2006-05-21, 16:07 h
Holger
Posts: 8116
User
|
php code:
<?
#global
$keyword = '[blue][b]${0}[/b][/blue]';
$libfunc = '[blue]${0}[/blue]';
$literal = '[green]${0}[/green]';
$escape = '[blue]${0}[/blue]';
$preproc = '[red]${0}[/red]';
$include = '${1}[green]${2}[/green]';
$comment = '[gray]${0}[/gray]';
$op_sep = '[black]${0}[/black]';
#...
while(preg_match("/^(.*?)\s*[code language="?([^"]*)"?\]\s*(.*?)\s*\[\/code\]\r?\n?(.*?)$/is", $out, $matches)) {
$orig_lang = $matches[2];
$lang = strtolower($orig_lang);
$code = $matches[3];
$code = preg_replace("/x1Ax1A/", "\[", $code);
$code = preg_replace("/\[(/?+(?:[biu]|black|red|green|blue|yellow|gray|code|img|url))b/", "x1Ax1A$1", $code);
$code = html_entity_decode($code);
highlight($code, $lang);
$code = htmlentities($code);
$out = $matches[1].'</font><blockquote>'.get_last_font().$orig_lang.' code:</font><hr><font face="Courier New, Courier, monospace" color="#400080"><pre>'.$code.'</pre></font><hr></blockquote>'.get_last_font().$matches[4];
}
// ---- Ohne Syntax-Highlighting
$out = preg_replace("/[code\]\s*(.*?)\s*\[\/code\]\r?\n?/is", '</font><blockquote>'.get_last_font().'code:</font><hr><font face="Courier New, Courier, monospace" color="#400080"><pre>${1}</pre></font><hr></blockquote>'.get_last_font(), $out);
#...
function highlight(&$code, $lang)
{
global $keyword;
global $libfunc;
global $literal;
global $escape;
global $preproc;
global $include;
global $comment;
global $op_sep;
//sgml document & html declaration tags
if(preg_match("/^(x?html|php|asp|hsc|jsp)$/", $lang)) {
$comments="<!--.*?-->";
$decl="<![^!>]*+>";
$word="(?:\w[\w:\-_]*+\b)";
$tags="(<\/\s*+$word\s*+>)|(<\s*+$word)((?:\s++$word(?:=(?:"[^"]*+"|$word))?+)*+\s*+>)";
$script="<%(?:[^%]*+(?:%(?!>))*+)*+%>|<\?(?:[^?]*+(?:\?(?!>))*+)*+\?".">";
$entity="\&(?:\#(?:[xX][0-9A-Fa-f]++|[0-9]++)|[a-zA-Z][a-zA-Z\-0-9]*+);";
preg_match_all("/($comments)|($decl)|($script)|$tags|(<?+[^<]*+)/", $code, $chunks);
$chunks[1]=preg_replace("/.+/s", $comment, $chunks[1]);
$chunks[2]=preg_replace("/.+/s", $preproc, $chunks[2]);
$keywords=&getKeywords('html');
$myKeyword=preg_replace("/\$\{0\}/", '\${1}', $op_sep)
.preg_replace("/\$\{0\}/", '\${2}', $keyword);
$chunks[5]=preg_replace("/^(<\s*\/?\s*)($keywords)\b/", $myKeyword, $chunks[5]);
$myKeyword.='${3}'.preg_replace("/\$\{0\}/", '>', $op_sep);
$chunks[4]=preg_replace("/^(<\s*\/?\s*)($keywords)\b(.*)>/", $myKeyword, $chunks[4]);
$chunks[6]=preg_replace("/[^>]++/", $libfunc, $chunks[6]);
$chunks[6]=preg_replace("/(?<==)(?:"[^"]*+"|[^\s\r\n=">]+)/", $literal, $chunks[6]);
$chunks[6]=preg_replace("/>$/", $op_sep, $chunks[6]);
$chunks[6]=preg_replace("/$entity/", $escape, $chunks[6]);
$chunks[7]=preg_replace("/$entity/", $escape, $chunks[7]);
$sublang=preg_match("/^(x?html)$/", $lang)? "javascript":
$sublang=='jsp'? 'java' :$lang.'1';
foreach($chunks[3] as &$subcode)
if($subcode) highlight($subcode, $sublang);
$chunks=array_map(NULL, $chunks[1], $chunks[2],
