mirror of
https://github.com/gtbu/Typesetter-5.3-p8.git
synced 2025-06-22 08:43:13 +02:00
improved Htmlparser
This commit is contained in:
gtbu
parent
796a988e9a
commit
2c712361a6
1 changed files with 272 additions and 326 deletions
|
|
@ -1,330 +1,276 @@
|
|||
<?php
|
||||
|
||||
declare(strict_types=1);
|
||||
namespace gp\tool\Editing;
|
||||
|
||||
defined('is_running') or die('Not an entry point...');
|
||||
|
||||
|
||||
class HTMLParse{
|
||||
|
||||
public $doc = '';
|
||||
public $dom_array = array();
|
||||
public $errors = array();
|
||||
|
||||
public $random;
|
||||
public $mark_double_slash;
|
||||
public $mark_escaped_single;
|
||||
public $mark_escaped_double;
|
||||
|
||||
public function __construct($text){
|
||||
$this->doc = $text;
|
||||
$this->Init_Parse();
|
||||
$this->Parse();
|
||||
}
|
||||
|
||||
|
||||
public function Parse(){
|
||||
|
||||
$offset = 0;
|
||||
|
||||
do{
|
||||
$continue = true;
|
||||
$pos = strpos($this->doc,'<',$offset);
|
||||
|
||||
|
||||
//no more tags
|
||||
if( $pos === false ){
|
||||
$continue = false;
|
||||
break;
|
||||
}
|
||||
|
||||
//comment
|
||||
if( substr($this->doc,$pos,4) === '<!--' ){
|
||||
$this->dom_array[] = substr($this->doc,$offset,$pos-$offset); //get content before comment
|
||||
$offset = $pos + 4;
|
||||
$this->CommentContent($offset);
|
||||
continue;
|
||||
}
|
||||
|
||||
//get tag name
|
||||
$tag_name = $this->TagName($pos+1,$name_len);
|
||||
if( !$tag_name ){
|
||||
$this->doc = substr_replace($this->doc,'<',$pos,1);
|
||||
continue;
|
||||
}
|
||||
|
||||
//content
|
||||
if( $pos > $offset ){
|
||||
$this->dom_array[] = substr($this->doc,$offset,$pos-$offset);
|
||||
}
|
||||
|
||||
|
||||
//advance offset to just after <tag_name
|
||||
$offset = $pos+1+$name_len;
|
||||
$new_element = array();
|
||||
$new_element['tag'] = $tag_name;
|
||||
|
||||
//attributes
|
||||
if( $tag_name[0] != '/' ){
|
||||
$new_element['attributes'] = $this->GetAttributes($offset);
|
||||
}
|
||||
|
||||
//tag closing and self-closing?
|
||||
$pos = strpos($this->doc,'>',$offset);
|
||||
if( $pos !== false ){
|
||||
if( $pos > 0 && $this->doc[$pos-1] == '/' ){
|
||||
$new_element['self_closing'] = true;
|
||||
}
|
||||
$offset = $pos+1;
|
||||
}
|
||||
|
||||
$this->dom_array[] = $new_element;
|
||||
|
||||
|
||||
//content
|
||||
switch(strtolower($tag_name)){
|
||||
case 'script':
|
||||
case 'style';
|
||||
$this->NonHtmlContent($offset,$tag_name);
|
||||
break;
|
||||
}
|
||||
|
||||
}while($continue);
|
||||
|
||||
//content after
|
||||
if( $offset < strlen($this->doc) ){
|
||||
$this->dom_array[] = substr($this->doc,$offset);
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Parse an html tag name
|
||||
*
|
||||
*/
|
||||
public function TagName($pos,&$name_len){
|
||||
$tag_name = false;
|
||||
$name_len = strspn($this->doc,'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890/!:--',$pos);
|
||||
if( $name_len > 0 ){
|
||||
$tag_name = substr($this->doc,$pos,$name_len);
|
||||
}
|
||||
return $tag_name;
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Parse html attributes
|
||||
* match name="value", name=value or name
|
||||
* accounts for name="value ' value"
|
||||
*
|
||||
*/
|
||||
public function GetAttributes(&$offset){
|
||||
|
||||
$this->doc = substr($this->doc,$offset);
|
||||
$pattern_name = '#^\s+([^\'"<>=\s/]+)\s*(=?)\s*#';
|
||||
$attributes = array();
|
||||
|
||||
|
||||
while( preg_match($pattern_name, $this->doc, $matches) ){
|
||||
|
||||
$attr_name = $matches[1];
|
||||
$attr_value = null;
|
||||
$offset = strlen($matches[0]);
|
||||
|
||||
//get attribute value
|
||||
