parseStringTextAndPlaceholders = $parseStringTextAndPlaceholders; } /** * @param array $gettextStrings * * @return array */ public function run( array $gettextStrings, string $htmlString ): array { $matchedGettextStrings = []; foreach ( $gettextStrings as $gettextString ) { if ( $this->doesHtmlStringMatchGettextString( $gettextString, $htmlString ) ) { $matchedGettextStrings[] = $gettextString; continue; } $value = StringItem::filterOnlyTextFromValue( $gettextString[0] ); $customPlaceholderStartPos = strpos( $value, '[' ); $customPlaceholderEndPos = strpos( $value, ']' ); $hasCustomPlaceholder = ( $customPlaceholderStartPos !== false && $customPlaceholderEndPos !== false && $customPlaceholderStartPos < $customPlaceholderEndPos ); if ( ! $hasCustomPlaceholder ) { continue; } $pattern = '/(?doesHtmlStringMatchGettextString( $gettextStringCopy, $htmlString ) ) { $matchedGettextStrings[] = $gettextString; } } return $matchedGettextStrings; } /** * @param array{string, string, string|null} $gettextString */ private function doesHtmlStringMatchGettextString( array $gettextString, string $htmlString ): bool { $nodes = $this->parseStringTextAndPlaceholders->run( StringItem::filterOnlyTextFromValue( $gettextString[0] ) ); $nodes = $this->parseTextNodes( $nodes, $htmlString ); if ( is_null( $nodes ) ) { return false; } $nodes = $this->parsePlaceholderNodes( $nodes, $htmlString ); $placeholdersCount = 0; $notEmptyPlaceholdersCount = 0; $emptyPlaceholdersCount = 0; foreach ( $nodes as $node ) { if ( $node['type'] !== 'placeholder' ) { continue; } $placeholdersCount++; $words = array_values( array_filter( explode( ' ', $node['htmlStringText'] ) ) ); if ( count( $words ) === 0 ) { $emptyPlaceholdersCount++; } else { $notEmptyPlaceholdersCount++; } } if ( $placeholdersCount === 0 ) { /* * When we match html string to gettext string we should allow redundant text in the start or in the end of the html string. * This should cover the case when gettext string is outputted together with some text near it. * From other side we should disable this behaviour for posts - we could match by accident part of the some post sentence * with some gettext string even if it was never outputted there. Such collisions almost always happen only on gettext strings * which contains only single word, so we should allow redundant text for such cases only for short strings, which should * cover most of the cases. * Example: * gettext string = 'Older Posts' outputted into html as string and some extra text: * 1) '

ExtraText

'. // Should match * 2) '

ExtraText

'. // Should match * 3) '

ExtraText ExtraText

'. // Should match * 4) '

Some part of the post with Older Posts text part as a part of the post

'. // Should not match */ $gettextStringWordsCount = count( explode( ' ', StringItem::filterOnlyTextFromValue( $gettextString[0] ) ) ); $htmlStringWordsCount = count( explode( ' ', $htmlString ) ); return ( $gettextStringWordsCount >= 2 ) || ( $gettextStringWordsCount === 1 && $htmlStringWordsCount <= 3 ); } /* * Should allow following case: * Gettext string: 'Ready to publish your first post? %1$sGet started here%2$s.' * Matches to html string: 'Ready to publish your first post? Get started here.' * So, placeholders can be empty in such case, but we should not allow this on short strings, * otherwise extra redundant matches can occur. * We have >= in condition to cover case when only some of the placeholders are empty. */ if ( $emptyPlaceholdersCount > 0 && $notEmptyPlaceholdersCount >= 0 ) { $words = explode( ' ', $htmlString ); return count( $words ) >= 3; } // Normally we do not allow empty placeholders. return $emptyPlaceholdersCount === 0; } /** * @return array|null */ private function parseTextNodes( array $nodes, string $htmlString ) { $lastMatchPos = -1; $lastMatchStr = ''; for ( $i = 0; $i < count( $nodes ); $i++ ) { $node = $nodes[ $i ]; if ( $node['type'] !== 'text' ) { continue; } // All non-placeholder nodes from gettext string should have direct match in html string in the same order. $minSearchOffset = ( $lastMatchPos !== -1 ) ? $lastMatchPos + strlen( $lastMatchStr ) - 1 : 0; $matchPos = strpos( $htmlString, $node['text'], $minSearchOffset ); if ( $matchPos === false || $matchPos < $lastMatchPos ) { return null; } $nodes[ $i ]['offsetInHtmlString'] = $matchPos; $nodes[ $i ]['offsetEndInHtmlString'] = $matchPos + strlen( $node['text'] ) - 1; $nodes[ $i ]['htmlStringText'] = $node['text']; $lastMatchPos = $matchPos; $lastMatchStr = $node['text']; } return $nodes; } private function parsePlaceholderNodes( array $nodes, string $htmlString ): array { $lastTextNode = null; for ( $i = 0; $i < count( $nodes ); $i++ ) { $node = $nodes[ $i ]; if ( $node['type'] === 'text' ) { $lastTextNode = $node; continue; } $nextTextNode = null; for ( $j = $i + 1; $j < count( $nodes ); $j++ ) { if ( $nodes[ $j ]['type'] === 'text' ) { $nextTextNode = $nodes[ $j ]; break; } } // Two placeholders in a row '%s%s' will capture the same text(as we cannot determine boundaries in that case). $nodes[ $i ]['offsetInHtmlString'] = $lastTextNode ? $lastTextNode['offsetEndInHtmlString'] + 1 : 0; $nodes[ $i ]['offsetEndInHtmlString'] = $nextTextNode ? $nextTextNode['offsetInHtmlString'] - 1 : strlen( $htmlString ) - 1; $nodes[ $i ]['htmlStringText'] = substr( $htmlString, $nodes[ $i ]['offsetInHtmlString'], $nodes[ $i ]['offsetEndInHtmlString'] - $nodes[ $i ]['offsetInHtmlString'] + 1 ); } return $nodes; } }