X-Git-Url: https://git.immae.eu/?a=blobdiff_plain;f=src%2FWallabag%2FCoreBundle%2FHelper%2FContentProxy.php;h=1a2a330ff085f5e922da295fbd5c908611936545;hb=fc040c749dec0275e562182562c1c1cb89e6cfa1;hp=fab052685ccef748c6aba14591e05fc7866780cb;hpb=f80f16dfc858ec90da76daacd405b0cfdaa32f74;p=github%2Fwallabag%2Fwallabag.git diff --git a/src/Wallabag/CoreBundle/Helper/ContentProxy.php b/src/Wallabag/CoreBundle/Helper/ContentProxy.php index fab05268..1a2a330f 100644 --- a/src/Wallabag/CoreBundle/Helper/ContentProxy.php +++ b/src/Wallabag/CoreBundle/Helper/ContentProxy.php @@ -66,49 +66,14 @@ class ContentProxy // so we'll be able to refetch it in the future $content['url'] = !empty($content['url']) ? $content['url'] : $url; - $this->stockEntry($entry, $content); - } - - /** - * Try to sanitize the title of the fetched content from wrong character encodings and invalid UTF-8 character. - * @param $title - * @param $contentType - * @return string - */ - private function sanitizeContentTitle($title, $contentType) { - if ('application/pdf' === $contentType) { - $convertedTitle = $this->convertPdfEncodingToUTF8($title); - return $this->sanitizeUTF8Text($convertedTitle); - } - return $this->sanitizeUTF8Text($title); - } - - /** - * If the title from the fetched content comes from a PDF, then its very possible that the character encoding is not - * UTF-8. This methods tries to identify the character encoding and translate the title to UTF-8. - * @param $title - * @return string (maybe contains invalid UTF-8 character) - */ - private function convertPdfEncodingToUTF8($title) { - // first try UTF-16 (then UTF-8) because its easier to detect its present/absence - foreach (array('UTF-16BE', 'UTF-16LE', 'UTF-8', 'WINDOWS-1252') as $encoding) { - if (mb_check_encoding($title, $encoding)) { - return mb_convert_encoding($title, 'UTF-8', $encoding); - } + // In one case (at least in tests), url is empty here + // so we set it using $url provided in the updateEntry call. + // Not sure what are the other possible cases where this property is empty + if (empty($entry->getUrl()) && !empty($url)) { + $entry->setUrl($url); } - return $title; - } - /** - * Remove invalid UTF-8 characters from the given string. - * @param String $rawText - * @return string - */ - private function sanitizeUTF8Text($rawText) { - if (mb_check_encoding($rawText, 'UTF-8')) { - return $rawText; - } - return iconv("UTF-8", "UTF-8//IGNORE", $rawText); + $this->stockEntry($entry, $content); } /** @@ -219,6 +184,59 @@ class ContentProxy $entry->setTitle($path); } + /** + * Try to sanitize the title of the fetched content from wrong character encodings and invalid UTF-8 character. + * + * @param $title + * @param $contentType + * + * @return string + */ + private function sanitizeContentTitle($title, $contentType) + { + if ('application/pdf' === $contentType) { + $title = $this->convertPdfEncodingToUTF8($title); + } + + return $this->sanitizeUTF8Text($title); + } + + /** + * If the title from the fetched content comes from a PDF, then its very possible that the character encoding is not + * UTF-8. This methods tries to identify the character encoding and translate the title to UTF-8. + * + * @param $title + * + * @return string (maybe contains invalid UTF-8 character) + */ + private function convertPdfEncodingToUTF8($title) + { + // first try UTF-8 because its easier to detect its present/absence + foreach (['UTF-8', 'UTF-16BE', 'WINDOWS-1252'] as $encoding) { + if (mb_check_encoding($title, $encoding)) { + return mb_convert_encoding($title, 'UTF-8', $encoding); + } + } + + return $title; + } + + /** + * Remove invalid UTF-8 characters from the given string. + * + * @param string $rawText + * + * @return string + */ + private function sanitizeUTF8Text($rawText) + { + if (mb_check_encoding($rawText, 'UTF-8')) { + return $rawText; + } + + return iconv('UTF-8', 'UTF-8//IGNORE', $rawText); + } + /** * Stock entry with fetched or imported content. * Will fall back to OpenGraph data if available. @@ -228,7 +246,7 @@ class ContentProxy */ private function stockEntry(Entry $entry, array $content) { - $entry->setUrl($content['url']); + $this->updateOriginUrl($entry, $content['url']); $this->setEntryDomainName($entry); @@ -294,6 +312,51 @@ class ContentProxy } } + /** + * Update the origin_url field when a redirection occurs + * This field is set if it is empty and new url does not match ignore list. + * + * @param Entry $entry + * @param string $url + */ + private function updateOriginUrl(Entry $entry, $url) + { + if (!empty($url) && $entry->getUrl() !== $url) { + $parsed_entry_url = parse_url($entry->getUrl()); + $parsed_content_url = parse_url($url); + + $diff_ec = array_diff_assoc($parsed_entry_url, $parsed_content_url); + $diff_ce = array_diff_assoc($parsed_content_url, $parsed_entry_url); + + $diff = array_merge($diff_ec, $diff_ce); + $diff_keys = array_keys($diff); + sort($diff_keys); + + switch ($diff_keys) { + case ['path']: + if (($parsed_entry_url['path'] . '/' === $parsed_content_url['path']) // diff is trailing slash, we only replace the url of the entry + || ($url === urldecode($entry->getUrl()))) { // we update entry url if new url is a decoded version of it, see EntryRepository#findByUrlAndUserId + $entry->setUrl($url); + } + break; + case ['scheme']: + $entry->setUrl($url); + break; + case ['fragment']: + case ['query']: + case ['fragment', 'query']: + // noop + break; + default: + if (empty($entry->getOriginUrl())) { + $entry->setOriginUrl($entry->getUrl()); + } + $entry->setUrl($url); + break; + } + } + } + /** * Validate that the given content has at least a title, an html and a url. *