geshi.php

一款基于PHP的网络日记程序。WikyBlog支持：多用户的 BLOG

    					    $attributes = ' class="st0"';
    					}
    					$char = "<span$attributes>" . $hq;
    					$i += strlen($hq) - 1;
    					$HARDQUOTE_OPEN = true;
    					$result .= $this->parse_non_string_part( $stuff_to_parse );
    					$stuff_to_parse = '';
    				    } elseif ($char == $this->language_data['ESCAPE_CHAR'] && $STRING_OPEN != '')
    				    {
                            // An escape character
							if (!$ESCAPE_CHAR_OPEN) {
    							$ESCAPE_CHAR_OPEN = !$HARDQUOTE_OPEN;  // true unless $HARDQUOTE_OPEN
    							if ($HARDQUOTE_OPEN)
    								foreach ($this->language_data['HARDESCAPE'] as $hard)
    								{
    									if (substr($part, $i, strlen($hard)) == $hard)
    									{
								$ESCAPE_CHAR_OPEN = true;
    										break;
    									}
    								}
    							if ($ESCAPE_CHAR_OPEN && $this->lexic_permissions['ESCAPE_CHAR']) {
									if (!$this->use_classes) {
										$attributes = ' style="' . $this->language_data['STYLES']['ESCAPE_CHAR'][0] . '"';
									} else {
										$attributes = ' class="es0"';
									}
									$char = "<span$attributes>" . $char;
                                    if (substr($code, $i + 1, 1) == "\n") {
                                        // escaping a newline, what's the point in putting the span around
                                        // the newline? It only causes hassles when inserting line numbers
                                        $char .= '</span>';
                                        $ESCAPE_CHAR_OPEN = false;
                                    }
								}
							} else {
								$ESCAPE_CHAR_OPEN = false;
								if ($this->lexic_permissions['ESCAPE_CHAR']) {
									$char .= '</span>';
								}
							}
						} elseif ($ESCAPE_CHAR_OPEN) {
							if ($this->lexic_permissions['ESCAPE_CHAR']) {
								$char .= '</span>';
							}
							$ESCAPE_CHAR_OPEN = false;
							$test_str = $char;
						} elseif ($STRING_OPEN == '') {
							// Is this a multiline comment?
							foreach ($this->language_data['COMMENT_MULTI'] as $open => $close) {
								$com_len = strlen($open);
								$test_str = substr( $part, $i, $com_len );
								$test_str_match = $test_str;
								if ($open == $test_str) {
									$COMMENT_MATCHED = true;
                                    //@todo If remove important do remove here
									if ($this->lexic_permissions['COMMENTS']['MULTI'] ||
                                        $test_str == GESHI_START_IMPORTANT) {
										if ($test_str != GESHI_START_IMPORTANT) {
											if (!$this->use_classes) {
												$attributes = ' style="' . $this->language_data['STYLES']['COMMENTS']['MULTI'] . '"';
											} else {
												$attributes = ' class="coMULTI"';
											}
											$test_str = "<span$attributes>" . @htmlspecialchars($test_str, ENT_COMPAT, $this->encoding);
										} else {
											if (!$this->use_classes) {
												$attributes = ' style="' . $this->important_styles . '"';
											} else {
												$attributes = ' class="imp"';
											}
											// We don't include the start of the comment if it's an
											// "important" part
											$test_str = "<span$attributes>";
										}
									} else {
										$test_str = @htmlspecialchars($test_str, ENT_COMPAT, $this->encoding);
									}

