8: Disabling styling of some Lexics

One disadvantage of GeSHi is that for large source files using complex languages, it can be quite slow with every option turned on. Although future releases will concentrate on the speed/resource side of highlighting, for now you can gain speed increases by disabling some of the highlighting options. This is done by using a series of set_*_highlighting methods:

• set_keyword_group_highlighting($group, $flag): Sets whether a particular $group of keywords is to be highlighted or not. Consult the necessary language file(s) to see what $group should be for each group (typically a positive integer). $flag is false if you want to disable highlighting of this group, and true if you want to re-enable higlighting of this group. If you disable a keyword group then even if the keyword group has a related URL one will not be generated for that keyword.
• set_comments_highlighting($group, $flag): Sets whether a particular $group of comments is to be highlighted or not. Consult the necessary language file(s) to see what $group should be for each group (typically a positive integer, or the string 'MULTI' for multiline comments. $flag is false if you want to disable highlighting of this group, and true if you want to re-enable highlighting of this group.
• set_regexps_highlighting($regexp, $flag): Sets whether a particular $regexp is to be highlighted or not. Consult the necessary language file(s) to see what $regexp should be for each regexp (typically a positive integer, or the string 'MULTI' for multiline comments. $flag is false if you want to disable highlighting of this group, and true if you want to re-enable highlighting of this group.
• The following methods:
  • set_escape_characters_highlighting($flag)
  • set_symbols_highlighting($flag)
  • set_strings_highlighting($flag)
  • set_numbers_highlighting($flag)
  • set_methods_highlighting($flag)
  Work on their respective lexics (e.g. set_methods_highlighting will disable/enable highlighting of methods). For each method, if $flag is false then the related lexics will not be highlighted at all (this means no HTML will surround the lexic like usual, saving on time and bandwidth.

3.9: Setting the Tab Width

If you're using the <pre> header, tabs are handled automatically by your browser, and in general you can count on good results. However, if you're using the <div> header, you may want to specify a tab width explicitly.

Note that tabs created in this fashion won't be like normal tabs - there won't be "tab-stops" as such, instead tabs will be replaced with the specified number of spaces.

To change the tab width, you call the set_tab_width() method:

Where $width is the width in spaces that you'd like tabs to be.

3.10: Using Strict Mode

Some languages like to get tricky, and jump in and out of the file that they're in. For example, the vast majority of you reading this will have used a PHP file. And you know that PHP code is only executed if it's within delimiters like <?php and ?> (there are others of course...). So what happens if you do the following in a php file?

Well normally using GeSHi with PHP, or using a bad highlighter, you'll end up with this:

What a mess! Especially if you're being slack about where you're putting your quotes, you could end up with the rest of your file as bright red. Fortunately, you can tell GeSHi to be "strict" about just when it highlights and when it does not, using the enable_strict_mode method:

Where $mode is true or not specified to enable strict mode, or false to disable strict mode if you've already turned it and don't want it now.

Here's the result: much better!

3.11: Adding/Removing Keywords

Lets say that you're working on a large project, with many files, many classes and many functions. Perhaps also you have the source code on the web and highlighted by GeSHi, perhaps as a front end to CVS, as a learning tool, something to refer to, whatever. Well, why not highlight the names of the functions and classes *your* project uses, as well as the standard functions and classes? Or perhaps you're not interested in highlighting certain functions, and would like to remove them? Or maybe you don't mind if an entire function group goes west in the interest of speed? GeSHi can handle all of this!

3.11.1: Adding a Keyword

If you want to add a keyword to an existing keyword group, you use the add_keyword method:

Where $key is the index of the group of keywords you want to add this keyword to, and $word is the word to add.

This implies knowledge of the language file to know the correct index.

3.11.2: Removing a Keyword

Perhaps you want to remove a keyword from an existing group. Maybe you don't use it and want to save yourself some time. Whatever the reason, you can remove it using the remove_keyword method:

Where $key is the index of the gropu of keywords that you want to remove this keyword from, and $word is the word to remove.

This implies knowledge of the language file to know the correct index - most of the time the keywords you'll want to remove will be in group 3, but this is not guaranteed and you should check the language file first.

This function is silent - if the keyword is not in the group you specified, nothing awful will happen ;)

3.11.3: Adding a Keyword Group

Lets say for your big project you have several main functions and classes that you'd like highlighted. Why not add them as their own group instead of having them highlighted the same way as other keywords? Then you can make them stand out, and people can instantly see which functions and classes are user defined or inbuilt. Furthermore, you could set the URL for this group to point at the API documentation of your project.

You add a keyword group by using the add_keyword_group method:

Where $key is the key that you want to use to refer to this group, $styles is the styles that you want to use to style this group, $case_sensitive is true or false depending on whether you want this group of keywords to be case sensitive or not and $words is an array of words (or a string) of which words to add to this group. For example:

Adds a keyword group referenced by index 10, of which all keywords in the group will be dark red, each keyword can be in any case and which contains the keywords "myfunc_1", "myfunc_2" and "myfunc_3".

After creating such a keyword group, you may call other GeSHi methods on it, just as you would for any other keyword group.

3.11.4: Removing a Keyword Group

Perhaps you *really* need speed? Why not just remove an entire keyword group? GeSHi won't have to loop through each keyword checking for its existance, saving much time. You removing a keyword group by using the remove_keyword_group method:

Where $key is the key of the group you wish to remove. This implies knowleged of the language file.

3.12: Headers and Footers for Your Code

So you want to add some special information to the highlighted source? GeSHi can do that too! You can specify headers and footers for your code, style them, and insert information from the highlighted source into your header or footer.

3.12.1: Keyword Substitution

In your header and footer, you can put special keywords that will be replaced with actual configuration values for this GeSHi object. The keywords you can use are:

• <TIME>: Is replaced by the time it took for the parse_code method - i.e., how long it took for your code to be highlighted. The time is returned to three decimal places.
• <LANGUAGE>: Is replaced by a nice, friendly version of the language name used to highlight this code.
• <VERSION>: The GeSHi version used to highlight the code.

3.12.2: Setting Header Content

The header for your code is a <div>, which is inside the containing block. Therefore, it is affected by the method set_overall_style, and should contain the sort of HTML that belongs in a <div>. You may use any HTML you like, and format it as an HTML document. You should use valid HTML - convert to entities any quotemarks or angle brackets you want displayed. You set the header content using the method set_header_content:

Where $content is the HTML you want to use for the header.

3.12.3: Setting Footer Content

The footer for your code is a <div>, which is inside the containing block. Therefore, it is affected by the method set_overall_style, and should contain the sort of HTML that belongs in a <div>. You may use any HTML you like, and format it as an HTML document. You should use valid HTML - convert to entities any quotemarks or angle brackets you want displayed. You set the footer content using the method set_footer_content:

Where $content is the HTML you want to use for the footer.

3.12.4: Styling Header Content

You can apply styles to the header content you have set with the set_header_content_style:

Where $styles is the stylesheet declarations you want to use to style the header content.

3.12.5: Styling Footer Content

You can apply styles to the footer content you have set with the set_footer_content_style:

Where $styles is the stylesheet declarations you want to use to style the footer content.

3.13: Keyword URLs

As of version 1.0.2, GeSHi allows you to specify a URL for keyword groups. This URL is used by GeSHi to convert the keywords in that group into URLs to appropriate documentation. And using add_keyword_group you can add functions and classes from your own projects and use the URL functionality to provide a link to your own API documentation.

