Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
98.02% |
99 / 101 |
|
85.71% |
6 / 7 |
CRAP | |
0.00% |
0 / 1 |
| AnnotatedHtmlDiffFormatter | |
98.02% |
99 / 101 |
|
85.71% |
6 / 7 |
31 | |
0.00% |
0 / 1 |
| format | |
100.00% |
63 / 63 |
|
100.00% |
1 / 1 |
14 | |||
| rTrimmedWordLevelDiff | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
1 | |||
| commonRTrim | |
80.00% |
8 / 10 |
|
0.00% |
0 / 1 |
5.20 | |||
| getOriginalInlineDiff | |
100.00% |
7 / 7 |
|
100.00% |
1 / 1 |
4 | |||
| getClosingInlineDiff | |
100.00% |
7 / 7 |
|
100.00% |
1 / 1 |
4 | |||
| getWordAccumulator | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
1 | |||
| composeHtml | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
2 | |||
| 1 | <?php |
| 2 | |
| 3 | namespace TwoColConflict; |
| 4 | |
| 5 | use Wikimedia\Diff\ComplexityException; |
| 6 | use Wikimedia\Diff\Diff; |
| 7 | use Wikimedia\Diff\WordAccumulator; |
| 8 | use Wikimedia\Diff\WordLevelDiff; |
| 9 | |
| 10 | /** |
| 11 | * @license GPL-2.0-or-later |
| 12 | * @author Christoph Jauera <christoph.jauera@wikimedia.de> |
| 13 | */ |
| 14 | class AnnotatedHtmlDiffFormatter { |
| 15 | |
| 16 | /** |
| 17 | * @param string[] $oldLines |
| 18 | * @param string[] $newLines |
| 19 | * @param string[] $preSaveTransformedLines |
| 20 | * |
| 21 | * @throws ComplexityException |
| 22 | * @return array[] List of changes, each of which include an HTML representation of the diff, |
| 23 | * and the original wikitext. Note the HTML does not use <br> but relies on `white-space: |
| 24 | * pre-line` being set! |
| 25 | * TODO: "preSavedTransformedLines" is still warty. |
| 26 | */ |
| 27 | public function format( |
| 28 | array $oldLines, |
| 29 | array $newLines, |
| 30 | array $preSaveTransformedLines |
| 31 | ): array { |
| 32 | $changes = []; |
| 33 | $oldLine = 0; |
| 34 | $newLine = 0; |
| 35 | $diff = new Diff( $oldLines, $preSaveTransformedLines ); |
| 36 | |
| 37 | foreach ( $diff->getEdits() as $edit ) { |
| 38 | switch ( $edit->getType() ) { |
| 39 | case 'add': |
| 40 | $changes[] = [ |
| 41 | 'action' => 'add', |
| 42 | 'oldhtml' => "\u{00A0}", |
| 43 | 'oldtext' => null, |
| 44 | 'newhtml' => '<ins class="mw-twocolconflict-diffchange">' . |
| 45 | $this->composeHtml( $edit->getClosing() ) . '</ins>', |
| 46 | 'newtext' => implode( "\n", |
| 47 | array_slice( $newLines, $newLine, $edit->nclosing() ) ), |
| 48 | ]; |
| 49 | break; |
| 50 | |
| 51 | case 'delete': |
| 52 | $changes[] = [ |
| 53 | 'action' => 'delete', |
| 54 | 'oldhtml' => '<del class="mw-twocolconflict-diffchange">' . |
| 55 | $this->composeHtml( $edit->getOrig() ) . '</del>', |
| 56 | 'oldtext' => implode( "\n", $edit->getOrig() ), |
| 57 | 'newhtml' => "\u{00A0}", |
| 58 | 'newtext' => null, |
| 59 | ]; |
| 60 | break; |
| 61 | |
| 62 | case 'change': |
| 63 | $wordLevelDiff = $this->rTrimmedWordLevelDiff( $edit->getOrig(), $edit->getClosing() ); |
| 64 | $changes[] = [ |
| 65 | 'action' => 'change', |
| 66 | 'oldhtml' => $this->getOriginalInlineDiff( $wordLevelDiff ), |
| 67 | 'oldtext' => implode( "\n", $edit->getOrig() ), |
| 68 | 'newhtml' => $this->getClosingInlineDiff( $wordLevelDiff ), |
| 69 | 'newtext' => implode( "\n", |
| 70 | array_slice( $newLines, $newLine, $edit->nclosing() ) ), |
| 71 | ]; |
| 72 | break; |
| 73 | |
| 74 | case 'copy': |
| 75 | $changes[] = [ |
| 76 | 'action' => 'copy', |
| 77 | // Warning, this must be unescaped Wikitext, not escaped HTML! |
| 78 | 'copytext' => implode( "\n", $edit->getOrig() ), |
| 79 | ]; |
| 80 | break; |
| 81 | } |
| 82 | |
| 83 | $oldLine += $edit->norig(); |
| 84 | $newLine += $edit->nclosing(); |
| 85 | } |
| 86 | |
| 87 | // Try to merge unchanged newline-only rows into a more meaningful row |
| 88 | foreach ( $changes as $i => $row ) { |
| 89 | if ( !isset( $row['copytext'] ) || trim( $row['copytext'], "\n" ) !== '' ) { |
| 90 | continue; |
| 91 | } |
| 92 | |
| 93 | // Prefer adding extra empty lines to the end of the previous row |
| 94 | foreach ( [ -1, 1 ] as $offset ) { |
| 95 | if ( !isset( $changes[$i + $offset] ) ) { |
| 96 | continue; |
| 97 | } |
| 98 | |
| 99 | $target = &$changes[$i + $offset]; |
