X-Git-Url: https://git.immae.eu/?a=blobdiff_plain;f=application%2Fbookmark%2FLinkUtils.php;h=cf97e3b0a904a08d72a5ee7d3467efb1536f5c96;hb=d9d71b10c3bc70a0881d630b37dc4e918c9e812f;hp=8837943037dd52468ff6e73bbbc39f8e669b1b04;hpb=3fb29fdda04ca86e04422d49b86cf646d53c4f9d;p=github%2Fshaarli%2FShaarli.git diff --git a/application/bookmark/LinkUtils.php b/application/bookmark/LinkUtils.php index 88379430..cf97e3b0 100644 --- a/application/bookmark/LinkUtils.php +++ b/application/bookmark/LinkUtils.php @@ -2,112 +2,6 @@ use Shaarli\Bookmark\Bookmark; -/** - * Get cURL callback function for CURLOPT_WRITEFUNCTION - * - * @param string $charset to extract from the downloaded page (reference) - * @param string $title to extract from the downloaded page (reference) - * @param string $description to extract from the downloaded page (reference) - * @param string $keywords to extract from the downloaded page (reference) - * @param bool $retrieveDescription Automatically tries to retrieve description and keywords from HTML content - * @param string $curlGetInfo Optionally overrides curl_getinfo function - * - * @return Closure - */ -function get_curl_download_callback( - &$charset, - &$title, - &$description, - &$keywords, - $retrieveDescription, - $curlGetInfo = 'curl_getinfo' -) { - $isRedirected = false; - $currentChunk = 0; - $foundChunk = null; - - /** - * cURL callback function for CURLOPT_WRITEFUNCTION (called during the download). - * - * While downloading the remote page, we check that the HTTP code is 200 and content type is 'html/text' - * Then we extract the title and the charset and stop the download when it's done. - * - * @param resource $ch cURL resource - * @param string $data chunk of data being downloaded - * - * @return int|bool length of $data or false if we need to stop the download - */ - return function (&$ch, $data) use ( - $retrieveDescription, - $curlGetInfo, - &$charset, - &$title, - &$description, - &$keywords, - &$isRedirected, - &$currentChunk, - &$foundChunk - ) { - $currentChunk++; - $responseCode = $curlGetInfo($ch, CURLINFO_RESPONSE_CODE); - if (!empty($responseCode) && in_array($responseCode, [301, 302])) { - $isRedirected = true; - return strlen($data); - } - if (!empty($responseCode) && $responseCode !== 200) { - return false; - } - // After a redirection, the content type will keep the previous request value - // until it finds the next content-type header. - if (! $isRedirected || strpos(strtolower($data), 'content-type') !== false) { - $contentType = $curlGetInfo($ch, CURLINFO_CONTENT_TYPE); - } - if (!empty($contentType) && strpos($contentType, 'text/html') === false) { - return false; - } - if (!empty($contentType) && empty($charset)) { - $charset = header_extract_charset($contentType); - } - if (empty($charset)) { - $charset = html_extract_charset($data); - } - if (empty($title)) { - $title = html_extract_title($data); - $foundChunk = ! empty($title) ? $currentChunk : $foundChunk; - } - if ($retrieveDescription && empty($description)) { - $description = html_extract_tag('description', $data); - $foundChunk = ! empty($description) ? $currentChunk : $foundChunk; - } - if ($retrieveDescription && empty($keywords)) { - $keywords = html_extract_tag('keywords', $data); - if (! empty($keywords)) { - $foundChunk = $currentChunk; - // Keywords use the format tag1, tag2 multiple words, tag - // So we format them to match Shaarli's separator and glue multiple words with '-' - $keywords = implode(' ', array_map(function($keyword) { - return implode('-', preg_split('/\s+/', trim($keyword))); - }, explode(',', $keywords))); - } - } - - // We got everything we want, stop the download. - // If we already found either the title, description or keywords, - // it's highly unlikely that we'll found the other metas further than - // in the same chunk of data or the next one. So we also stop the download after that. - if ((!empty($responseCode) && !empty($contentType) && !empty($charset)) && $foundChunk !