3.13.1: Setting a URL for a Keyword Group

To set the URL to be used for a keyword group, you use the set_url_for_keyword_group method:

Where $group is the keyword group you want to assign the URL for, and $url is the URL for this group of keywords.

You may be wondering how to make each keyword in the group point to the correct URL. You do this by putting {FNAME} in the URL at the correct place. For example, PHP makes it easy by linking www.php.net/function-name to the documentation for that function, so the URL used is http://www.php.net/{FNAME}.

Of course, when you get to a language like Java, that puts its class documentation in related folders, it gets a little trickier to work out an appropriate URL (see the Java language file!). I hope to provide some kind of redirection service at the GeSHi website in the future.

3.13.2: Disabling a URL for a Keyword Group

It's easy to disable a URL for a keyword group: Simply use the method set_url_for_keyword_group to pass an empty string as the URL:

3.13.3: Styling Links

You can also style the function links. You can style their default status, hovered, active and visited status. All of this is controlled by one method, set_link_styles:

Where $mode is one of four values:

• GESHI_LINK: The default style of the links.
• GESHI_HOVER: The style of the links when they have focus (the mouse is hovering over them).
• GESHI_ACTIVE: The style of the links when they are being clicked.
• GESHI_VISITED: The style of links that the user has already visited.

And $styles is the stylesheet declarations to apply to the links.

3.13.4: Setting the Link Target

Perhaps you want to set the target of link attributes, so the manual pages open in a new window? Use the set_link_target method:

Where $target is any valid (X)HTML target value - _blank or _top for example.

3.14: Using Contextual Importance

3.15: Highlighting Special Lines "Extra"

An alternative (and more stable) method of highlighting code that is important is to use extra highlighting by line. Although you may not know what line numbers contain the important lines, if you do this method is a much more flexible way of making important lines stand out.

3.15.1: Specifying the Lines to Highlight Extra

To specify which lines to highlight extra, you past an array containing the line numbers to highlight_lines_extra:

The array could be in the form array(2, 3, 4, 7, 12, 344, 4242), made from a DB query, generated from looking through the source for certain important things and working out what line those things are... however you get the line numbers, the array should simply be an array of integers.

Here's an example, using the same source as before:

Which produces:

What's more, as you can see the code on a highlighted line is still actually highlighted itself.

3.15.2: Styles for the Highlighted Lines

Again as with contextual importance, you're not chained to the yellow theme that is the default. You can use the set_highlight_lines_extra_style method:

Where $styles is the stylesheet declarations that you want to apply to highlighted lines.

3.16: Adding IDs to Each Line

Perhaps you're a javascript junkie? GeSHi provides a way to give each line an ID so you can access that line with javascript, or perhaps just by plain CSS (though if you want to access lines by CSS you should use the method in the previous section). To enable IDs you call the enable_ids method:

Where $flag is true or not present to enable IDs, and false to disable them again if you need.

The ID generated is in the form {overall-css-id}-{line-number}. So for example, if you set the overall CSS id to be "mycode", then the IDs for each line would by "mycode-1", "mycode-2" etc. If there is no CSS ID set, then the CSS class is used, but this isn't a good idea if you're going to have more than one block of code on any one page.

3.17: Getting the Time of Styling

Once you've called parse_code, you can get the time it took to run the highlighting by calling the get_time method:

4: Language Files

So now you know what features GeSHi offers, and perhaps you've even meddled with the source. Or perhaps you'd like a language file for language X but it doesn't seem to be supported? Rubbish! GeSHi will highlight anything, what do you think I coded this for? ^_^ You'll just have to learn how to make a language file yourself. And I promise it's not too hard - and if you're here you're in the right place!

4.1: An Example Language File

Let's begin by looking at an example language file - the language file for the first language ever supported, PHP.

<?php
/*************************************************************************************
 * php.php
 * --------
 * Author: Nigel McNie (nigel@geshi.org)
 * Copyright: (c) 2004 Nigel McNie (http://qbnz.com/highlighter/)
 * Release Version: 1.0.2
 * CVS Revision Version: $Revision$
 * Date Started: 2004/06/20
 * Last Modified: $Date$
 *
 * PHP language file for GeSHi.
 *
 * CHANGES
 * -------
 * 2004/XX/XX (1.0.2)
 *  -  Added URL support
 *  -  Added extra constants
 * 2004/08/05 (1.0.1)
 *  -  Added support for symbols
 * 2004/07/14 (1.0.0)
 *  -  First Release
 *
 * TODO (updated 2004/07/14)
 * -------------------------
 * * Make sure the last few function I may have missed
 *   (like eval()) are included for highlighting
 * * Split to several files - php4, php5 etc
 *
 *************************************************************************************
 *
 *     This file is part of GeSHi.
 *
 *   GeSHi is free software; you can redistribute it and/or modify
 *   it under the terms of the GNU General Public License as published by
 *   the Free Software Foundation; either version 2 of the License, or
 *   (at your option) any later version.
 *
 *   GeSHi is distributed in the hope that it will be useful,
 *   but WITHOUT ANY WARRANTY; without even the implied warranty of
 *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *   GNU General Public License for more details.
 *
 *   You should have received a copy of the GNU General Public License
 *   along with GeSHi; if not, write to the Free Software
 *   Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 *
 ************************************************************************************/
 