									$close_pos = strpos( $part, $close, $i + strlen($close) );

									if ($close_pos === false) {
										$close_pos = strlen($part);
									}

									// Short-cut through all the multiline code
									$rest_of_comment = @htmlspecialchars(substr($part, $i + $com_len, $close_pos - $i), ENT_COMPAT, $this->encoding);
									if (($this->lexic_permissions['COMMENTS']['MULTI'] ||
                                        $test_str_match == GESHI_START_IMPORTANT) &&
                                        ($this->line_numbers != GESHI_NO_LINE_NUMBERS ||
                                        count($this->highlight_extra_lines) > 0)) {
										// strreplace to put close span and open span around multiline newlines
										$test_str .= str_replace("\n", "</span>\n<span$attributes>", $rest_of_comment);
									} else {
										$test_str .= $rest_of_comment;
									}

									if ($this->lexic_permissions['COMMENTS']['MULTI'] ||
                                        $test_str_match == GESHI_START_IMPORTANT) {
										$test_str .= '</span>';
									}
									$i = $close_pos + $com_len - 1;
									// parse the rest
									$result .= $this->parse_non_string_part($stuff_to_parse);
									$stuff_to_parse = '';
									break;
								}
							}
							// If we haven't matched a multiline comment, try single-line comments
							if (!$COMMENT_MATCHED) {
								foreach ($this->language_data['COMMENT_SINGLE'] as $comment_key => $comment_mark) {
									$com_len = strlen($comment_mark);
									$test_str = substr($part, $i, $com_len);
									if ($this->language_data['CASE_SENSITIVE'][GESHI_COMMENTS]) {
										$match = ($comment_mark == $test_str);
									} else {
										$match = (strtolower($comment_mark) == strtolower($test_str));
									}
									if ($match) {
										$COMMENT_MATCHED = true;
										if ($this->lexic_permissions['COMMENTS'][$comment_key]) {
											if (!$this->use_classes) {
												$attributes = ' style="' . $this->language_data['STYLES']['COMMENTS'][$comment_key] . '"';
											} else {
												$attributes = ' class="co' . $comment_key . '"';
											}
											$test_str = "<span$attributes>" . @htmlspecialchars($this->change_case($test_str), ENT_COMPAT, $this->encoding);
										} else {
											$test_str = @htmlspecialchars($test_str, ENT_COMPAT, $this->encoding);
										}
										$close_pos = strpos($part, "\n", $i);
                                        $oops = false;
										if ($close_pos === false) {
											$close_pos = strlen($part);
                                            $oops = true;
										}
										$test_str .= @htmlspecialchars(substr($part, $i + $com_len, $close_pos - $i - $com_len), ENT_COMPAT, $this->encoding);
										if ($this->lexic_permissions['COMMENTS'][$comment_key]) {
											$test_str .= "</span>";
										}
                                        // Take into account that the comment might be the last in the source
                                        if (!$oops) { 
										  $test_str .= "\n";
                                        }
										$i = $close_pos;
										// parse the rest
										$result .= $this->parse_non_string_part($stuff_to_parse);
										$stuff_to_parse = '';
										break;
									}
								}
							}
						} elseif ($STRING_OPEN != '') {
                            // Otherwise, convert it to HTML form
							if (strtolower($this->encoding) == 'utf-8') {
								//only escape <128 (we don't want to break multibyte chars)
								if (ord($char) < 128) {
									$char = @htmlspecialchars($char, ENT_COMPAT, $this->encoding);
								}
							} else {
								//encode everthing
								$char = @htmlspecialchars($char, ENT_COMPAT, $this->encoding);
							}
						}
						// Where are we adding this char?
						if (!$COMMENT_MATCHED) {
							if (($STRING_OPEN == '') && !$CLOSE_STRING) {
								$stuff_to_parse .= $char;
							} else {
								$result .= $char;
								$CLOSE_STRING = false;
							}
						} else {
							$result .= $test_str;
							$COMMENT_MATCHED = false;
						}
					}
					// Parse the last bit
					$result .= $this->parse_non_string_part($stuff_to_parse);
					$stuff_to_parse = '';
				} else {
					$result .= @htmlspecialchars($part, ENT_COMPAT, $this->encoding);
				}
				// Close the <span> that surrounds the block
				if ($this->strict_mode && $this->language_data['STYLES']['SCRIPT'][$script_key] != '' &&
                    $this->lexic_permissions['SCRIPT']) {
					$result .= '</span>';
				}
			} else {
                // Else not a block to highlight
				$result .= @htmlspecialchars($part, ENT_COMPAT, $this->encoding);
			}
		}