$chunks[3], $chunks[4],$chunks[5], $chunks[6], $chunks[7]);
$code=""; foreach($chunks as $part) $code.=join("", $part);
}
// c-style syntax, comments, literals
if(preg_match("/^(?:c(?:\+\+)?|java(?:script)?|perl|php1|asp1)$/", $lang)) {
$comments ="\/\/[^\r\n]*";
$comments.="|\/\*(?:[^*]*+(?:\*(?!\/))*+)*+\*\/";
if("php1"==$lang) {
$comments.="|\#[^rn]*";
}
// string literals include " escapes, single line only
$literals =""(?:[^"\n\r\\]*+(?:\\[^\r\n\\]|\\\\)?+)*+"";
// 'x' char literals or 'xyz' string literals, depending on language
$literals.="|'(?:\\'|[^'\n\r])*+'";
// numerical values, e.g. 0 1.2 .3 4.5E6 78f 9D 0.12d 3.4E5F 6789L
$literals.="|-?+(?:\.\d++|\d++(?:\.\d++)?+)(?:[eE]\d++|[Ll]\b)?+[dDfF]?";
$literals.="|0[xX][0-9a-fA-F]++[Ll]?";
// now everything not a comment, literal or preprocessor instruction
$keywords=&getKeywords($lang);
$preprocs=&getPreprocessor($lang);
$libfuncs=&getLibFuncs($lang);
if("c++"==$lang) {
$literals.="|(?:\b(?:true|false)\b)";
} elseif(preg_match("/^java(?:script)?$/", $lang)) {
$literals.="|(?:\b(?:true|false|null)\b)";
}
preg_match_all("/($preprocs)|($comments)|($literals)|([\w_$]+)|(x1Ax1A|\S)|([\n\r\s]+)/", $code, $chunks);
$chunks[1]=preg_replace("/.+/s", $preproc, $chunks[1]);
$chunks[1]=preg_replace("/(#include\s*)(<[^>]+>|"[^"]*")/s", $include, $chunks[1]);
$chunks[2]=preg_replace("/.+/s", $comment, $chunks[2]);
$chunks[3]=preg_replace("/.+/", $literal, $chunks[3]);
$chunks[3]=preg_replace("/(?:\\(?:x1Ax1A|.))+/", $escape, $chunks[3]);
$chunks[4]=preg_replace("/^(?:$keywords)$/", $keyword, $chunks[4]);
$chunks[4]=preg_replace("/^(?:$libfuncs)$/", $libfunc, $chunks[4]);
if("php1"==$lang) { // case insensitive keywords
$keywords=&getKeywords("php2");
$chunks[4]=preg_replace("/^(?:$keywords)$/i", $keyword, $chunks[4]);
}
$chunks[5]=preg_replace("/.+/", $op_sep, $chunks[5]);
$chunks=array_map(NULL, $chunks[1], $chunks[2],
$chunks[3], $chunks[4],$chunks[5], $chunks[6]);
$code=""; foreach($chunks as $part) $code.=join("", $part);
} elseif("basic"==$lang) {
$comments ="(?:\bREM\b|'|x60)[^\r\n]*";
$literals =""[^"\n\r]*+"";
// 'x' char literals or 'xyz' string literals, depending on language
$literals.="|'(?:\\'|[^'\n\r])*+'";
// numerical values, e.g. 0 1.2 .3 4.5E6 78f 9D 0.12d 3.4E5F 6789L
$literals.="|-?+(?:\.\d++|\d++(?:\.\d++)?+)(?:[eE]\d++|[%&]\b)?+!?";
$literals.="|[&]h[0-9a-fA-F]++\b[%&]?|%[01]++\b[%&]?";
// now everything not a comment, literal or preprocessor instruction
$keywords=&getKeywords($lang);
$preprocs="(?:\bREM\b|'|x60)s*\$"."i(?:nclude)?+\b[^\r\n]*|^[a-z]*:";
preg_match_all("/($preprocs)|($comments)|($literals)|([\w_]+[%!&$]?)|(x1Ax1A|\S)|([\n\r\s]+)/i", $code, $chunks);
$chunks[1]=preg_replace("/.+/s", $preproc, $chunks[1]);
$chunks[1]=preg_replace("/(\x24i(?:nclude)?+\s*+)("[^"]*"|.*)/is", $include, $chunks[1]);
$chunks[2]=preg_replace("/.+/s", $comment, $chunks[2]);
$chunks[3]=preg_replace("/.+/", $literal, $chunks[3]);
$chunks[4]=preg_replace("/^(?:$keywords)[%!&$]?$/i", $keyword, $chunks[4]);