$language_data = array (
	'LANG_NAME' => 'PHP',
	'COMMENT_SINGLE' => array(1 => '//', 2 => '#'),
	'COMMENT_MULTI' => array('/*' => '*/'),
	'CASE_KEYWORDS' => GESHI_CAPS_NO_CHANGE,
	'QUOTEMARKS' => array("'", '"'),
	'ESCAPE_CHAR' => '\\',
	'KEYWORDS' => array(
		1 => array(
			'include', 'require', 'include_once', 'require_once',
			'for', 'foreach', 'as', 'if', 'elseif', 'else', 'while', 'do', 'endwhile', 'endif', 'switch', 'case', 'endswitch',
			'return', 'break'
			),
		2 => array(
			'null', '__LINE__', '__FILE__',
			'false', '&lt;?php', '?&gt;', '&lt;?',
			'&lt;script language=', '&lt;/script&gt;',
			'true', 'var', 'default',
			'function', 'class', 'new',
			'__FUNCTION__', '__CLASS__', '__METHOD__', 'PHP_VERSION',
			'PHP_OS', 'DEFAULT_INCLUDE_PATH', 'PEAR_INSTALL_DIR', 'PEAR_EXTENSION_DIR',
			'PHP_EXTENSION_DIR', 'PHP_BINDIR', 'PHP_LIBDIR', 'PHP_DATADIR', 'PHP_SYSCONFDIR',
			'PHP_LOCALSTATEDIR', 'PHP_CONFIG_FILE_PATH', 'PHP_OUTPUT_HANDLER_START', 'PHP_OUTPUT_HANDLER_CONT',
			'PHP_OUTPUT_HANDLER_END', 'E_ERROR', 'E_WARNING', 'E_PARSE', 'E_NOTICE',
			'E_CORE_ERROR', 'E_CORE_WARNING', 'E_COMPILE_ERROR', 'E_COMPILE_WARNING', 'E_USER_ERROR',
			'E_USER_WARNING', 'E_USER_NOTICE', 'E_ALL'
			),
		3 => array(
			'zlib_get_coding_type','zend_version','zend_logo_guid','yp_order','yp_next',
			'yp_match','yp_master','yp_get_default_domain','yp_first','yp_errno','yp_err_string',
			'yp_cat','yp_all','xml_set_unparsed_entity_decl_handler','xml_set_start_namespace_decl_handler','xml_set_processing_instruction_handler','xml_set_object',
			'xml_set_notation_decl_handler','xml_set_external_entity_ref_handler','xml_set_end_namespace_decl_handler','xml_set_element_handler','xml_set_default_handler','xml_set_character_data_handler',
			'xml_parser_set_option','xml_parser_get_option','xml_parser_free','xml_parser_create_ns','xml_parser_create','xml_parse_into_struct',
			'xml_parse','xml_get_error_code','xml_get_current_line_number','xml_get_current_column_number','xml_get_current_byte_index','xml_error_string',
			'wordwrap','wddx_serialize_vars','wddx_serialize_value','wddx_packet_start','wddx_packet_end','wddx_deserialize',
			'wddx_add_vars','vsprintf','vprintf','virtual','version_compare','var_export',
			'var_dump','utf8_encode','utf8_decode','usort','usleep','user_error',
			'urlencode','urldecode','unserialize','unregister_tick_function','unpack','unlink',
			'unixtojd','uniqid','umask','uksort','ucwords','ucfirst',
			'uasort','trim','trigger_error','touch','token_name','token_get_all',
			'tmpfile','time','textdomain','tempnam','tanh','tan',
			'system','syslog','symlink','substr_replace','substr_count','substr',
			'strval','strtr','strtoupper','strtotime','strtolower','strtok',
			'strstr','strspn','strrpos','strrev','strrchr','strpos',
			'strncmp','strncasecmp','strnatcmp','strnatcasecmp','strlen','stristr',
			'stripslashes','stripcslashes','strip_tags','strftime','stream_wrapper_register','stream_set_write_buffer',
			'stream_set_timeout','stream_set_blocking','stream_select','stream_register_wrapper','stream_get_meta_data','stream_filter_prepend',
			'stream_filter_append','stream_context_set_params','stream_context_set_option','stream_context_get_options','stream_context_create','strcspn',
			'strcoll','strcmp','strchr','strcasecmp','str_word_count','str_shuffle',
			'str_rot13','str_replace','str_repeat','str_pad','stat','sscanf',
			'srand','sqrt','sql_regcase','sprintf','spliti','split',
			'soundex','sort','socket_writev','socket_write','socket_strerror','socket_shutdown',
			'socket_setopt','socket_set_timeout','socket_set_option','socket_set_nonblock','socket_set_blocking','socket_set_block',
			'socket_sendto','socket_sendmsg','socket_send','socket_select','socket_recvmsg','socket_recvfrom',
			'socket_recv','socket_readv','socket_read','socket_listen','socket_last_error','socket_iovec_set',
			'socket_iovec_free','socket_iovec_fetch','socket_iovec_delete','socket_iovec_alloc','socket_iovec_add','socket_getsockname',
			'socket_getpeername','socket_getopt','socket_get_status','socket_get_option','socket_create_pair','socket_create_listen',
			'socket_create','socket_connect','socket_close','socket_clear_error','socket_bind','socket_accept',
			'sleep','sizeof','sinh','sin','similar_text','shuffle',
			'show_source','shmop_write','shmop_size','shmop_read','shmop_open','shmop_delete',
			'shmop_close','shm_remove_var','shm_remove','shm_put_var','shm_get_var','shm_detach',
			'shm_attach','shell_exec','sha1_file','sha1','settype','setlocale',
			'setcookie','set_time_limit','set_socket_blocking','set_magic_quotes_runtime','set_include_path','set_file_buffer',
			'set_error_handler','session_write_close','session_unset','session_unregister','session_start','session_set_save_handler',
			'session_set_cookie_params','session_save_path','session_register','session_regenerate_id','session_name','session_module_name',
			'session_is_registered','session_id','session_get_cookie_params','session_encode','session_destroy','session_decode',
			'session_cache_limiter','session_cache_expire','serialize','sem_remove','sem_release','sem_get',
			'sem_acquire','rtrim','rsort','round','rmdir','rewinddir',
			'rewind','restore_include_path','restore_error_handler','reset','rename','register_tick_function',
			'register_shutdown_function','realpath','readlink','readgzfile','readfile','readdir',
			'read_exif_data','rawurlencode','rawurldecode','range','rand','rad2deg',
			'quotemeta','quoted_printable_decode','putenv','proc_open','proc_close','printf',
			'print_r','prev','preg_split','preg_replace_callback','preg_replace','preg_quote',
			'preg_match_all','preg_match','preg_grep','pow','posix_uname','posix_ttyname',
			'posix_times','posix_strerror','posix_setuid','posix_setsid','posix_setpgid','posix_setgid',
			'posix_seteuid','posix_setegid','posix_mkfifo','posix_kill','posix_isatty','posix_getuid',
			'posix_getsid','posix_getrlimit','posix_getpwuid','posix_getpwnam','posix_getppid','posix_getpid',
			'posix_getpgrp','posix_getpgid','posix_getlogin','posix_getgroups','posix_getgrnam','posix_getgrgid',
			'posix_getgid','posix_geteuid','posix_getegid','posix_getcwd','posix_get_last_error','posix_errno',
			'posix_ctermid','pos','popen','pi','phpversion','phpinfo',
			'phpcredits','php_uname','php_sapi_name','php_logo_guid','php_ini_scanned_files','pg_update',
			'pg_untrace','pg_unescape_bytea','pg_tty','pg_trace','pg_setclientencoding','pg_set_client_encoding',
			'pg_send_query','pg_select','pg_result_status','pg_result_seek','pg_result_error','pg_result',
			'pg_query','pg_put_line','pg_port','pg_ping','pg_pconnect','pg_options',
			'pg_numrows','pg_numfields','pg_num_rows','pg_num_fields','pg_meta_data','pg_lowrite',
			'pg_lounlink','pg_loreadall','pg_loread','pg_loopen','pg_loimport','pg_loexport',
			'pg_locreate','pg_loclose','pg_lo_write','pg_lo_unlink','pg_lo_tell','pg_lo_seek',
			'pg_lo_read_all','pg_lo_read','pg_lo_open','pg_lo_import','pg_lo_export','pg_lo_create',
			'pg_lo_close','pg_last_oid','pg_last_notice','pg_last_error','pg_insert','pg_host',
			'pg_getlastoid','pg_get_result','pg_get_pid','pg_get_notify','pg_freeresult','pg_free_result',
			'pg_fieldtype','pg_fieldsize','pg_fieldprtlen','pg_fieldnum','pg_fieldname','pg_fieldisnull',
			'pg_field_type','pg_field_size','pg_field_prtlen','pg_field_num','pg_field_name','pg_field_is_null',
			'pg_fetch_row','pg_fetch_result','pg_fetch_object','pg_fetch_assoc','pg_fetch_array','pg_fetch_all',
			'pg_exec','pg_escape_string','pg_escape_bytea','pg_errormessage','pg_end_copy','pg_delete',
			'pg_dbname','pg_copy_to','pg_copy_from','pg_convert','pg_connection_status','pg_connection_reset',
			'pg_connection_busy','pg_connect','pg_cmdtuples','pg_close','pg_clientencoding','pg_client_encoding',
			'pg_cancel_query','pg_affected_rows','pfsockopen','pclose','pathinfo','passthru',
			'parse_url','parse_str','parse_ini_file','pack','overload','output_reset_rewrite_vars',
			'output_add_rewrite_var','ord','openssl_x509_read','openssl_x509_parse','openssl_x509_free','openssl_x509_export_to_file',
			'openssl_x509_export','openssl_x509_checkpurpose','openssl_x509_check_private_key','openssl_verify','openssl_sign','openssl_seal',
			'openssl_public_encrypt','openssl_public_decrypt','openssl_private_encrypt','openssl_private_decrypt','openssl_pkey_new','openssl_pkey_get_public',
			'openssl_pkey_get_private','openssl_pkey_free','openssl_pkey_export_to_file','openssl_pkey_export','openssl_pkcs7_verify','openssl_pkcs7_sign',
			'openssl_pkcs7_encrypt','openssl_pkcs7_decrypt','openssl_open','openssl_get_publickey','openssl_get_privatekey','openssl_free_key',
			'openssl_error_string','openssl_csr_sign','openssl_csr_new','openssl_csr_export_to_file','openssl_csr_export','openlog',
			'opendir','octdec','ob_start','ob_list_handlers','ob_implicit_flush','ob_iconv_handler',
			'ob_gzhandler','ob_get_status','ob_get_level','ob_get_length','ob_get_flush','ob_get_contents',
			'ob_get_clean','ob_flush','ob_end_flush','ob_end_clean','ob_clean','number_format',
			'nl_langinfo','nl2br','ngettext','next','natsort','natcasesort',
			'mysql_unbuffered_query','mysql_thread_id','mysql_tablename','mysql_table_name','mysql_stat','mysql_selectdb',
			'mysql_select_db','mysql_result','mysql_real_escape_string','mysql_query','mysql_ping','mysql_pconnect',
			'mysql_numrows','mysql_numfields','mysql_num_rows','mysql_num_fields','mysql_listtables','mysql_listfields',
			'mysql_listdbs','mysql_list_tables','mysql_list_processes','mysql_list_fields','mysql_list_dbs','mysql_insert_id',
			'mysql_info','mysql_get_server_info','mysql_get_proto_info','mysql_get_host_info','mysql_get_client_info','mysql_freeresult',