		// Parse the last stuff (redundant?)
		$result .= $this->parse_non_string_part($stuff_to_parse);

		// Lop off the very first and last spaces
		$result = substr($result, 1, -1);

		// Are we still in a string?
		if ($STRING_OPEN) {
			$result .= '</span>';
		}

		// We're finished: stop timing
		$this->set_time($start_time, microtime());

		return $this->finalise($result);
	}

	/**
	 * Swaps out spaces and tabs for HTML indentation. Not needed if
	 * the code is in a pre block...
     * 
     * @param  string The source to indent
     * @return string The source with HTML indenting applied
     * @since  1.0.0
     * @access private
	 */
	function indent ($result)
	{
            /// Replace tabs with the correct number of spaces
            if (false !== strpos($result, "\t")) {
                $lines = explode("\n", $result);
                foreach ($lines as $key => $line) {
                    if (false === strpos($line, "\t")) {
                        $lines[$key] = $line;
                        continue;
                    }
                    
                    $pos = 0;
                    $tab_width = $this->tab_width;
                    $length = strlen($line);
                    $result_line = '';

                    $IN_TAG = false;
                    for ($i = 0; $i < $length; $i++) {
                        $char = substr($line, $i, 1);
                        // Simple engine to work out whether we're in a tag.
                        // If we are we modify $pos. This is so we ignore HTML
                        // in the line and only workout the tab replacement
                        // via the actual content of the string
                        // This test could be improved to include strings in the
                        // html so that < or > would be allowed in user's styles
                        // (e.g. quotes: '<' '>'; or similar)
                        if ($IN_TAG && '>' == $char) {
                            $IN_TAG = false;
                            $result_line .= '>';
                            ++$pos;
                        } elseif (!$IN_TAG && '<' == $char) {
                            $IN_TAG = true;
                            $result_line .= '<';
                            ++$pos;
                        } elseif (!$IN_TAG && '&' == $char) {                            $substr = substr($line, $i + 3, 4);
                            //$substr_5 = substr($line, 5, 1);
                            $posi = strpos($substr, ';');
                            if (false !== $posi) {
                                $pos += $posi + 3;
                            }
                            $result_line .= '&';
                        } elseif (!$IN_TAG && "\t" == $char) {
                            $str = '';
                            // OPTIMISE - move $strs out. Make an array:
                            // $tabs = array(
                            //  1 => '&nbsp;',
                            //  2 => '&nbsp; ',
                            //  3 => '&nbsp; &nbsp;' etc etc
                            // to use instead of building a string every time
                            $strs = array(0 => '&nbsp;', 1 => ' ');
                            for ($k = 0; $k < ($tab_width - (($i - $pos) % $tab_width)); $k++) $str .= $strs[$k % 2];
                            $result_line .= $str;
                            //$pos--;
                            $pos++;
                            //$pos -= $tab_width-1;

                            if (false === strpos($line, "\t", $i + 1)) {
                                //$lines[$key] = $result_line;
                                $result_line .= substr($line, $i + 1);
                                break;
                            }
                        } elseif ( $IN_TAG ) {
                            ++$pos;
                            $result_line .= $char;
                        } else {
                            $result_line .= $char;
                            //++$pos;
                        }
                    }
                    $lines[$key] = $result_line;
                }
                $result = implode("\n", $lines);
            }
		// Other whitespace
		$result = str_replace('  ', '&nbsp; ', $result);
		$result = str_replace('  ', ' &nbsp;', $result);
		$result = str_replace("\n ", "\n&nbsp;", $result);

		if ($this->line_numbers == GESHI_NO_LINE_NUMBERS) {
			$result = nl2br($result);
		}
		return $result;
	}