if( !empty($matches[2]) ){
|
||||
|
||||
$attr_match = $this->MatchAttribute($offset);
|
||||
if( $attr_match ){
|
||||
$offset += strlen($attr_match[0]);
|
||||
$attr_value = $attr_match[1];
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
if( !isset($attributes[$attr_name]) ){
|
||||
$attributes[$attr_name] = $attr_value;
|
||||
}
|
||||
|
||||
$this->doc = substr($this->doc,$offset);
|
||||
$offset = 0;
|
||||
}
|
||||
|
||||
|
||||
$offset = 0;
|
||||
|
||||
return $attributes;
|
||||
}
|
||||
|
||||
/**
|
||||
* Get an html attribute value
|
||||
*
|
||||
*/
|
||||
protected function MatchAttribute(&$offset){
|
||||
|
||||
$char = $this->doc[$offset];
|
||||
|
||||
//double quote
|
||||
if( $char === '"' ){
|
||||
if( preg_match('#\\G"([^"]*)"#', $this->doc, $matches, 0, $offset) ){
|
||||
return $matches;
|
||||
}
|
||||
return;
|
||||
}
|
||||
|
||||
//single quote
|
||||
if( $char == "'" ){
|
||||
if( preg_match('#\\G\'([^\']*)\'#', $this->doc, $matches, 0, $offset) ){
|
||||
return $matches;
|
||||
}
|
||||
return;
|
||||
}
|
||||
|
||||
|
||||
|
||||
//not quoted
|
||||
if( preg_match('#\\G\s*([^\'"<>=\s/]+)#', $this->doc, $matches, 0, $offset) ){
|
||||
return $matches;
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
|
||||
/**
|
||||
* Parse HTML comments <!-- comments go here -->
|
||||
* Does not support full sgml comments <!------> second comment -->
|
||||
*
|
||||
*/
|
||||
public function CommentContent(&$offset){
|
||||
|
||||
$this->doc = substr($this->doc,$offset);
|
||||
$offset = 0;
|
||||
|
||||
$pos = strpos($this->doc,'-->');
|
||||
if( $pos === false ){
|
||||
$pos = strlen($this->doc);
|
||||
}
|
||||
|
||||
$comment_content = substr($this->doc,0,$pos);
|
||||
$this->doc = substr($this->doc,$pos+3);
|
||||
|
||||
$new_element = array();
|
||||
$new_element['comment'] = $comment_content;
|
||||
$this->dom_array[] = $new_element;
|
||||
}
|
||||
|
||||
public function NonHtmlContent(&$offset,$untill){
|
||||
|
||||
$this->doc = substr($this->doc,$offset);
|
||||
$offset = 0;
|
||||
$this->doc = $this->EscapeQuotes($this->doc);
|
||||
$full_length = strlen($this->doc);
|
||||
$untill_length = strlen($untill);
|
||||
|
||||
do{
|
||||
|
||||
$continue = false;
|
||||
$end_string = false;
|
||||
|
||||
$pos_quote1 = $this->strpos_min("'",$offset,$full_length);
|
||||
$pos_quote2 = $this->strpos_min('"',$offset,$full_length);
|
||||
$pos_scomment = $this->strpos_min('//',$offset,$full_length);
|
||||
$pos_mcomment = $this->strpos_min('/*',$offset,$full_length);
|
||||
|
||||
$min_pos = min($pos_quote1, $pos_quote2, $pos_scomment, $pos_mcomment);
|
||||
|
||||
$pos_close = strpos($this->doc,'</',$offset);
|
||||
|
||||
// found </script>
|
||||
if( ($pos_close !== false)
|
||||
&& ($pos_close <= $min_pos)
|
||||
&& (strtolower(substr($this->doc,$pos_close+2,$untill_length)) == $untill)
|
||||
){
|
||||
$offset = $pos_close;
|
||||
break;
|
||||
}
|
||||
|
||||
// nothing else found
|
||||
if( $min_pos === $full_length ){
|
||||
$offset = $full_length;
|
||||
break;
|
||||
}
|
||||
|
||||
|
||||
if( $min_pos === $pos_quote1 ){
|
||||
$end_string = "'";
|
||||
}elseif( $min_pos === $pos_quote2 ){
|
||||
$end_string = '"';
|
||||
}elseif( $min_pos === $pos_scomment ){
|
||||
$end_string = "\n";
|
||||
}elseif( $min_pos === $pos_mcomment ){
|
||||
$end_string = '*/';
|
||||
}
|
||||
|
||||
$end_pos = strpos($this->doc,$end_string,$min_pos+1);
|
||||
if( $end_pos === false ){
|
||||
$offset = $full_length;
|
||||
}else{
|
||||
$offset = $full_length;
|
||||
$offset = $end_pos + strlen($end_string);
|
||||
$continue = true;
|
||||
}
|
||||
|
||||
|
||||
}while($continue);