			'mysql_free_result','mysql_fieldtype','mysql_fieldtable','mysql_fieldname','mysql_fieldlen','mysql_fieldflags',
			'mysql_field_type','mysql_field_table','mysql_field_seek','mysql_field_name','mysql_field_len','mysql_field_flags',
			'mysql_fetch_row','mysql_fetch_object','mysql_fetch_lengths','mysql_fetch_field','mysql_fetch_assoc','mysql_fetch_array',
			'mysql_escape_string','mysql_error','mysql_errno','mysql_dropdb','mysql_drop_db','mysql_dbname',
			'mysql_db_query','mysql_db_name','mysql_data_seek','mysql_createdb','mysql_create_db','mysql_connect',
			'mysql_close','mysql_client_encoding','mysql_affected_rows','mysql','mt_srand','mt_rand',
			'mt_getrandmax','move_uploaded_file','money_format','mktime','mkdir','min',
			'microtime','method_exists','metaphone','memory_get_usage','md5_file','md5',
			'mbsubstr','mbstrrpos','mbstrpos','mbstrlen','mbstrcut','mbsplit',
			'mbregex_encoding','mberegi_replace','mberegi','mbereg_search_setpos','mbereg_search_regs','mbereg_search_pos',
			'mbereg_search_init','mbereg_search_getregs','mbereg_search_getpos','mbereg_search','mbereg_replace','mbereg_match',
			'mbereg','mb_substr_count','mb_substr','mb_substitute_character','mb_strwidth','mb_strtoupper',
			'mb_strtolower','mb_strrpos','mb_strpos','mb_strlen','mb_strimwidth','mb_strcut',
			'mb_split','mb_send_mail','mb_regex_set_options','mb_regex_encoding','mb_preferred_mime_name','mb_parse_str',
			'mb_output_handler','mb_language','mb_internal_encoding','mb_http_output','mb_http_input','mb_get_info',
			'mb_eregi_replace','mb_eregi','mb_ereg_search_setpos','mb_ereg_search_regs','mb_ereg_search_pos','mb_ereg_search_init',
			'mb_ereg_search_getregs','mb_ereg_search_getpos','mb_ereg_search','mb_ereg_replace','mb_ereg_match','mb_ereg',
			'mb_encode_numericentity','mb_encode_mimeheader','mb_detect_order','mb_detect_encoding','mb_decode_numericentity','mb_decode_mimeheader',
			'mb_convert_variables','mb_convert_kana','mb_convert_encoding','mb_convert_case','max','mail',
			'magic_quotes_runtime','ltrim','lstat','long2ip','log1p','log10',
			'log','localtime','localeconv','linkinfo','link','levenshtein',
			'lcg_value','ksort','krsort','key_exists','key','juliantojd',
			'join','jewishtojd','jdtounix','jdtojulian','jdtojewish','jdtogregorian',
			'jdtofrench','jdmonthname','jddayofweek','is_writeable','is_writable','is_uploaded_file',
			'is_subclass_of','is_string','is_scalar','is_resource','is_real','is_readable',
			'is_object','is_numeric','is_null','is_nan','is_long','is_link',
			'is_integer','is_int','is_infinite','is_float','is_finite','is_file',
			'is_executable','is_double','is_dir','is_callable','is_bool','is_array',
			'is_a','iptcparse','iptcembed','ip2long','intval','ini_set',
			'ini_restore','ini_get_all','ini_get','ini_alter','in_array','import_request_variables',
			'implode','image_type_to_mime_type','ignore_user_abort','iconv_set_encoding','iconv_get_encoding','iconv',
			'i18n_mime_header_encode','i18n_mime_header_decode','i18n_ja_jp_hantozen','i18n_internal_encoding','i18n_http_output','i18n_http_input',
			'i18n_discover_encoding','i18n_convert','hypot','htmlspecialchars','htmlentities','html_entity_decode',
			'highlight_string','highlight_file','hexdec','hebrevc','hebrev','headers_sent',
			'header','gzwrite','gzuncompress','gztell','gzseek','gzrewind',
			'gzread','gzputs','gzpassthru','gzopen','gzinflate','gzgetss',
			'gzgets','gzgetc','gzfile','gzeof','gzencode','gzdeflate',
			'gzcompress','gzclose','gregoriantojd','gmstrftime','gmmktime','gmdate',
			'glob','gettype','gettimeofday','gettext','getservbyport','getservbyname',
			'getrusage','getrandmax','getprotobynumber','getprotobyname','getopt','getmyuid',
			'getmypid','getmyinode','getmygid','getmxrr','getlastmod','getimagesize',
			'gethostbynamel','gethostbyname','gethostbyaddr','getenv','getdate','getcwd',
			'getallheaders','get_resource_type','get_required_files','get_parent_class','get_object_vars','get_meta_tags',
			'get_magic_quotes_runtime','get_magic_quotes_gpc','get_loaded_extensions','get_included_files','get_include_path','get_html_translation_table',
			'get_extension_funcs','get_defined_vars','get_defined_functions','get_defined_constants','get_declared_classes','get_current_user',
			'get_class_vars','get_class_methods','get_class','get_cfg_var','get_browser','fwrite',
			'function_exists','func_num_args','func_get_args','func_get_arg','ftruncate','ftp_systype',
			'ftp_ssl_connect','ftp_size','ftp_site','ftp_set_option','ftp_rmdir','ftp_rename',
			'ftp_rawlist','ftp_quit','ftp_pwd','ftp_put','ftp_pasv','ftp_nlist',
			'ftp_nb_put','ftp_nb_get','ftp_nb_fput','ftp_nb_fget','ftp_nb_continue','ftp_mkdir',
			'ftp_mdtm','ftp_login','ftp_get_option','ftp_get','ftp_fput','ftp_fget',
			'ftp_exec','ftp_delete','ftp_connect','ftp_close','ftp_chdir','ftp_cdup',
			'ftok','ftell','fstat','fsockopen','fseek','fscanf',
			'frenchtojd','fread','fputs','fpassthru','fopen','fnmatch',
			'fmod','flush','floor','flock','floatval','filetype',
			'filesize','filepro_rowcount','filepro_retrieve','filepro_fieldwidth','filepro_fieldtype','filepro_fieldname',
			'filepro_fieldcount','filepro','fileperms','fileowner','filemtime','fileinode',
			'filegroup','filectime','fileatime','file_get_contents','file_exists','file',
			'fgetss','fgets','fgetcsv','fgetc','fflush','feof',
			'fclose','ezmlm_hash','extract','extension_loaded','expm1','explode',
			'exp','exif_thumbnail','exif_tagname','exif_read_data','exif_imagetype','exec',
			'escapeshellcmd','escapeshellarg','error_reporting','error_log','eregi_replace','eregi',
			'ereg_replace','ereg','end','easter_days','easter_date','each',
			'doubleval','dngettext','dl','diskfreespace','disk_total_space','disk_free_space',
			'dirname','dir','dgettext','deg2rad','defined','define_syslog_variables',
			'define','decoct','dechex','decbin','debug_zval_dump','debug_backtrace',
			'deaggregate','dcngettext','dcgettext','dba_sync','dba_replace','dba_popen',
			'dba_optimize','dba_open','dba_nextkey','dba_list','dba_insert','dba_handlers',
			'dba_firstkey','dba_fetch','dba_exists','dba_delete','dba_close','date',
			'current','ctype_xdigit','ctype_upper','ctype_space','ctype_punct','ctype_print',
			'ctype_lower','ctype_graph','ctype_digit','ctype_cntrl','ctype_alpha','ctype_alnum',
			'crypt','create_function','crc32','count_chars','count','cosh',
			'cos','copy','convert_cyr_string','constant','connection_status','connection_aborted',
			'compact','closelog','closedir','clearstatcache','class_exists','chunk_split',
			'chr','chown','chop','chmod','chgrp','checkdnsrr',
			'checkdate','chdir','ceil','call_user_method_array','call_user_method','call_user_func_array',
			'call_user_func','cal_to_jd','cal_info','cal_from_jd','cal_days_in_month','bzwrite',
			'bzread','bzopen','bzflush','bzerrstr','bzerror','bzerrno',
			'bzdecompress','bzcompress','bzclose','bindtextdomain','bindec','bind_textdomain_codeset',
			'bin2hex','bcsub','bcsqrt','bcscale','bcpow','bcmul',
			'bcmod','bcdiv','bccomp','bcadd','basename','base_convert',
			'base64_encode','base64_decode','atanh','atan2','atan','assert_options',
			'assert','asort','asinh','asin','arsort','array_walk',
			'array_values','array_unshift','array_unique','array_sum','array_splice','array_slice',
			'array_shift','array_search','array_reverse','array_reduce','array_rand','array_push',
			'array_pop','array_pad','array_multisort','array_merge_recursive','array_merge','array_map',
			'array_keys','array_key_exists','array_intersect_assoc','array_intersect','array_flip','array_filter',
			'array_fill','array_diff_assoc','array_diff','array_count_values','array_chunk','array_change_key_case',
			'apache_setenv','apache_response_headers','apache_request_headers','apache_note','apache_lookup_uri','apache_get_version',
			'apache_child_terminate','aggregation_info','aggregate_properties_by_regexp','aggregate_properties_by_list','aggregate_properties','aggregate_methods_by_regexp',
			'aggregate_methods_by_list','aggregate_methods','aggregate','addslashes','addcslashes','acosh',
			'acos','abs','_','echo', 'print', 'global', 'static', 'exit', 'array', 'empty', 'eval', 'isset', 'unset', 'die'
			)
		),
	'SYMBOLS' => array(
		'(', ')', '[', ']', '{', '}', '!', '@', '%', '&', '*', '|', '/', '<', '>'
		),
	'CASE_SENSITIVE' => array(
		GESHI_COMMENTS => false,
		1 => false,
		2 => false,
		3 => false,
		),
	'STYLES' => array(
		'KEYWORDS' => array(
			1 => 'color: #b1b100;',
			2 => 'color: #000000; font-weight: bold;',
			3 => 'color: #000066;'
			),
		'COMMENTS' => array(
			1 => 'color: #808080; font-style: italic;',
			2 => 'color: #808080; font-style: italic;',
			'MULTI' => 'color: #808080; font-style: italic;'
			),
		'ESCAPE_CHAR' => array(
			0 => 'color: #000099; font-weight: bold;'
			),
		'BRACKETS' => array(
			0 => 'color: #66cc66;'
			),
		'STRINGS' => array(
			0 => 'color: #ff0000;'
			),
		'NUMBERS' => array(
			0 => 'color: #cc66cc;'
			),
		'METHODS' => array(
			0 => 'color: #006600;'
			),
		'SYMBOLS' => array(
			0 => 'color: #66cc66;'
			),
		'REGEXPS' => array(
			0 => 'color: #0000ff;'
			),
		'SCRIPT' => array(
			0 => '',
			1 => '',
			2 => '',
			3 => ''
			)
		),
	'URLS' => array(
		1 => '',
		2 => '',
		3 => 'http://www.php.net/{FNAME}',
		4 => ''
		),
	'OOLANG' => true,
	'OBJECT_SPLITTER' => '-&gt;',
	'REGEXPS' => array(
		0 => "[\\$]{1,2}[a-zA-Z_][a-zA-Z0-9_]*"
		),
	'STRICT_MODE_APPLIES' => GESHI_MAYBE,
	'SCRIPT_DELIMITERS' => array(
		0 => array(
			'<?php' => '?>'
			),
		1 => array(
			'<?' => '?>'
			),
		2 => array(
			'<%' => '%>'
			),
		3 => array(
			'<script language="php">' => '</script>'
			)
		),
	'HIGHLIGHT_STRICT_BLOCK' => array(
		0 => true,
		1 => true,
		2 => true,
		3 => true
		)
);
 
