Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
94.23% |
49 / 52 |
|
75.00% |
3 / 4 |
CRAP | |
0.00% |
0 / 1 |
| Validator | |
94.23% |
49 / 52 |
|
75.00% |
3 / 4 |
32.20 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
1 | |||
| validateRef | |
85.00% |
17 / 20 |
|
0.00% |
0 / 1 |
19.09 | |||
| validateRefOutsideOfReferenceList | |
100.00% |
11 / 11 |
|
100.00% |
1 / 1 |
6 | |||
| validateRefInReferenceList | |
100.00% |
17 / 17 |
|
100.00% |
1 / 1 |
7 | |||
| 1 | <?php |
| 2 | |
| 3 | namespace Cite; |
| 4 | |
| 5 | use MediaWiki\Parser\Sanitizer; |
| 6 | use StatusValue; |
| 7 | |
| 8 | /** |
| 9 | * Context-aware, detailed validation of the arguments and content of a <ref> tag. |
| 10 | * |
| 11 | * @license GPL-2.0-or-later |
| 12 | */ |
| 13 | class Validator { |
| 14 | |
| 15 | private ReferenceStack $referenceStack; |
| 16 | private ?string $inReferencesGroup; |
| 17 | private bool $isSectionPreview; |
| 18 | private bool $isExtendsEnabled; |
| 19 | |
| 20 | /** |
| 21 | * @param ReferenceStack $referenceStack |
| 22 | * @param string|null $inReferencesGroup Group name of the <references> context to consider |
| 23 | * during validation. Null if we are currently not in a <references> context. |
| 24 | * @param bool $isSectionPreview Validation is relaxed when previewing parts of a page |
| 25 | * @param bool $isExtendsEnabled Temporary feature flag |
| 26 | */ |
| 27 | public function __construct( |
| 28 | ReferenceStack $referenceStack, |
| 29 | ?string $inReferencesGroup = null, |
| 30 | bool $isSectionPreview = false, |
| 31 | bool $isExtendsEnabled = false |
| 32 | ) { |
| 33 | $this->referenceStack = $referenceStack; |
| 34 | $this->inReferencesGroup = $inReferencesGroup; |
| 35 | $this->isSectionPreview = $isSectionPreview; |
| 36 | $this->isExtendsEnabled = $isExtendsEnabled; |
| 37 | } |
| 38 | |
| 39 | public function validateRef( |
| 40 | ?string $text, |
| 41 | string $group, |
| 42 | ?string $name, |
| 43 | ?string $extends, |
| 44 | ?string $follow, |
| 45 | ?string $dir |
| 46 | ): StatusValue { |
| 47 | if ( ctype_digit( (string)$name ) |
| 48 | || ctype_digit( (string)$extends ) |
| 49 | || ctype_digit( (string)$follow ) |
| 50 | ) { |
| 51 | // Numeric names mess up the resulting id's, potentially producing |
| 52 | // duplicate id's in the XHTML. The Right Thing To Do |
| 53 | // would be to mangle them, but it's not really high-priority |
| 54 | // (and would produce weird id's anyway). |
| 55 | return StatusValue::newFatal( 'cite_error_ref_numeric_key' ); |
| 56 | } |
| 57 | |
| 58 | if ( $extends ) { |
| 59 | // Temporary feature flag until mainstreamed, see T236255 |
| 60 | if ( !$this->isExtendsEnabled ) { |
| 61 | return StatusValue::newFatal( 'cite_error_ref_too_many_keys' ); |
| 62 | } |
| 63 | |
| 64 | $groupRefs = $this->referenceStack->getGroupRefs( $group ); |
| 65 | if ( $name !== null && isset( $groupRefs[$name] ) && $groupRefs[$name]->extends === null ) { |
| 66 | // T242141: A top-level <ref> can't be changed into a sub-reference |
| 67 | return StatusValue::newFatal( 'cite_error_references_duplicate_key', $name ); |
| 68 | } elseif ( isset( $groupRefs[$extends] ) && $groupRefs[$extends]->extends !== null ) { |
| 69 | // A sub-reference can not be extended a second time (no nesting) |
| 70 | return StatusValue::newFatal( 'cite_error_ref_nested_extends', $extends, |
| 71 | $groupRefs[$extends]->extends ?? '' ); |
| 72 | } |
| 73 | } |
| 74 | |
