Pull in the RemexHtml library, which is an HTML 5 library I recently created. RemexCompatMunger mutates the event stream, inserting <mw:p-wrap> elements where necessary, and occasionally taking even more invasive action such as reparenting and removing nodes maintained in Serializer's tree. RemexCompatFormatter produces a MediaWiki-style serialization which is relatively compatible with existing parser tests. It also does final empty element handling, including translating <mw:p-wrap> to <p> Tests are imported from both Html5Depurate and Subbu's pwrap.js. Depends-On: I864f31d9afdffdde49bfd39f07a0fb7f4df5c5d9 Change-Id: I900155b7dd199b0ae2a3b9cdb6db5136fc4f35a8
57 lines
1.4 KiB
PHP
57 lines
1.4 KiB
PHP
<?php
|
|
|
|
namespace MediaWiki\Tidy;
|
|
|
|
use RemexHtml\Serializer\Serializer;
|
|
use RemexHtml\Tokenizer\Tokenizer;
|
|
use RemexHtml\TreeBuilder\Dispatcher;
|
|
use RemexHtml\TreeBuilder\TreeBuilder;
|
|
use RemexHtml\TreeBuilder\TreeMutationTracer;
|
|
|
|
class RemexDriver extends TidyDriverBase {
|
|
private $trace;
|
|
private $pwrap;
|
|
|
|
public function __construct( array $config ) {
|
|
$config += [
|
|
'treeMutationTrace' => false,
|
|
'pwrap' => true
|
|
];
|
|
$this->trace = $config['treeMutationTrace'];
|
|
$this->pwrap = $config['pwrap'];
|
|
parent::__construct( $config );
|
|
}
|
|
|
|
public function tidy( $text ) {
|
|
$formatter = new RemexCompatFormatter;
|
|
$serializer = new Serializer( $formatter );
|
|
if ( $this->pwrap ) {
|
|
$munger = new RemexCompatMunger( $serializer );
|
|
} else {
|
|
$munger = $serializer;
|
|
}
|
|
if ( $this->trace ) {
|
|
$tracer = new TreeMutationTracer( $munger, function ( $msg ) {
|
|
wfDebug( "RemexHtml: $msg" );
|
|
} );
|
|
} else {
|
|
$tracer = $munger;
|
|
}
|
|
$treeBuilder = new TreeBuilder( $tracer, [
|
|
'ignoreErrors' => true,
|
|
'ignoreNulls' => true,
|
|
] );
|
|
$dispatcher = new Dispatcher( $treeBuilder );
|
|
$tokenizer = new Tokenizer( $dispatcher, $text, [
|
|
'ignoreErrors' => true,
|
|
'ignoreCharRefs' => true,
|
|
'ignoreNulls' => true,
|
|
'skipPreprocess' => true,
|
|
] );
|
|
$tokenizer->execute( [
|
|
'fragmentNamespace' => \RemexHtml\HTMLData::NS_HTML,
|
|
'fragmentName' => 'body'
|
|
] );
|
|
return $serializer->getResult();
|
|
}
|
|
}
|