?>

If you're remotely familiar with PHP (or even if you're not), you can see that all that a language file consists of is a glorified variable assignment. Easy! All a language file does is assign a variable $language_data. Though still, there's a lot of indices to that array... but this section is here to break each index down and explain it to you.

4.2: Language File Conventions

There are several conventions that are used in language files. For ease of use and readability, your language files should obey the following rules:

• Indentation is in tabs, not spaces: Use tabs! There's no excuse for using spaces in this day and age, where almost every good editor allows you to change the tab width. Tabs also take up less space, and ensure proper alignment. When you indent, use tabs, and if your editor converts tabs to spaces, tell it not to.
• Strings are in single quotes: Every string in a language file should be in single quotes ('), unless you are specifying a single quote as a quotemark or escape character, in which case they can be in double quotes for readability; or if you are specifying a REGEXP (see below)
• Large arrays are multi-lined: An array with more than three or four values should be broken into multiple lines. In any case, lines should not be wider than a full-screen window (about 100 chars per line max).
• Ending brackets for multi-lined arrays on a new line: Also with a comma after them, unless the array is the last one in a parent array. See the PHP language file for examples of where to use commas.
• Use GeSHi's constants: For capatalisation, regular expressions etc. use the GeSHi constants, not their actual values.

There are more notes on each convention where it may appear in the language file sections below.

4.3: Language File Sections

This section will look at all the sections of a language file, and how they relate to the final highlighting result.