| 75 | if ( $follow && ( $name || $extends ) ) { |
| 76 | return StatusValue::newFatal( 'cite_error_ref_follow_conflicts' ); |
| 77 | } |
| 78 | |
| 79 | if ( $dir !== null && $dir !== 'rtl' && $dir !== 'ltr' ) { |
| 80 | return StatusValue::newFatal( 'cite_error_ref_invalid_dir', $dir ); |
| 81 | } |
| 82 | |
| 83 | return $this->inReferencesGroup === null ? |
| 84 | $this->validateRefOutsideOfReferenceList( $text, $name ) : |
| 85 | $this->validateRefInReferenceList( $text, $group, $name ); |
| 86 | } |
| 87 | |
| 88 | private function validateRefOutsideOfReferenceList( |
| 89 | ?string $text, |
| 90 | ?string $name |
| 91 | ): StatusValue { |
| 92 | if ( !$name ) { |
| 93 | if ( $text === null ) { |
| 94 | // Completely empty ref like <ref /> is forbidden. |
| 95 | return StatusValue::newFatal( 'cite_error_ref_no_key' ); |
| 96 | } elseif ( trim( $text ) === '' ) { |
| 97 | // Must have content or reuse another ref by name. |
| 98 | return StatusValue::newFatal( 'cite_error_ref_no_input' ); |
| 99 | } |
| 100 | } |
| 101 | |
| 102 | if ( $text !== null && preg_match( |
| 103 | '/<ref(erences)?\b[^>]*+>/i', |
| 104 | preg_replace( '#<(\w++)[^>]*+>.*?</\1\s*>|<!--.*?-->#s', '', $text ) |
| 105 | ) ) { |
| 106 | // (bug T8199) This most likely implies that someone left off the |
| 107 | // closing </ref> tag, which will cause the entire article to be |
| 108 | // eaten up until the next <ref>. So we bail out early instead. |
| 109 | // The fancy regex above first tries chopping out anything that |
| 110 | // looks like a comment or SGML tag, which is a crude way to avoid |
| 111 | // false alarms for <nowiki>, <pre>, etc. |
| 112 | // |
| 113 | // Possible improvement: print the warning, followed by the contents |
| 114 | // of the <ref> tag. This way no part of the article will be eaten |
| 115 | // even temporarily. |
| 116 | return StatusValue::newFatal( 'cite_error_included_ref' ); |
| 117 | } |
| 118 | |
| 119 | return StatusValue::newGood(); |
| 120 | } |
| 121 | |
| 122 | private function validateRefInReferenceList( |
| 123 | ?string $text, |
| 124 | string $group, |
| 125 | ?string $name |
| 126 | ): StatusValue { |
| 127 | if ( $group !== $this->inReferencesGroup ) { |
| 128 | // <ref> and <references> have conflicting group attributes. |
| 129 | return StatusValue::newFatal( 'cite_error_references_group_mismatch', |
| 130 | Sanitizer::safeEncodeAttribute( $group ) ); |
| 131 | } |
| 132 | |
| 133 | if ( !$name ) { |
| 134 | // <ref> calls inside <references> must be named |
| 135 | return StatusValue::newFatal( 'cite_error_references_no_key' ); |
| 136 | } |
| 137 | |
| 138 | if ( $text === null || trim( $text ) === '' ) { |
| 139 | // <ref> called in <references> has no content. |
| 140 | return StatusValue::newFatal( |
| 141 | 'cite_error_empty_references_define', |
| 142 | Sanitizer::safeEncodeAttribute( $name ), |
| 143 | Sanitizer::safeEncodeAttribute( $group ) |
| 144 | ); |
| 145 | } |
| 146 | |
| 147 | // Section previews are exempt from some rules. |
| 148 | if ( !$this->isSectionPreview ) { |
| 149 | $groupRefs = $this->referenceStack->getGroupRefs( $group ); |
| 150 | |
| 151 | if ( !isset( $groupRefs[$name] ) ) { |
| 152 | // No such named ref exists in this group. |
| 153 | return StatusValue::newFatal( 'cite_error_references_missing_key', |
| 154 | Sanitizer::safeEncodeAttribute( $name ) ); |
| 155 | } |
| 156 | } |
| 157 | |
| 158 | return StatusValue::newGood(); |
| 159 | } |
| 160 | |
| 161 | } |