// 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) {
- $title = $this->convertPdfEncodingToUTF8($title);
+ // 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 $this->sanitizeUTF8Text($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 (array('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);
+ $this->stockEntry($entry, $content);
}
/**
$entry->setTitle($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.
*/
private function stockEntry(Entry $entry, array $content)
{
- $entry->setUrl($content['url']);
+ $this->updateOriginUrl($entry, $content['url']);
$this->setEntryDomainName($entry);
}
}
+ /**
+ * 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) {
+ return false;
+ }
+
+ $parsed_entry_url = parse_url($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($diff_keys);
+
+ if ($this->ignoreUrl($entry->getUrl())) {
+ $entry->setUrl($url);
+ return false;
+ }
+
+ 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;
+ }
+ }
+
+ /**
+ * Check entry url against an ignore list to replace with content url.
+ *
+ * XXX: move the ignore list in the database to let users handle it
+ *
+ * @param string $url url to test
+ *
+ * @return bool true if url matches ignore list otherwise false
+ */
+ private function ignoreUrl($url)
+ {
+ $ignored_hosts = ['feedproxy.google.com', 'feeds.reuters.com'];
+ $ignored_patterns = ['https?://www\.lemonde\.fr/tiny.*'];
+
+ $parsed_url = parse_url($url);
+
+ $filtered = array_filter($ignored_hosts, function ($var) use ($parsed_url) {
+ return $var === $parsed_url['host'];
+ });
+
+ if ([] !== $filtered) {
+ return true;
+ }
+
+ $filtered = array_filter($ignored_patterns, function ($var) use ($url) {
+ return preg_match("`$var`i", $url);
+ });
+
+ if ([] !== $filtered) {
+ return true;
+ }
+
+ return false;
+ }
+
/**
* Validate that the given content has at least a title, an html and a url.
*
projects / github / wallabag / wallabag.git / blobdiff