4.3.1: The Header

The header of a language file is the first lines with the big comment and the start of the variable $language_data:

<?php
/*************************************************************************************
 * <name-of-language-file.php>
 * ---------------------------------
 * Author: <name> (<e-mail address>)
 * Copyright: (c) 2004 <name> (<website URL>)
 * Release Version: 1.0.0
 * CVS Revision Version: $Revision: 1.11 $
 * Date Started: <date started>
 * Last Modified: $Date: 2006/02/25 00:43:30 $
 *
 * <name-of-language> language file for GeSHi.
 *
 * <any-comments...>
 *
 * CHANGES
 * -------
 * <date-of-release> (1.0.0)
 *  -  First Release
 *
 * TODO (updated <date-of-release>)
 * -------------------------
 * <things-to-do>
 *
 *************************************************************************************
 *
 *     This file is part of GeSHi.
 *
 *   GeSHi is free software; you can redistribute it and/or modify
 *   it under the terms of the GNU General Public License as published by
 *   the Free Software Foundation; either version 2 of the License, or
 *   (at your option) any later version.
 *
 *   GeSHi is distributed in the hope that it will be useful,
 *   but WITHOUT ANY WARRANTY; without even the implied warranty of
 *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 *   GNU General Public License for more details.
 *
 *   You should have received a copy of the GNU General Public License
 *   along with GeSHi; if not, write to the Free Software
 *   Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 *
 ************************************************************************************/

$language_data = array (

The parts in angle brackets are the parts that you change for your language file. Everything else *must* remain the same!

Here are the parts you should change:

• <name-of-language-file.php> - This should become the name of your language file. Language file names are in lower case and contain only alphanumeric characters, dashes and underscores. Language files end with .php (which you should put with the name of your language file, eg language.php)
• <name> - Your name, or alias.
• <e-mail address> - Your e-mail address. If you want your language file included with GeSHi you must include an e-mail address that refers to an inbox controlled by you.
• <website> - A URL of a website of yours (perhaps to a page that deals with your contribution to GeSHi, or your home page/blog)
• <date-started> - The date you started working on the language file. If you can't remember, guestimate.
• <name-of-language> - The name of the language you made this language file for (probably similar to the language file name).
• <any-comments> - Any comments you have to make about this language file, perhaps on where you got the keywords for, what dialect of the language this language file is for etc etc. If you don't have any comments, remove the space for them.
• <date-of-release - The date you released the language file to the public. If you simply send it to me for inclusion in a new GeSHi and don't release it, leave this blank, and I'll replace it with the date of the GeSHi release that it is first added to.

Everything should remain the same, including $Revision: 1.11 $ and $Date: 2006/02/25 00:43:30 $ (I know these may look funny but they have their purpose for those of you who don't know about CVS).

Also: I'm not sure about the copyright on a new language file. I'm not a lawyer, could someone contact me about whether the copyright for a new language file should be exclusivly the authors, or joint with me (if included in a GeSHi release)?

4.3.2: The First Indices

Here is an example from the php language file of the first indices:

	'LANG_NAME' => 'PHP',
	'COMMENT_SINGLE' => array(1 => '//', 2 => '#'),
	'COMMENT_MULTI' => array('/*' => '*/'),
	'CASE_KEYWORDS' => GESHI_CAPS_NO_CHANGE,
	'QUOTEMARKS' => array("'", '"'),
	'ESCAPE_CHAR' => '\\',

The first indices are the first few lines of a language file before the KEYWORDS index. These indices specify:

• 'LANG_NAME': The name of the language. This name should be a human-readable version of the name (e.g. HTML 4 (transitional) instead of html4trans)
• 'COMMENT_SINGLE': An array of single-line comments in your language, indexed by integers starting from 1. A single line comment is a comment that starts at the marker and goes until the end of the line. These comments may be any length > 0, and since they can be styled individually, can be used for other things than comments (for example the Java language file defines "import" as a single line comment). If you are making a language that uses a ' (apostrophe) as a comment (or in the comment marker somewhere), use double quotes. e.g.: "'"
• 'COMMENT_MULTI': Used to specify multiline comments, an array in the form 'OPEN' => 'CLOSE'. Unfortunately, all of these comments you add here will be styled the same way (an area of improvement for GeSHi 1.2.X). These comment markers may be any length > 0.
• 'CASE_KEYWORDS': Used to set whether the case of keywords should be changed automatically as they are found. For example, in an SQL or BASIC dialect you may want all keywords to be upper case. The accepted values for this are:
  • GESHI_CAPS_UPPER: Convert the case of all keywords to upper case.
  • GESHI_CAPS_LOWER: Convert the case of all keywords to lower case.
  • GESHI_CAPS_NO_CHANGE: Don't change the case of any keyword.
• 'QUOTEMARKS': Specifies the characters that mark the beginning and end of a string. This is another example where if your language includes the ' string delimiter you should use double quotes around it.
• 'ESCAPE_CHAR': Specifies the escape character used in all strings. If your language does not have an escape character then make this the empty string (''). This is not an array! If found, any character after an escape character and the escape character itself will be highlighted differently, and the character after the escape character cannot end a string.

4.3.3: Keywords

Keywords will make up the bulk of a language file. In this part you add keywords for your language, including inbuilt functions, data types, predefined constants etc etc.

Here's a (shortened) example from the php language file:

	'KEYWORDS' => array(
		1 => array(
			'include', 'require', 'include_once', 'require_once',
			'for', 'foreach', 'as', 'if', 'elseif', 'else', 'while', 'do', 'endwhile', 'endif', 'switch', 'case', 'endswitch',
			'return', 'break'
			),
		2 => array(
			'null', '__LINE__', '__FILE__',
			'false', '<?php', '?>', '<?',
			'<script language=', '</script>',
			'true', 'var', 'default',
			'function', 'class', 'new',
			'__FUNCTION__', '__CLASS__', '__METHOD__', 'PHP_VERSION',
			'PHP_OS', 'DEFAULT_INCLUDE_PATH', 'PEAR_INSTALL_DIR', 'PEAR_EXTENSION_DIR',
			'PHP_EXTENSION_DIR', 'PHP_BINDIR', 'PHP_LIBDIR', 'PHP_DATADIR', 'PHP_SYSCONFDIR',
			'PHP_LOCALSTATEDIR', 'PHP_CONFIG_FILE_PATH', 'PHP_OUTPUT_HANDLER_START', 'PHP_OUTPUT_HANDLER_CONT',
			'PHP_OUTPUT_HANDLER_END', 'E_ERROR', 'E_WARNING', 'E_PARSE', 'E_NOTICE',
			'E_CORE_ERROR', 'E_CORE_WARNING', 'E_COMPILE_ERROR', 'E_COMPILE_WARNING', 'E_USER_ERROR',
			'E_USER_WARNING', 'E_USER_NOTICE', 'E_ALL'
			),
		3 => array(
			'zlib_get_coding_type','zend_version','zend_logo_guid','yp_order','yp_next',
			'yp_match','yp_master','yp_get_default_domain','yp_first','yp_errno','yp_err_string',
			'yp_cat','yp_all','xml_set_unparsed_entity_decl_handler','xml_set_start_namespace_decl_handler','xml_set_processing_instruction_handler','xml_set_object',
			'xml_set_notation_decl_handler','xml_set_external_entity_ref_handler','xml_set_end_namespace_decl_handler','xml_set_element_handler','xml_set_default_handler','xml_set_character_data_handler',
			'xml_parser_set_option','xml_parser_get_option','xml_parser_free','xml_parser_create_ns','xml_parser_create','xml_parse_into_struct'
			)
		),