== null - && (! $retrieveDescription - || $foundChunk < $currentChunk - || (!empty($title) && !empty($description) && !empty($keywords)) - ) - ) { - return false; - } - - return strlen($data); - }; -} - /** * Extract title from an HTML document. * @@ -132,7 +26,7 @@ function html_extract_title($html) */ function header_extract_charset($header) { - preg_match('/charset="?([^; ]+)/i', $header, $match); + preg_match('/charset=["\']?([^; "\']+)/i', $header, $match); if (! empty($match[1])) { return strtolower(trim($match[1])); } @@ -172,16 +66,18 @@ function html_extract_tag($tag, $html) { $propertiesKey = ['property', 'name', 'itemprop']; $properties = implode('|', $propertiesKey); - // Try to retrieve OpenGraph image. - $ogRegex = '#]+(?:'. $properties .')=["\']?(?:og:)?'. $tag .'["\'\s][^>]*content=["\']?(.*?)["\'/>]#'; + // We need a OR here to accept either 'property=og:noquote' or 'property="og:unrelated og:my-tag"' + $orCondition = '["\']?(?:og:)?'. $tag .'["\']?|["\'][^\'"]*?(?:og:)?' . $tag . '[^\'"]*?[\'"]'; + // Try to retrieve OpenGraph tag. + $ogRegex = '#]+(?:'. $properties .')=(?:'. $orCondition .')[^>]*content=(["\'])([^\1]*?)\1.*?>#'; // If the attributes are not in the order property => content (e.g. Github) // New regex to keep this readable... more or less. - $ogRegexReverse = '#]+content=["\']([^"\']+)[^>]+(?:'. $properties .')=["\']?(?:og)?:'. $tag .'["\'\s/>]#'; + $ogRegexReverse = '#]+content=(["\'])([^\1]*?)\1[^>]+(?:'. $properties .')=(?:'. $orCondition .').*?>#'; if (preg_match($ogRegex, $html, $matches) > 0 || preg_match($ogRegexReverse, $html, $matches) > 0 ) { - return $matches[1]; + return $matches[2]; } return false; @@ -220,7 +116,7 @@ function hashtag_autolink($description, $indexUrl = '') * \p{Mn} - any non marking space (accents, umlauts, etc) */ $regex = '/(^|\s)#([\p{Pc}\p{N}\p{L}\p{Mn}]+)/mui'; - $replacement = '$1#$2'; + $replacement = '$1#$2'; return preg_replace($regex, $replacement, $description); } @@ -242,12 +138,17 @@ function space2nbsp($text) * * @param string $description shaare's description. * @param string $indexUrl URL to Shaarli's index. - + * @param bool $autolink Turn on/off automatic linkifications of URLs and hashtags + * * @return string formatted description. */ -function format_description($description, $indexUrl = '') +function format_description($description, $indexUrl = '', $autolink = true) { - return nl2br(space2nbsp(hashtag_autolink(text2clickable($description), $indexUrl))); + if ($autolink) { + $description = hashtag_autolink(text2clickable($description), $indexUrl); + } + + return nl2br(space2nbsp($description)); } /** @@ -275,3 +176,49 @@ function is_note($linkUrl) { return isset($linkUrl[0]) && $linkUrl[0] === '?'; } + +/** + * Extract an array of tags from a given tag string, with provided separator. + * + * @param string|null $tags String containing a list of tags separated by $separator. + * @param string $separator Shaarli's default: ' ' (whitespace) + * + * @return array List of tags + */ +function tags_str2array(?string $tags, string $separator): array +{ + // For whitespaces, we use the special \s regex character + $separator = $separator === ' ' ? '\s' : $separator; + + return preg_split('/\s*' . $separator . '+\s*/', trim($tags) ?? '', -1, PREG_SPLIT_NO_EMPTY); +} + +/** + * Return a tag string with provided separator from a list of tags. + * Note that given array is clean up by tags_filter(). + * + * @param array|null $tags List of tags + * @param string $separator + * + * @return string + */ +function tags_array2str(?array $tags, string $separator): string +{ + return implode($separator, tags_filter($tags, $separator)); +} + +/** + * Clean an array of tags: trim + remove empty entries + * + * @param array|null $tags List of tags + * @param string $separator + * + * @return array + */ +function tags_filter(?array $tags, string $separator): array +{ + $trimDefault = " \t\n\r\0\x0B"; + return array_values(array_filter(array_map(function (string $entry) use ($separator, $trimDefault): string { + return trim($entry, $trimDefault . $separator); + }, $tags ?? []))); +}