|
||||
|
||||
$code = substr($this->doc,0,$offset);
|
||||
$this->doc = substr($this->doc,$offset);
|
||||
$this->doc = $this->UnescapeQuotes($this->doc);
|
||||
$this->dom_array[] = $this->UnescapeQuotes($code);
|
||||
$offset = 0;
|
||||
}
|
||||
|
||||
|
||||
public function strpos_min($needle,$offset,$length){
|
||||
$pos = strpos($this->doc,$needle,$offset);
|
||||
if( $pos === false ){
|
||||
return $length;
|
||||
}
|
||||
return $pos;
|
||||
}
|
||||
|
||||
|
||||
public function EscapeQuotes($string){
|
||||
|
||||
$search = array('\\\\','\\\'','\\"');
|
||||
$replace = array( $this->mark_double_slash, $this->mark_escaped_single, $this->mark_escaped_double);
|
||||
|
||||
return str_replace($search, $replace, $string);
|
||||
}
|
||||
|
||||
public function UnescapeQuotes($string){
|
||||
$search = array( $this->mark_double_slash, $this->mark_escaped_single, $this->mark_escaped_double);
|
||||
$replace = array('\\\\','\\\'','\\"');
|
||||
return str_replace($search, $replace, $string);
|
||||
}
|
||||
|
||||
/*
|
||||
* Init
|
||||
*
|
||||
*/
|
||||
public function Init_Parse(){
|
||||
$this->GetRandom();
|
||||
$this->mark_double_slash = $this->GetMarker();
|
||||
$this->mark_escaped_single = $this->GetMarker();
|
||||
$this->mark_escaped_double = $this->GetMarker();
|
||||
}
|
||||
|
||||
|
||||
public function GetRandom(){
|
||||
do{
|
||||
$this->random = dechex(mt_rand(0, 0x7fffff));
|
||||
}while(strpos($this->doc,$this->random) !== false);
|
||||
}
|
||||
|
||||
public function GetMarker(){
|
||||
static $n = 0;
|
||||
return $this->random . sprintf('%08X', $n++);
|
||||
}
|
||||
}
|
||||
/**
|
||||
* A custom, non-validating HTML parser that converts an HTML string into an array structure.
|
||||
* It's designed to be fast and handle real-world, often imperfect, HTML.
|
||||
*/
|
||||
class HTMLParse
|
||||
{
|
||||
public string $doc = '';
|
||||
public array $dom_array = [];
|
||||
public array $errors = [];
|
||||
|
||||
private int $doc_length;
|
||||
private int $position = 0;
|
||||
|
||||
private string $mark_double_slash;
|
||||
private string $mark_escaped_single;
|
||||
private string $mark_escaped_double;
|
||||
|
||||
public function __construct(string $text)
|
||||
{
|
||||
$this->doc = $text;
|
||||
$this->doc_length = strlen($text);
|
||||
$this->Init_Parse();
|
||||
$this->Parse();
|
||||
}
|
||||
|
||||
public function Init_Parse(): void
|
||||
{
|
||||
$this->generateMarkers();
|
||||
}
|
||||
|
||||
private function generateMarkers(): void
|
||||
{
|
||||
$this->mark_double_slash = $this->uniqueMarker();
|
||||
$this->mark_escaped_single = $this->uniqueMarker();
|
||||
$this->mark_escaped_double = $this->uniqueMarker();
|
||||
}
|
||||
|
||||
private function uniqueMarker(): string
|
||||
{
|
||||
static $counter = 0;
|
||||
return "\x01".hash('xxh3', microtime().$counter++)."\x02";
|
||||
}
|
||||
|
||||
private function addError(string $message): void
|
||||
{
|
||||
$this->errors[] = "Error at position {$this->position}: {$message}";
|
||||
}
|
||||
|
||||
public function Parse(): void
|
||||
{
|
||||
while ($this->position < $this->doc_length) {
|
||||
$char = $this->doc[$this->position];
|
||||
|
||||
if ($char !== '<') {
|
||||
$this->parseTextContent();
|
||||
continue;
|
||||
}
|
||||
|
||||
if ($this->handleCommentIfAny()) {
|
||||
continue;
|
||||
}
|
||||
|
||||
$tag_info = $this->parseTag();
|
||||
if ($tag_info === null) {
|
||||
// If parseTag fails, treat the '<' as literal text
|
||||
$this->dom_array[] = '<';
|
||||
$this->position++;
|
||||
continue;
|
||||
}
|
||||
|
||||
// If it's an opening tag for a special content element...