You can see that the index 'KEYWORDS' refers to an array of arrays, indexed by positive integers. In each array, there are some keywords (in the actual php language file there is in fact many more keywords in the array indexed by 3). Here are some points to note about these keywords:

• Indexed by positive integers: Use nothing else! I may change this in 1.2.X, but for the 1.0.X series, use positive integers only.
• Keywords sorted in reverse: Keywords *should* be sorted in reverse order. I know that many of the language files I've made do not follow this rule, but that's because I made the files before I discovered the following issue with GeSHi: If you have two keywords, as and ascfor example, then when GeSHi encounters the "as" keyword it will be highlighted even if the "as" is part of "asc". I would get GeSHi to reverse-sort keyword arrays, but there's no harm in you doing this yourself and saving some processing time. At the least (and in fact this is sufficient) you should swap any keywords you come across that would be caught by this issue.
• Keywords are case sensitive (sometimes): If your language is case-sensitive, the correct casing of the keywords is defined as the case of the keywords in these keyword arrays. If you check the java language file you will see that everything is in exact casing. So if any of these keyword arrays are case sensitive, put the keywords in as their correct case! (note that which groups are case sensitive and which are not is configurable, see later on)
• Keywords must be in htmlentities() form: All keywords should be written as if they had been run through the php function htmlentities(). E.g, the keyword is &lt;foo&gt;, not <foo>
• Don't use keywords to highlight symbols: Just don't. It doesn't work, and there will be seperate support for symbols later.
• Markup Languages are special cases: Check the html4strict language file for an example: keywords have to be specified twice for opening tags and twice for each closing tag.

4.3.4: Symbols and Case Sensitivity

So you've put all the keywords for your language in? Now for a breather before we style them :). Symbols define what symbols your language uses, these are things like colons, brackets/braces, and other such general punctuation (it's important to note that in the 1.0.X series these will not be used), and case sensitivity alludes to the previous section: here you can set which keyword groups are case sensitive.

Here's an example:

	'SYMBOLS' => array(
		'(', ')', '[', ']', '{', '}', '!', '@', '%', '&', '*', '|', '/', '<', '>'
		),
	'CASE_SENSITIVE' => array(
		GESHI_COMMENTS => false,
		1 => false,
		2 => false,
		3 => false,
		),

• 'SYMBOLS': An array of the symbols used in your language. Please note that in 1.0.X these are not used. Hopefully they will be used in 1.2.X.
• 'CASE_SENSITIVE': Note the GESHI_COMMENTS! This is used to set whether comments are case sensitive or not (for example, BASIC has the REM statement which while not being case sensitive is still alphanumeric, and as in the example given before about the Java language file using "import" as a single line comment, this can be useful sometimes. true if comments are case sensitive, false otherwise. All of the other indices correspond to indices in the 'KEYWORDS' section (see above).

4.3.5: Styles for your Language File

This is the fun part! Here you get to choose the colours, fonts, backgrounds and anything else you'd like for your language file.

Here's an example:

	'STYLES' => array(
		'KEYWORDS' => array(
			1 => 'color: #b1b100;',
			2 => 'color: #000000; font-weight: bold;',
			3 => 'color: #000066;'
			),
		'COMMENTS' => array(
			1 => 'color: #808080; font-style: italic;',
			2 => 'color: #808080; font-style: italic;',
			'MULTI' => 'color: #808080; font-style: italic;'
			),
		'ESCAPE_CHAR' => array(
			0 => 'color: #000099; font-weight: bold;'
			),
		'BRACKETS' => array(
			0 => 'color: #66cc66;'
			),
		'STRINGS' => array(
			0 => 'color: #ff0000;'
			),
		'NUMBERS' => array(
			0 => 'color: #cc66cc;'
			),
		'METHODS' => array(
			0 => 'color: #006600;'
			),
		'SYMBOLS' => array(
			0 => 'color: #66cc66;'
			),
		'REGEXPS' => array(
			0 => 'color: #0000ff;'
			),
		'SCRIPT' => array(
			0 => '',
			1 => '',
			2 => '',
			3 => ''
			)
		),

Note that all style rules should end with a semi-colon! This is important: GeSHi may add extra rules to the rules you specify (and will do so if a user tries to change your styles on the fly), so the last semi-colon in any stylesheet rule is important.

All strings here should contain valid stylesheet declarations (it's also find to have the empty string).

• 'KEYWORDS': This is an array, from keyword index to style. The index you use is the index you used in the keywords section to specify the keywords belonging to that group.
• 'COMMENTS': This is an array, from single-line comment index to style for that index. The index 'MULTI' is used for multiline comments (and cannot be an array)
• 'ESCAPE_CHAR' down to 'REGEXPS': These are arrays with only one index: 0. You cannot add other indices to these arrays.
• 'SCRIPT': For languages that use script delimiters, this is where you can style each block of script. For example, HTML and XML have blocks that begin with < and end with > (i.e. tags) and blocks that begin with & and end with ; (i.e. character entities), and you can set a style to apply to each whole block. You specify the delimiters for the blocks below. Note that many languages will not need this feature.

4.3.6: URLs for Functions

This section lets you specify a url to visit for each keyword group. Useful for pointing functions at their online manual entries.

Here is an example:

	'URLS' => array(
		1 => '',
		2 => '',
		3 => 'http://www.php.net/{FNAME}',
		4 => ''
		),

The indices of this array correspond to the keyword groups you specified in the keywords section. The string {FNAME} marks where the name of the function is substituted in. So for the example above, if the keyword being highlighted is "echo", then the keyword will be a URL pointing to http://www.php.net/echo. Because some languages (Java!) don't keep a uniform URL for functions/classes, you may have trouble in creating a URL for that language (though look in the java language file for a novel solution to it's problem)

4.3.7: Object Orientation Support

Now we're reaching the most little-used section of a language file, which includes such goodies as object orientation support and context support. GeSHi can highlight methods and data fields of objects easily, all you need to do is to tell it to do so and what the "splitter" is between object/method etc.

Here's an example:

	'OOLANG' => true,
	'OBJECT_SPLITTER' => '->',

If your language has object orientation, the value of 'OOLANG' is true, otherwise it is false. If it is object orientated, in the 'OBJECT_SPLITTER' value you put the htmlentities() version of the "splitter" between objects and methods/fields. If it is not, then make this the empty string.

4.3.8: Using Regular Expressions

Regular expressions are a good way to catch any other lexic that fits certain rules but can't be listed as a keyword. A good example is variables in PHP: variables always start with either one or two "$" signs, then alphanumeric characters (a simplification). This is easy to catch with regular expressions.

And new to version 1.0.2, there is an advanced way of using regular expressions to catch certain things but highlight only part of those things. This is particularly useful for languages like XML.

Here is an example (this time the PHP file merged with the XML file):

		0 => array(
			GESHI_SEARCH => '(((xml:)?[a-z\-]+))(=)',
			GESHI_REPLACE => '\\1',
			GESHI_MODIFIERS => '',
			GESHI_BEFORE => '',
			GESHI_AFTER => '\\4'
			),
		1 => array(
			GESHI_SEARCH => '(</?[a-z0-9]*(>)?)',
			GESHI_REPLACE => '\\1',
			GESHI_MODIFIERS => '',
			GESHI_BEFORE => '',
			GESHI_AFTER => ''
			),
		2 => "[\\$]{1,2}[a-zA-Z_][a-zA-Z0-9_]*"

As you can see there are two formats. One is the "simple" format used in GeSHi < 1.0.2, and the other is a more advanced syntax. Firstly, the simple syntax:

• May be in double quotes: To make it easier for those who always place their regular expressions in double quotes, you may place any regular expression here in double quotes if you wish.
• Don't use curly brackets where possible: If you want to use curly brackets (()) then by all means give it a try, but I'm not sure whether under some circumstances GeSHi may throw a wobbly. You have been warned! If you want to use brackets, it would be better to use the advanced syntax.
• Don't use the "everything" regex: (That's the .*? regex). Use advanced syntax instead.

