2006-04-15 01:15:56 +00:00
|
|
|
<?php
|
|
|
|
|
|
|
|
/*
|
2006-04-15 02:19:27 +00:00
|
|
|
Forgivingly lexes SGML style documents, aka HTML, XML, XHMTML, you name it.
|
2006-04-15 01:15:56 +00:00
|
|
|
|
|
|
|
TODO:
|
|
|
|
* Reread the XML spec and make sure I got everything right
|
2006-04-15 02:19:27 +00:00
|
|
|
* Add support for CDATA sections
|
|
|
|
* Have comments output with the leading and trailing --s
|
|
|
|
* Optimize and benchmark
|
|
|
|
* Check MF_Text behavior: shouldn't the info in there be raw (entities parsed?)
|
2006-04-15 01:15:56 +00:00
|
|
|
|
|
|
|
*/
|
|
|
|
|
2006-04-15 01:47:12 +00:00
|
|
|
class HTML_Lexer
|
2006-04-15 01:15:56 +00:00
|
|
|
{
|
|
|
|
|
|
|
|
function nextQuote($string, $offset = 0) {
|
|
|
|
$quotes = array('"', "'");
|
|
|
|
return $this->next($string, $quotes, $offset);
|
|
|
|
}
|
|
|
|
|
|
|
|
function nextWhiteSpace($string, $offset = 0) {
|
|
|
|
$spaces = array(chr(0x20), chr(0x9), chr(0xD), chr(0xA));
|
|
|
|
return $this->next($string, $spaces, $offset);
|
|
|
|
}
|
|
|
|
|
|
|
|
function next($haystack, $needles, $offset = 0) {
|
|
|
|
if (is_string($needles)) {
|
|
|
|
$string_needles = $needles;
|
|
|
|
$needles = array();
|
|
|
|
$size = strlen($string_needles);
|
|
|
|
for ($i = 0; $i < $size; $i++) {
|
|
|
|
$needles[] = $string_needles{$i};
|
|
|
|
}
|
|
|
|
}
|
|
|
|
$positions = array();
|
|
|
|
foreach ($needles as $needle) {
|
|
|
|
$position = strpos($haystack, $needle, $offset);
|
|
|
|
if ($position !== false) {
|
|
|
|
$positions[] = $position;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return empty($positions) ? false : min($positions);
|
|
|
|
}
|
|
|
|
|
|
|
|
function tokenizeHTML($string) {
|
|
|
|
|
|
|
|
// some quick checking (if empty, return empty)
|
|
|
|
$string = (string) $string;
|
|
|
|
if ($string == '') return array();
|
|
|
|
|
|
|
|
$cursor = 0; // our location in the text
|
|
|
|
$inside_tag = false; // whether or not we're parsing the inside of a tag
|
|
|
|
$array = array(); // result array
|
|
|
|
|
|
|
|
while(true) {
|
|
|
|
|
|
|
|
$position_next_lt = strpos($string, '<', $cursor);
|
|
|
|
$position_next_gt = strpos($string, '>', $cursor);
|
|
|
|
|
|
|
|
// triggers on "<b>asdf</b>" but not "asdf <b></b>"
|
|
|
|
if ($position_next_lt === $cursor) {
|
|
|
|
$inside_tag = true;
|
|
|
|
$cursor++;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!$inside_tag && $position_next_lt !== false) {
|
|
|
|
// We are not inside tag and there still is another tag to parse
|
2006-04-15 01:30:26 +00:00
|
|
|
$array[] = new MF_Text(substr($string, $cursor, $position_next_lt - $cursor));
|
2006-04-15 01:15:56 +00:00
|
|
|
$cursor = $position_next_lt + 1;
|
|
|
|
$inside_tag = true;
|
|
|
|
continue;
|
|
|
|
} elseif (!$inside_tag) {
|
|
|
|
// We are not inside tag but there are no more tags
|
|
|
|
// If we're already at the end, break
|
|
|
|
if ($cursor === strlen($string)) break;
|
|
|
|
// Create Text of rest of string
|
2006-04-15 01:30:26 +00:00
|
|
|
$array[] = new MF_Text(substr($string, $cursor));
|
2006-04-15 01:15:56 +00:00
|
|
|
break;
|
|
|
|
} elseif ($inside_tag && $position_next_gt !== false) {
|
|
|
|
// We are in tag and it is well formed