|
||||
if ($tag_info['name'][0] !== '/' && !$tag_info['self_closing']) {
|
||||
$this->handleSpecialContent($tag_info['name']);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private function parseTextContent(): void
|
||||
{
|
||||
$next_tag_pos = strpos($this->doc, '<', $this->position);
|
||||
if ($next_tag_pos === false) {
|
||||
$text = substr($this->doc, $this->position);
|
||||
$this->position = $this->doc_length;
|
||||
} else {
|
||||
$text = substr($this->doc, $this->position, $next_tag_pos - $this->position);
|
||||
$this->position = $next_tag_pos;
|
||||
}
|
||||
|
||||
if ($text !== '') {
|
||||
$this->dom_array[] = $text;
|
||||
}
|
||||
}
|
||||
|
||||
private function handleCommentIfAny(): bool
|
||||
{
|
||||
if (substr_compare($this->doc, '<!--', $this->position, 4) === 0) {
|
||||
$end_pos = strpos($this->doc, '-->', $this->position + 4);
|
||||
if ($end_pos === false) {
|
||||
$content = substr($this->doc, $this->position + 4);
|
||||
$this->position = $this->doc_length;
|
||||
$this->addError("Unclosed HTML comment.");
|
||||
} else {
|
||||
$content = substr($this->doc, $this->position + 4, $end_pos - ($this->position + 4));
|
||||
$this->position = $end_pos + 3;
|
||||
}
|
||||
$this->dom_array[] = ['comment' => $content];
|
||||
return true;
|
||||
}
|
||||
return false;
|
||||
}
|
||||
|
||||
/** @return ?array{name: string, self_closing: bool} */
|
||||
private function parseTag(): ?array
|
||||
{
|
||||
$original_tag_start_pos = $this->position;
|
||||
$this->position++; // Skip '<'
|
||||
|
||||
if ($this->position >= $this->doc_length) {
|
||||
$this->position = $original_tag_start_pos; // backtrack
|
||||
return null;
|
||||
}
|
||||
|
||||
$is_closing_tag_char = ($this->doc[$this->position] === '/');
|
||||
if ($is_closing_tag_char) {
|
||||
$this->position++; // Skip '/' for tag name parsing
|
||||
}
|
||||
|
||||
$tag_name = $this->parseTagName();
|
||||
|
||||
if ($tag_name === null || $tag_name === '') {
|
||||
$this->position = $original_tag_start_pos; // backtrack
|
||||
return null;
|
||||
}
|
||||
|
||||
$element = ['tag' => $tag_name];
|
||||
$self_closing = false;
|
||||
|
||||
if ($is_closing_tag_char) {
|
||||
$element['tag'] = '/' . $tag_name;
|
||||
} else { // Only parse attributes for opening tags
|
||||
$element['attributes'] = $this->parseAttributes();
|
||||
}
|
||||
|
||||
// Find the end of the tag
|
||||
$gt_pos = strpos($this->doc, '>', $this->position);
|
||||
if ($gt_pos === false) {
|
||||
$this->addError("Unclosed tag '{$element['tag']}'.");
|
||||
$this->position = $this->doc_length; // Consume rest of document
|
||||
$element['self_closing'] = false;
|
||||
$this->dom_array[] = $element;
|
||||
return ['name' => $element['tag'], 'self_closing' => false];
|
||||
}
|
||||
|
||||
// Check for XML-style self-closing tags like <br />
|
||||
if (!$is_closing_tag_char) {
|
||||
$before_gt_segment = substr($this->doc, $this->position, $gt_pos - $this->position);
|
||||
$trimmed_before_gt = rtrim($before_gt_segment);
|
||||
|
||||
// CHANGE 3: Use substr() for a cleaner, more modern check.