	/**
	 * Changes the case of a keyword for those languages where a change is asked for
     * 
     * @param  string The keyword to change the case of
     * @return string The keyword with its case changed
     * @since  1.0.0
     * @access private
	 */
	function change_case ($instr)
	{
		if ($this->language_data['CASE_KEYWORDS'] == GESHI_CAPS_UPPER) {
			return strtoupper($instr);
		} elseif ($this->language_data['CASE_KEYWORDS'] == GESHI_CAPS_LOWER) {
			return strtolower($instr);
		}
		return $instr;
	}

	/**
	 * Adds a url to a keyword where needed.
	 * 
     * @param  string The keyword to add the URL HTML to
     * @param  int What group the keyword is from
     * @param  boolean Whether to get the HTML for the start or end
     * @return The HTML for either the start or end of the HTML &lt;a&gt; tag
     * @since  1.0.2
     * @access private
     * @todo   Get rid of ender
	 */
	function add_url_to_keyword ($keyword, $group, $start_or_end)
	{
        if (isset($this->language_data['URLS'][$group]) &&
            $this->language_data['URLS'][$group] != '' &&
            substr($keyword, 0, 5) != '&lt;/') {
			// There is a base group for this keyword
			if ($start_or_end == 'BEGIN') {
				// HTML workaround... not good form (tm) but should work for 1.0.X
				if ($keyword != '') {
                    // Old system: strtolower
					//$keyword = ( $this->language_data['CASE_SENSITIVE'][$group] ) ? $keyword : strtolower($keyword);
                    // New system: get keyword from language file to get correct case
                    foreach ($this->language_data['KEYWORDS'][$group] as $word) {
                        if (strtolower($word) == strtolower($keyword)) {
                            break;
                        }
                    }
                    $word = ( substr($word, 0, 4) == '&lt;' ) ? substr($word, 4) : $word;
                    $word = ( substr($word, -4) == '&gt;' ) ? substr($word, 0, strlen($word) - 4) : $word;
                    if (!$word) return '';
                    
					return '<|UR1|"' .
                        str_replace(
                            array('{FNAME}', '.'),
                            array(@htmlspecialchars($word, ENT_COMPAT, $this->encoding), '<DOT>'),
                            $this->language_data['URLS'][$group]
                        ) . '">';
				}
				return '';
            // HTML fix. Again, dirty hackage...
			} elseif (!($this->language == 'html4strict' && '&gt;' == $keyword)) {
				return '</a>';
			}
		}
	}

	/**
	 * 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.
     * @since 1.0.0
     * @access private
     * @todo BUGGY! Why? Why not build string and return?
	 */
	function parse_non_string_part (&$stuff_to_parse)
	{
		$stuff_to_parse = ' ' . @htmlspecialchars($stuff_to_parse, ENT_COMPAT, $this->encoding);
        $stuff_to_parse_pregquote = preg_quote($stuff_to_parse, '/');
		$func = '$this->change_case';
		$func2 = '$this->add_url_to_keyword';

		//
		// Regular expressions
		//
		foreach ($this->language_data['REGEXPS'] as $key => $regexp) {
			if ($this->lexic_permissions['REGEXPS'][$key]) {
				if (is_array($regexp)) {
					$stuff_to_parse = preg_replace(
                        "/" .
                        str_replace('/', '\/', $regexp[GESHI_SEARCH]) .
                        "/{$regexp[GESHI_MODIFIERS]}",
                        "{$regexp[GESHI_BEFORE]}<|!REG3XP$key!>{$regexp[GESHI_REPLACE]}|>{$regexp[GESHI_AFTER]}",
                        $stuff_to_parse
                    );
				} else {
					$stuff_to_parse = preg_replace( "/(" . str_replace('/', '\/', $regexp) . ")/", "<|!REG3XP$key!>\\1|>", $stuff_to_parse);
				}