Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
96.30% |
26 / 27 |
|
50.00% |
1 / 2 |
CRAP | |
0.00% |
0 / 1 |
| RemoveRedLinks | |
96.30% |
26 / 27 |
|
50.00% |
1 / 2 |
10 | |
0.00% |
0 / 1 |
| __construct | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| run | |
100.00% |
26 / 26 |
|
100.00% |
1 / 1 |
9 | |||
| 1 | <?php |
| 2 | declare( strict_types = 1 ); |
| 3 | |
| 4 | namespace Wikimedia\Parsoid\Html2Wt; |
| 5 | |
| 6 | use Wikimedia\Parsoid\Config\Env; |
| 7 | use Wikimedia\Parsoid\DOM\Element; |
| 8 | use Wikimedia\Parsoid\DOM\Node; |
| 9 | use Wikimedia\Parsoid\Utils\DOMCompat; |
| 10 | use Wikimedia\Parsoid\Utils\UrlUtils; |
| 11 | |
| 12 | class RemoveRedLinks { |
| 13 | |
| 14 | private Env $env; |
| 15 | |
| 16 | public function __construct( Env $env ) { |
| 17 | $this->env = $env; |
| 18 | } |
| 19 | |
| 20 | /** |
| 21 | * Remove redlinks from a document |
| 22 | * @param Element $root |
| 23 | */ |
| 24 | public function run( Node $root ): void { |
| 25 | '@phan-var Element|DocumentFragment $root'; // @var Element|DocumentFragment $root |
| 26 | $wikilinks = DOMCompat::querySelectorAll( $root, 'a[rel~="mw:WikiLink"].new' ); |
| 27 | foreach ( $wikilinks as $a ) { |
| 28 | $href = DOMCompat::getAttribute( $a, 'href' ); |
| 29 | $qmPos = strpos( $href ?? '', '?' ); |
| 30 | if ( $qmPos !== false ) { |
| 31 | $parsedURL = UrlUtils::parseUrl( $href ); |
| 32 | if ( !isset( $parsedURL['query'] ) ) { |
| 33 | // TODO this mitigates a bug in the AddRedLinks pass, which puts the query |
| 34 | // parameters AFTER a fragment; the parse_url then interprets these query parameters |
| 35 | // as part of the fragment. |
| 36 | // 2022-12-01: That issue is solved in the "wt2html" direction, but some |
| 37 | // RESTBase-stored content may still exist, so we'll have to remove this when |
| 38 | // 2.7.0 is not longer being stored. |
| 39 | $href = str_replace( |
| 40 | [ '?action=edit&redlink=1', '?action=edit&redlink=1', |
| 41 | '&action=edit&redlink=1', '&action=edit&redlink=1' ], |
| 42 | [ '', '', '', '' ], |
| 43 | $href |
| 44 | ); |
| 45 | } else { |
| 46 | $queryParams = $parsedURL['query']; |
| 47 | $queryElts = []; |
| 48 | parse_str( $queryParams, $queryElts ); |
| 49 | if ( isset( $queryElts['action'] ) && $queryElts['action'] === 'edit' ) { |
| 50 | unset( $queryElts['action'] ); |
| 51 | } |
| 52 | if ( isset( $queryElts['redlink'] ) && $queryElts['redlink'] === '1' ) { |
| 53 | unset( $queryElts['redlink'] ); |
| 54 | } |
| 55 | |
| 56 | // My understanding of this method and of PHP array handling makes me |
| 57 | // believe that the order of the parameters should not be modified here. |
| 58 | // There is however no guarantee whatsoever in the documentation or spec |
| 59 | // of these methods. |
| 60 | $newQueryParams = http_build_query( $queryElts ); |
| 61 | |
| 62 | if ( count( $queryElts ) === 0 ) { |
| 63 | // avoids the insertion of ? on empty query string |
| 64 | $parsedURL['query'] = null; |
| 65 | } else { |
| 66 | $parsedURL['query'] = http_build_query( $queryElts ); |
| 67 | } |
| 68 | $href = UrlUtils::assembleUrl( $parsedURL ); |
| 69 | } |
| 70 | $a->setAttribute( 'href', $href ); |
| 71 | } |
| 72 | } |
| 73 | } |
| 74 | } |