$chunks[5]=preg_replace("/.+/", $op_sep, $chunks[5]);
$chunks=array_map(NULL, $chunks[1], $chunks[2],
$chunks[3], $chunks[4],$chunks[5], $chunks[6]);
$code=""; foreach($chunks as $part) $code.=join("", $part);
}
}
function &getKeywords($lang)
{
if($lang=="html") {
$keywords ="a|abbr|acronym|address|applet|area|b|base|basefont";
$keywords.="|bdo|big|blockquote|body|br|button|caption|center";
$keywords.="|cite|code|col|colgroup|dd|del|dfn|dir|div|dl|dt|em";
$keywords.="|fieldset|font|form|frame|frameset|h[1-6]|head|hr";
$keywords.="|html|i|iframe|img|input|ins|isindex|kbd|label";
$keywords.="|legend|li|link|map|menu|meta|noframes|noscript";
$keywords.="|object|ol|optgroup|option|p|param|pre|q|s|samp";
$keywords.="|script|select|small|span|strike|strong|style|sub";
$keywords.="|sup|table|tbody|td|textarea|tfoot|th|thead|title";
$keywords.="|tr|tt|u|ul|var";
} elseif(preg_match("/^c(\+\+)?$/", $lang)) {
$keywords ="auto|break|case|char|const|continue|default|do|double|else";
$keywords.="|enum|extern|float|for|goto|if|int|long|register|return";
$keywords.="|short|signed|sizeof|static|struct|switch|typedef|union";
$keywords.="|unsigned|void|volatile|while";
if("c++"==$lang) // C++ only
{
$keywords.="|asm|bool|catch|class|const_cast|delete|dynamic_cast";
$keywords.="|explicit|friend|inline|mutable|namespace|new|operator";
$keywords.="|private|protected|public|reinterpret_cast|static_cast";
$keywords.="|template|this|throw|try|typeid|typename|using|virtual";
$keywords.="|wchar_t";
}
} elseif("java"==$lang) {
$keywords ="abstract|boolean|break|byte|case|catch|char|class";
$keywords.="|continue|default|do|double|else|extends|final|finally";
$keywords.="|float|for|if|implements|import|instanceof|int|interface";
$keywords.="|long|native|new|package|private|protected|public|return";
$keywords.="|short|static|strictfp|super|switch|synchronized|this|throw";
$keywords.="|throws|transient|try|void|volatile|while|assert";
// the following are reserved but unused
$keywords.="|goto|const";
} elseif("php1"==$lang) {
// case insensitive keywords
$keywords ="array|as|bool|boolean|break|case|compact|continue|count";
$keywords.="|current|default|define|defined|die|do|double|doubleval";
$keywords.="|each|echo|else|elseif|empty|end|eregi?(?:_replace)?+";
$keywords.="|eval|exit|extract|float|floatval|for|foreach|function";
$keywords.="|gettype|if|in_array|include(?:_once)?+|int|integer|intval";
$keywords.="|isset|key|leak|list|list|mail|next|object|pos|prev";
$keywords.="|print(?:f|_r)?+|range|readfile|real|require(?:_once)?+";
$keywords.="|reset|return|serialize|settype|shuffle|sizeof|sleep";
$keywords.="|(?:nat(?:case)?|u?[ak]?r?)?sort|spliti?|sql_regcase";
$keywords.="|string|strval|switch|uniqid|unserialize|unset|usleep";
$keywords.="|var_dump|var_export|virtual|while|preg_(?:match(?:_all)?+|replace)";
} elseif("php2"==$lang) {
// case sensitive keywords (constants)
$keywords ="NULL|PHP_OS|PHP_VERSION|TRUE|FALSE|E_ERROR|E_WARNING";