And now for advanced syntax, which gives you much more control over exactly what is highlighted:

• GESHI_SEARCH: This element specifies the regular expression to search for. If you plan to capture the output, use brackets (()). See how in the first example above, most of the regular expression is in one set of brackets (with the equals sign in other brackets). You should make sure that the part of the regular expression that is supposed to match what is highlighted is in brackets.
• GESHI_REPLACE: This is what the stuff matched by the regular expression will be replaced with. If you've grouped the stuff you want highlighted into brackets in the GESHI_SEARCH element, then you can use \\number to match that group, where number is a number corresponding to how many open brackets are between the open bracket of the group you want highlighted and the start of the GESHI_SEARCH string + 1. This may sound confusing, and it probably is, but if you're familiar with how PHP's regular expressions work you should understand. In the example above, the opening bracket for the stuff we want highlighted is the very first bracket in the string, so the number of brackets before that bracket and the start of the string is 0. So we add 1 and get our replacement string of \\1 (whew!).
   
  If you didn't understand a word of that, make sure that there are brackets around the string in GESHI_SEARCH and use \\1 for GESHI_REPLACE ;)
• GESHI_MODIFIERS: Specify modifiers for your regular expression. If your regular expression includes the everything matcher (.*?), then your modifiers should include "s" and "i" (e.g. use 'si' for this).
• GESHI_BEFORE:Specifies a bracket group that should appear before the highlighted match (this bracketed group will not be highlighted). Use this if you had to match what you wanted by matching part of your regexp string to something before what you wanted to highlight, and you don't want that part to disappear in the highlighted result.
• GESHI_AFTER:Specifies a bracket group that should appear after the highlighted match (this bracketed group will not be highlighted). Use this if you had to match what you wanted by matching part of your regexp string to something after what you wanted to highlight, and you don't want that part to disappear in the highlighted result.

Is that totally confusing? Here's the test for if you're an android or not: If you found that perfectly understandable then you're an android ;). Here's a better example:

Let's say that I'm making a language, and variables in this language always start with a dollar sign ($), are always written in lowercase letters and always end with an ampersand (&). eg:

I want to highlight only the text between the $ and the &. How do I do that? With simple regular expressions I can't, but with advanced, it's relatively easy:

		1 => array(
			GESHI_SEARCH => '(\$)([a-z]+)(&)',   // search for a dollar sign, then one or more of the characters a-z, then an ampersand
			GESHI_REPLACE => '\\2',                  // we wanna highlight the characters, which are in the second bracketed group
			GESHI_MODIFIERS => '',                   // no modifiers, since we're not matching the "anything" regex
			GESHI_BEFORE => '\\1',                   // before the highlighted characters should be the first bracketed group (always a dollar sign in this example)
			GESHI_AFTER => '\\3'                     // after the highlighted characters should be the third bracketed group (always an ampersand in this example)
			),

So if someone tried to highlight using my language, all cases of $foo& would turn into:

(which would of course be viewed in a browser to get something like $foo&)

4.3.9: Contextual Highlighting and Strict Mode

For languages like HTML, it's good if we can highlight a tag (like <a> for example). But how do we stop every single "a" in the source getting highlighted? What about for attributes? If I've got the word "colspan" in my text I don't want that highlighted! So how do you tell GeSHi not to highlight in that case? You do it with "Strict Blocks".

Here is an example:

	'STRICT_MODE_APPLIES' => GESHI_MAYBE,
	'SCRIPT_DELIMITERS' => array(
		0 => array(
			'<?php' => '?>'
			),
		1 => array(
			'<?' => '?>'
			),
		2 => array(
			'<%' => '%>'
			),
		3 => array(
			'<script language="php">' => '</script>'
			)
		),
	'HIGHLIGHT_STRICT_BLOCK' => array(
		0 => true,
		1 => true,
		2 => true,
		3 => true
		)

What is strict mode? Strict mode says that highlighting only occurs inside the blocks you specify. You can see from the example above that highlighting will only occur if the source is inside <?php ... ?> (though note the GESHI_MAYBE!). Here are some points about strict highlighting:

• 'STRICT_MODE_APPLIES': This takes three values (all constants):
  • GESHI_ALWAYS: Strict mode always applies for all of the blocks you specify. Users of your language file cannot turn strict mode off. This should be used for markup languages.
  • GESHI_NEVER: Strict mode is never used. Users of your language file cannot turn strict mode on. Use this value if there is no such thing as a block of code that would not be highlighted in your language (most languages, like C, Java etc. use this because anything in a C file should be highlighted).
  • GESHI_MAYBE: Strict mode *sometimes* applies. It defaults to "off". Users can turn strict mode on if they please. If strict mode is off then everything in the source will be highlighted, even things outside the strict block markers. If strict mode is on the nothing outside strict block markers will be highlighted.
• SCRIPT_DELIMITERS: This is an array of script delimiters, in the format of the above. The indices are use in the 'SCRIPT' part of the styles section for highlighting everything in a strict block in a certain way. For example, you could set up your language file to make the background yellow of any code inside a strict block this way. The delimiters are in the form 'OPEN' => 'CLOSE'. Delimiters can be of any length > 0. Delimiters are not formatted as if they were run through htmlentities()!
• 'HIGHLIGHT_STRICT_BLOCK': specifies whether any highlighting should go on inside each block. Most of the time this should be true, but for example, in the XML language file highlighting is turned off for blocks beginning with <!DOCTYPE and ending with >. However, you can still style the overall block using the method described above, and the XML language file does just that.

4.3.10: Tidying Up

All language files should end with:

);

?>

5: Method/Constant Reference

I'm afraid I have been lying for a little while about this now! Since 1.0.7 I have been including a phpdoc API for the sourcecode in the api directory, but have forgot to update the documentation! However, it is available, and may assist you in coding, especially for plugin coders.

That's all, folks!

I've improved the documentation greatly from version 1.0.1, but there may still be problems with it, or it may still be confusing for you. Or perhaps I was just plain wrong about one point! If so, contact me and I'll do my best to sort it out.

In case you were wondering, I've finished development of the 1.0.X thread of GeSHi. The only releases I'll make in this thread will be of the bug-fix/add language files type. In particular, version 1.0.2 was a "concept" release - testing how far I could take the highlighting idea (as well as ideas from others).

I'm planning a code rewrite for 1.2.X, which will be based on a new engine - a "psuedo-tokenizer" engine. Hopefully it will massively reduce the server load and time taken (by almost eliminating regexps), while providing superior highlighting. But fear not! The interface and method names should all remain the same ^_^ (though I can't say the same for language files!)

And finally, a couple of people have been asking me: how did you generate that documentation? The amazing answer is: my brain. And yes, it took a long time, and I don't reccommend doing it this way. And yes, you can borrow the styles if you like, though flick me an e-mail if you do.

Anyway, enough blather from me. Get GeSHi working for you already! :D

Nigel McNie
nigel@geshi.org
http://qbnz.com/highlighter/