Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
95.00% |
38 / 40 |
|
50.00% |
2 / 4 |
CRAP | |
0.00% |
0 / 1 |
| RemexCompatFormatter | |
95.00% |
38 / 40 |
|
50.00% |
2 / 4 |
20 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
7 / 7 |
|
100.00% |
1 / 1 |
1 | |||
| startDocument | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| characters | |
85.71% |
6 / 7 |
|
0.00% |
0 / 1 |
4.05 | |||
| element | |
96.00% |
24 / 25 |
|
0.00% |
0 / 1 |
14 | |||
| 1 | <?php |
| 2 | |
| 3 | namespace MediaWiki\Tidy; |
| 4 | |
| 5 | use MediaWiki\Parser\Sanitizer; |
| 6 | use Wikimedia\RemexHtml\HTMLData; |
| 7 | use Wikimedia\RemexHtml\Serializer\HtmlFormatter; |
| 8 | use Wikimedia\RemexHtml\Serializer\SerializerNode; |
| 9 | |
| 10 | /** |
| 11 | * @internal |
| 12 | * |
| 13 | * WATCH OUT! Unlike normal HtmlFormatter, this class requires the 'ignoreCharRefs' option |
| 14 | * in Tokenizer to be used. If that option is not used, it will produce wrong results (T354361). |
| 15 | */ |
| 16 | class RemexCompatFormatter extends HtmlFormatter { |
| 17 | private const MARKED_EMPTY_ELEMENTS = [ |
| 18 | 'li' => true, |
| 19 | 'p' => true, |
| 20 | 'tr' => true, |
| 21 | ]; |
| 22 | |
| 23 | /** @var ?callable */ |
| 24 | private $textProcessor; |
| 25 | |
| 26 | public function __construct( $options = [] ) { |
| 27 | parent::__construct( $options ); |
| 28 | // Escape non-breaking space |
| 29 | $this->attributeEscapes["\u{00A0}"] = ' '; |
| 30 | $this->textEscapes["\u{00A0}"] = ' '; |
| 31 | // Escape U+0338 (T387130) |
| 32 | $this->textEscapes["\u{0338}"] = '̸'; |
| 33 | // Disable escaping of '&', because we expect to see entities, due to 'ignoreCharRefs' |
| 34 | unset( $this->attributeEscapes["&"] ); |
| 35 | unset( $this->textEscapes["&"] ); |
| 36 | $this->textProcessor = $options['textProcessor'] ?? null; |
| 37 | } |
| 38 | |
| 39 | public function startDocument( $fragmentNamespace, $fragmentName ) { |
| 40 | return ''; |
| 41 | } |
| 42 | |
| 43 | /** |
| 44 | * WATCH OUT! Unlike normal HtmlFormatter, this class expects that the $text argument contains |
| 45 | * unexpanded character references (entities), as a result of using the 'ignoreCharRefs' option |
| 46 | * in Tokenizer. If that option is not used, this method will produce wrong results (T354361). |
| 47 | * |
| 48 | * @inheritDoc |
| 49 | */ |
| 50 | public function characters( SerializerNode $parent, $text, $start, $length ) { |
| 51 | $text = parent::characters( $parent, $text, $start, $length ); |
| 52 | |
| 53 | if ( $parent->namespace !== HTMLData::NS_HTML |
| 54 | || !isset( $this->rawTextElements[$parent->name] ) |
| 55 | ) { |
| 56 | if ( $this->textProcessor !== null ) { |
| 57 | $text = ( $this->textProcessor )( $text ); |
| 58 | } |
| 59 | } |
| 60 | |
| 61 | // Ensure a consistent representation for all entities |
| 62 | $text = Sanitizer::normalizeCharReferences( $text ); |
| 63 | return $text; |
| 64 | } |
| 65 | |
| 66 | public function element( SerializerNode $parent, SerializerNode $node, $contents ) { |
| 67 | $data = $node->snData; |
| 68 | if ( $data && $data->isPWrapper ) { |
| 69 | if ( $data->nonblankNodeCount ) { |
| 70 | return "<p>$contents</p>"; |
| 71 | } else { |
| 72 | return $contents; |
| 73 | } |
| 74 | } |
| 75 | |
| 76 | $name = $node->name; |
| 77 | $attrs = $node->attrs; |
| 78 | if ( isset( self::MARKED_EMPTY_ELEMENTS[$name] ) && $attrs->count() === 0 |
| 79 | && strspn( $contents, "\t\n\f\r " ) === strlen( $contents ) |
| 80 | ) { |
| 81 | return "<{$name} class=\"mw-empty-elt\">$contents</{$name}>"; |
| 82 | } |
| 83 | |
| 84 | $s = "<$name"; |
| 85 | foreach ( $attrs->getValues() as $attrName => $attrValue ) { |
| 86 | $encValue = strtr( $attrValue, $this->attributeEscapes ); |
| 87 | $encValue = Sanitizer::normalizeCharReferences( $encValue ); |
| 88 | $s .= " $attrName=\"$encValue\""; |
| 89 | } |
| 90 | if ( $node->namespace === HTMLData::NS_HTML && isset( $this->voidElements[$name] ) ) { |
| 91 | $s .= ' />'; |
| 92 | return $s; |
| 93 | } |
| 94 | |
| 95 | $s .= '>'; |
| 96 | if ( $node->namespace === HTMLData::NS_HTML |
| 97 | && isset( $contents[0] ) && $contents[0] === "\n" |
| 98 | && isset( $this->prefixLfElements[$name] ) |
| 99 | ) { |
| 100 | $s .= "\n$contents</$name>"; |
| 101 | } else { |
| 102 | $s .= "$contents</$name>"; |
| 103 | } |
| 104 | return $s; |
| 105 | } |
| 106 | } |