|
||||
if (substr($trimmed_before_gt, -1) === '/') {
|
||||
$self_closing = true;
|
||||
}
|
||||
}
|
||||
|
||||
$element['self_closing'] = $self_closing;
|
||||
$this->dom_array[] = $element;
|
||||
$this->position = $gt_pos + 1;
|
||||
|
||||
return ['name' => $element['tag'], 'self_closing' => $self_closing];
|
||||
}
|
||||
|
||||
private function parseTagName(): ?string
|
||||
{
|
||||
$name_len = strspn(
|
||||
$this->doc,
|
||||
'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789_:.-',
|
||||
$this->position
|
||||
);
|
||||
|
||||
if ($name_len === 0) {
|
||||
return null;
|
||||
}
|
||||
|
||||
$name = substr($this->doc, $this->position, $name_len);
|
||||
$this->position += $name_len;
|
||||
|
||||
return strtolower($name);
|
||||
}
|
||||
|
||||
/** @return array<string, string|null> */
|
||||
private function parseAttributes(): array
|
||||
{
|
||||
$attributes = [];
|
||||
// This regex finds attributes one by one, from the current position.
|
||||
$pattern = '/
|
||||
\G # Anchor to the current position in the string
|
||||
\s+ # Require at least one space before an attribute
|
||||
(?!/?>) # Negative lookahead: ensure we are not at the end of the tag (/> or >)
|
||||
([^\s=<>\/]+) # Capture group 1: The attribute name
|
||||
(?: # Optional group for the value part
|
||||
\s*=\s* # The equals sign, with optional whitespace
|
||||
(?:
|
||||
"([^"]*)" # Capture group 2: Double-quoted value
|
||||
| # OR
|
||||
\'([^\']*)\' # Capture group 3: Single-quoted value
|
||||
| # OR
|
||||
([^\s"\'=<>`]+) # Capture group 4: Unquoted value
|
||||
)
|
||||
)? # The entire value part is optional (for boolean attributes)
|
||||
/ix'; // Case-insensitive and extended mode
|
||||
|
||||
while (preg_match($pattern, $this->doc, $matches, PREG_OFFSET_CAPTURE, $this->position)) {
|
||||
$name = strtolower($matches[1][0]);
|
||||
|
||||
$value = $matches[2][0] ?? $matches[3][0] ?? $matches[4][0] ?? null;
|
||||
|
||||
if (!isset($attributes[$name])) {
|
||||
$attributes[$name] = $value !== null ?
|
||||
htmlspecialchars_decode($value, ENT_QUOTES) :
|
||||
null; // Store null for boolean attributes like 'disabled'
|
||||
}
|
||||
|
||||
$this->position = $matches[0][1] + strlen($matches[0][0]);
|
||||
}
|
||||
return $attributes;
|
||||
}
|
||||
|
||||
private function handleSpecialContent(string $tag_name_from_parser): void
|
||||
{
|
||||
if (!in_array($tag_name_from_parser, ['script', 'style'])) {
|
||||
return;
|
||||
}
|
||||
|
||||
$content_start_pos = $this->position;
|
||||
// Use the already-lowercased tag name
|
||||
$end_tag_to_find = "</{$tag_name_from_parser}>";
|
||||
|
||||
$remaining_doc_part = substr($this->doc, $content_start_pos);
|
||||
if ($remaining_doc_part === false || $remaining_doc_part === '') {
|
||||
$this->addError("Unclosed special tag '<{$tag_name_from_parser}>'.");
|
||||
return;
|
||||
}
|
||||
|
||||
// IMPORTANT: This logic correctly handles cases like `var x = "</script>";` inside a script tag.
|
||||
// It temporarily escapes certain sequences to prevent a premature match.
|
||||
$escaped_remaining_part = str_replace(
|
||||
['\\\\', '\\\'', '\\"', '</'],
|
||||
[$this->mark_double_slash, $this->mark_escaped_single, $this->mark_escaped_double, "<\\/"],
|
||||
$remaining_doc_part
|
||||
);
|
||||
|
||||
$end_tag_pos_in_escaped_part = stripos($escaped_remaining_part, $end_tag_to_find);
|
||||
|
||||
$actual_content = '';
|
||||
if ($end_tag_pos_in_escaped_part === false) {
|
||||
$this->addError("Unclosed special tag '<{$tag_name_from_parser}>'.");
|
||||
// Consume the rest of the document as content of this tag
|
||||
$actual_content = $remaining_doc_part;
|
||||
$this->position = $this->doc_length;
|
||||
} else {
|
||||
|
||||
$actual_content = substr($remaining_doc_part, 0, $end_tag_pos_in_escaped_part);
|
||||
|
||||
$this->position = $content_start_pos + strlen($actual_content) + strlen($end_tag_to_find);
|
||||
}
|
||||
|
||||
if ($actual_content !== '') {
|
||||
$this->dom_array[] = $actual_content;
|
||||
}
|
||||
}
|
||||
}
|
||||
Loading…
Reference in a new issue