| 100 | if ( isset( $target['oldtext'] ) && isset( $target['newtext'] ) ) { |
| 101 | $extra = "\n" . $row['copytext']; |
| 102 | if ( $offset < 0 ) { |
| 103 | $target['oldtext'] .= $extra; |
| 104 | $target['newtext'] .= $extra; |
| 105 | } else { |
| 106 | $target['oldtext'] = $extra . $target['oldtext']; |
| 107 | $target['newtext'] = $extra . $target['newtext']; |
| 108 | } |
| 109 | unset( $changes[$i] ); |
| 110 | break; |
| 111 | } |
| 112 | } |
| 113 | } |
| 114 | |
| 115 | return array_values( $changes ); |
| 116 | } |
| 117 | |
| 118 | /** |
| 119 | * @param string[] $before |
| 120 | * @param string[] $after |
| 121 | * |
| 122 | * @return WordLevelDiff |
| 123 | */ |
| 124 | private function rTrimmedWordLevelDiff( array $before, array $after ): WordLevelDiff { |
| 125 | end( $before ); |
| 126 | end( $after ); |
| 127 | $this->commonRTrim( $before[key( $before )], $after[key( $after )] ); |
| 128 | return new WordLevelDiff( $before, $after ); |
| 129 | } |
| 130 | |
| 131 | /** |
| 132 | * Trims identical sequences of whitespace from the end of both lines. |
| 133 | * |
| 134 | * @param string &$before |
| 135 | * @param string &$after |
| 136 | */ |
| 137 | private function commonRTrim( string &$before, string &$after ): void { |
| 138 | $uncommonBefore = strlen( $before ); |
| 139 | $uncommonAfter = strlen( $after ); |
| 140 | while ( $uncommonBefore > 0 && |
| 141 | $uncommonAfter > 0 && |
| 142 | $before[$uncommonBefore - 1] === $after[$uncommonAfter - 1] && |
| 143 | ctype_space( $after[$uncommonAfter - 1] ) |
| 144 | ) { |
| 145 | $uncommonBefore--; |
| 146 | $uncommonAfter--; |
| 147 | } |
| 148 | $before = substr( $before, 0, $uncommonBefore ); |
| 149 | $after = substr( $after, 0, $uncommonAfter ); |
| 150 | } |
| 151 | |
| 152 | /** |
| 153 | * Composes lines from a WordLevelDiff and marks removed words. |
| 154 | * |
| 155 | * @param WordLevelDiff $diff |
| 156 | * |
| 157 | * @return string Composed HTML string with inline markup |
| 158 | */ |
| 159 | private function getOriginalInlineDiff( WordLevelDiff $diff ): string { |
| 160 | $wordAccumulator = $this->getWordAccumulator(); |
| 161 | |
| 162 | foreach ( $diff->getEdits() as $edit ) { |
| 163 | if ( $edit->type === 'copy' ) { |
| 164 | $wordAccumulator->addWords( $edit->orig ); |
| 165 | } elseif ( $edit->orig ) { |
| 166 | $wordAccumulator->addWords( $edit->orig, 'del' ); |
| 167 | } |
| 168 | } |
| 169 | return implode( "\n", $wordAccumulator->getLines() ); |
| 170 | } |
| 171 | |
| 172 | /** |
| 173 | * Composes lines from a WordLevelDiff and marks added words. |
| 174 | * |
| 175 | * @param WordLevelDiff $diff |
| 176 | * |
| 177 | * @return string Composed HTML string with inline markup |
| 178 | */ |
| 179 | private function getClosingInlineDiff( WordLevelDiff $diff ): string { |
| 180 | $wordAccumulator = $this->getWordAccumulator(); |
| 181 | |
| 182 | foreach ( $diff->getEdits() as $edit ) { |
| 183 | if ( $edit->type === 'copy' ) { |
| 184 | $wordAccumulator->addWords( $edit->closing ); |
| 185 | } elseif ( $edit->closing ) { |
| 186 | $wordAccumulator->addWords( $edit->closing, 'ins' ); |
| 187 | } |
| 188 | } |
| 189 | return implode( "\n", $wordAccumulator->getLines() ); |
| 190 | } |
| 191 | |
| 192 | /** |
| 193 | * @return WordAccumulator |
| 194 | */ |
| 195 | private function getWordAccumulator(): WordAccumulator { |
| 196 | $wordAccumulator = new WordAccumulator(); |
| 197 | $wordAccumulator->insClass = ' class="mw-twocolconflict-diffchange"'; |
| 198 | $wordAccumulator->delClass = ' class="mw-twocolconflict-diffchange"'; |
| 199 | return $wordAccumulator; |
| 200 | } |
| 201 | |
| 202 | /** |
| 203 | * @param string[] $lines |
| 204 | * |
| 205 | * @return string HTML without <br>, relying on `white-space: pre-line` being set |
| 206 | */ |
| 207 | private function composeHtml( array $lines ): string { |
| 208 | return htmlspecialchars( implode( "\n", array_map( |
| 209 | static function ( string $line ): string { |
| 210 | // Replace empty lines with a non-breaking space |
| 211 | return $line === '' ? "\u{00A0}" : $line; |
| 212 | }, |
| 213 | $lines |
| 214 | ) ) ); |
| 215 | } |
| 216 | |
| 217 | } |