|
|
|
|
// Grab the internals of the tag
|
|
|
|
$segment = substr($string, $cursor, $position_next_gt - $cursor);
|
|
|
|
|
|
|
|
// Check if it's a comment
|
|
|
|
if (substr($segment,0,3) == '!--' && substr($segment,strlen($segment)-2,2) == '--') {
|
2006-04-15 01:30:26 +00:00
|
|
|
$array[] = new MF_Comment(substr($segment,3,strlen($segment)-5));
|
2006-04-15 01:15:56 +00:00
|
|
|
$inside_tag = false;
|
|
|
|
$cursor = $position_next_gt + 1;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Check if it's an end tag
|
|
|
|
$is_end_tag = (strpos($segment,'/') === 0);
|
|
|
|
if ($is_end_tag) {
|
|
|
|
$type = substr($segment, 1);
|
2006-04-15 01:30:26 +00:00
|
|
|
$array[] = new MF_EndTag($type);
|
2006-04-15 01:15:56 +00:00
|
|
|
$inside_tag = false;
|
|
|
|
$cursor = $position_next_gt + 1;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
2006-04-15 01:47:12 +00:00
|
|
|
// Check if it is explicitly self closing, if so, remove
|
|
|
|
// trailing slash. Remember, we could have a tag like <br>, so
|
|
|
|
// any later token processing scripts must convert improperly
|
|
|
|
// classified EmptyTags from StartTags.
|
2006-04-15 01:15:56 +00:00
|
|
|
$is_self_closing = (strpos($segment,'/') === strlen($segment) - 1);
|
|
|
|
if ($is_self_closing) {
|
|
|
|
$segment = substr($segment, 0, strlen($segment) - 1);
|
|
|
|
}
|
|
|
|
|
|
|
|
// Check if there are any attributes
|
|
|
|
$position_first_space = $this->nextWhiteSpace($segment);
|
|
|
|
if ($position_first_space === false) {
|
|
|
|
if ($is_self_closing) {
|
2006-04-15 01:30:26 +00:00
|
|
|
$array[] = new MF_EmptyTag($segment);
|
2006-04-15 01:15:56 +00:00
|
|
|
} else {
|
2006-04-15 01:30:26 +00:00
|
|
|
$array[] = new MF_StartTag($segment, array());
|
2006-04-15 01:15:56 +00:00
|
|
|
}
|
|
|
|
$inside_tag = false;
|
|
|
|
$cursor = $position_next_gt + 1;
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
// Grab out all the data
|
|
|
|
$type = substr($segment, 0, $position_first_space);
|
|
|
|
$attribute_string = trim(substr($segment, $position_first_space));
|
|
|
|
$attributes = $this->tokenizeAttributeString($attribute_string);
|
|
|
|
if ($is_self_closing) {
|
2006-04-15 01:30:26 +00:00
|
|
|
$array[] = new MF_EmptyTag($type, $attributes);
|
2006-04-15 01:15:56 +00:00
|
|
|
} else {
|
2006-04-15 01:30:26 +00:00
|
|
|
$array[] = new MF_StartTag($type, $attributes);
|
2006-04-15 01:15:56 +00:00
|
|
|
}
|
|
|
|
$cursor = $position_next_gt + 1;
|
|
|
|
$inside_tag = false;
|
|
|
|
continue;
|
|
|
|
} else {
|
2006-04-15 01:30:26 +00:00
|
|
|
$array[] = new MF_Text('<' . substr($string, $cursor));
|
2006-04-15 01:15:56 +00:00
|
|
|
break;
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return $array;
|
|
|
|
}
|
|
|
|
|
|
|
|
function tokenizeAttributeString($string) {
|
|
|
|
$string = (string) $string;
|
|
|
|
if ($string == '') return array();
|
|
|
|
|
|
|
|
$array = array();
|
|
|
|
$cursor = 0;
|
|
|
|
$in_value = false;
|
|
|
|
$i = 0;
|
|
|
|
$size = strlen($string);
|
|
|
|
while(true) {
|
|
|
|
if ($cursor >= $size) {
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
$position_next_space = $this->nextWhiteSpace($string, $cursor);
|
|
|
|
//scroll to the last whitespace before text