$keywords.="|E_PARSE|E_NOTICE|E_CORE_ERROR|E_CORE_WARNING|E_STRICT";
} elseif("perl"==$lang) {
$keywords ="abs|atan2|bless|caller|chomp|chop|chr|continue";
$keywords.="|cos|crypt|dbmclose|dbmopen|defined|delete|die";
$keywords.="|do|dump|each|eval|exists|exit|exp|formline";
$keywords.="|gmtime|goto|grep|hex|import|index|int|join";
$keywords.="|keys|last|lc|lcfirst|length|local|localtime|log";
$keywords.="|map|my|next|no|oct|ord|our|pack|package|pop|pos";
$keywords.="|push|quotemeta|rand|redo|ref|require|reset";
$keywords.="|return|reverse|rindex|scalar|shift|sin|sort";
$keywords.="|splice|split|sprintf|sqrt|srand|study|sub|substr";
$keywords.="|tie|tied|time|times|uc|ucfirst|undef|unpack";
$keywords.="|unshift|untie|use|values|wantarray";
} elseif("javascript"==$lang) {
$keywords ="as|break|case|catch|class|const|continue|default|delete|";
$keywords.="do|else|export|extends|finally|for|function|if|import";
$keywords.="|in|instanceof|is|namespace|new|package|private|public";
$keywords.="|return|super|switch|this|throw|try|typeof|use|var";
$keywords.="|void|while|with";
// the following are reserved but unused
$keywords.="|abstract|debugger|enum|goto|implements|interface|native";
$keywords.="|protected|synchronized|throws|transient|volatile";
} elseif("basic"==$lang) {
$keywords ="if|then|else|elsif|end|for|to|step|next|loop|repeat|until";
$keywords.="|do|while|wend|declare|sub|shared|common|call|return";
$keywords.="|print|at|using|locate|screen|color|palette|input|timer|sleep";
$keywords.="|dim|as|integer|string|float|open|input|output|close|seek";
$keywords.="|function|let|on|error|goto|gosub|on|off|or|and|xor|not|mod";
$keywords.="|str|chr|asc|bin|oct|hex|mid|left|right|peek[bwl]?+|poke[bwl]?";
$keywords.="|csrlin|pos|cls|line|circle|paint|inkey|beep|varptr|sadd";
$keywords.="|randomize|int|rnd|resume|library|data|read|restore|val";
} else {//unknown... most likely keywords in every language
$keywords ="break|case|continue|default|do|else|for|if|return|switch|while";
}
return $keywords;
}
function &getPreprocessor($lang) {
if(preg_match("/^c(\+\+)?$/", $lang)) {
$preprocs="\s*+\#(?:include|define|ifdef|endif|else|pragma)\b";
// macros may span multiple lines if linebreak is escaped with
$preprocs.="(?:(?:[^\n\r\\]*+(?:\\[^\n\r])*+)*+(?:\\\r?\n?)?+)*+";
return $preprocs;
} else return "(?!X)X";//no preprocessor
}
function &getLibFuncs($lang) {
if(preg_match("/^c(\+\+)?$/", $lang)) {
$libfuncs="U?(?:BYTE|WORD|LONG)|[AB]PTR|BOOL|TRUE|FALSE|NULL";
$libfuncs.="|TAG_(?:END|IGNORE|MORE)";
$libfuncs.="|(?:Open|Close)(?:Library|Device)";
$libfuncs.="|f?+printf|fopen|fclose|fwrite";
if("c++"==$lang) {
$libfuncs.="|std|cout|cin|endl|map|vector|pair";
}
return $libfuncs;
} elseif("java"==$lang) {
return "Object|String|Class(?:Loader)?+";
}
return "(?!X)X";//no libfuncs yet
}
?>
--
Good coders do not comment. What was hard to write should be hard to read too.
[ - Answer - Quote - Direct link - ]
|