|
|
|
|
while ($position_next_space === $cursor) {
|
|
|
|
$cursor++;
|
|
|
|
$position_next_space = $this->nextWhiteSpace($string, $cursor);
|
|
|
|
}
|
|
|
|
$position_next_equal = strpos($string, '=', $cursor);
|
|
|
|
if ($position_next_equal !== false &&
|
|
|
|
($position_next_equal < $position_next_space ||
|
|
|
|
$position_next_space === false)) {
|
|
|
|
//attr="asdf"
|
|
|
|
$key = trim(substr($string, $cursor, $position_next_equal - $cursor));
|
|
|
|
$position_next_quote = $this->nextQuote($string, $cursor);
|
|
|
|
$quote = $string{$position_next_quote};
|
|
|
|
$position_end_quote = strpos($string, $quote, $position_next_quote + 1);
|
|
|
|
$value = substr($string, $position_next_quote + 1,
|
|
|
|
$position_end_quote - $position_next_quote - 1);
|
|
|
|
if ($key) {
|
|
|
|
$array[$key] = $value;
|
|
|
|
}
|
|
|
|
$cursor = $position_end_quote + 1;
|
|
|
|
} else {
|
|
|
|
//boolattr
|
|
|
|
if ($position_next_space === false) {
|
|
|
|
$position_next_space = $size;
|
|
|
|
}
|
|
|
|
$key = substr($string, $cursor, $position_next_space - $cursor);
|
|
|
|
if ($key) {
|
|
|
|
$array[$key] = $key;
|
|
|
|
}
|
|
|
|
$cursor = $position_next_space + 1;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
return $array;
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2006-04-15 01:47:12 +00:00
|
|
|
// uses the PEAR class XML_HTMLSax3 to parse XML
|
2006-04-15 02:19:27 +00:00
|
|
|
// only shares the tokenizeHTML() function
|
2006-04-15 01:47:12 +00:00
|
|
|
class HTML_Lexer_Sax extends HTML_Lexer
|
|
|
|
{
|
|
|
|
|
|
|
|
var $tokens = array();
|
|
|
|
|
|
|
|
function tokenizeHTML($html) {
|
|
|
|
$this->tokens = array();
|
|
|
|
$parser=& new XML_HTMLSax3();
|
|
|
|
$parser->set_object($this);
|
|
|
|
$parser->set_element_handler('openHandler','closeHandler');
|
|
|
|
$parser->set_data_handler('dataHandler');
|
|
|
|
$parser->set_escape_handler('escapeHandler');
|
|
|
|
$parser->parse($html);
|
|
|
|
return $this->tokens;
|
|
|
|
}
|
|
|
|
|
2006-04-15 02:19:27 +00:00
|
|
|
function openHandler(&$parser, $name, $attrs, $closed) {
|
|
|
|
if ($closed) {
|
|
|
|
$this->tokens[] = new MF_EmptyTag($name, $attrs);
|
|
|
|
} else {
|
|
|
|
$this->tokens[] = new MF_StartTag($name, $attrs);
|
|
|
|
}
|
2006-04-15 01:47:12 +00:00
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
function closeHandler(&$parser, $name) {
|
2006-04-15 02:19:27 +00:00
|
|
|
// HTMLSax3 seems to always send empty tags an extra close tag
|
|
|
|
// check and ignore if you see it:
|
|
|
|
// [TESTME] to make sure it doesn't overreach
|
|
|
|
if (is_a($this->tokens[count($this->tokens)-1], 'MF_EmptyTag')) {
|
|
|
|
return true;
|
|
|
|
}
|
2006-04-15 01:47:12 +00:00
|
|
|
$this->tokens[] = new MF_EndTag($name);
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
function dataHandler(&$parser, $data) {
|
|
|
|
$this->tokens[] = new MF_Text($data);
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
function escapeHandler(&$parser, $data) {
|
|
|
|
if (strpos($data, '-') === 0) {
|
|
|
|
$this->tokens[] = new MF_Comment($data);
|
|
|
|
}
|
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|
|
|
|
|
2006-04-15 01:15:56 +00:00
|
|
|
?>
|