Add shortened URL link label stripping to PageInfo::stripTrailingUrlFromBody

- Add test cases for shortened URL link labels
This commit is contained in:
Hypolite Petovan 2020-07-17 17:16:22 -04:00
parent da50456675
commit 8de66c0274
2 changed files with 33 additions and 5 deletions

View file

@ -252,22 +252,35 @@ class PageInfo
/**
* Remove the provided URL from the body if it is at the end of it.
* Keep the link label if it isn't the full URL.
* Keep the link label if it isn't the full URL or a shortened version of it.
*
* @param string $body
* @param string $url
* @return string|string[]|null
* @return string
*/
protected static function stripTrailingUrlFromBody(string $body, string $url)
{
$quotedUrl = preg_quote($url, '#');
$body = preg_replace("#(?:
$body = preg_replace_callback("#(?:
\[url]$quotedUrl\[/url]|
\[url=$quotedUrl]$quotedUrl\[/url]|
\[url=$quotedUrl]([^[]*?)\[/url]|
$quotedUrl
)$#isx", '$1', $body);
)$#isx", function ($match) use ($url) {
// Stripping URLs with no label
if (!isset($match[1])) {
return '';
}
return $body;
// Stripping link labels that include a shortened version of the URL
if (strpos($url, trim($match[1], '.…')) !== false) {
return '';
}
// Keep all other labels
return $match[1];
}, $body);
return rtrim($body);
}
}

View file

@ -108,6 +108,21 @@ class PageInfoTest extends MockedTest
'body' => '[url=https://example.com]link label[/url]',
'url' => 'https://example.com',
],
'task-8797-shortened-link-label' => [
'expected' => 'content',
'body' => 'content [url=https://example.com/page]example.com/[/url]',
'url' => 'https://example.com/page',
],
'task-8797-shortened-link-label-ellipsis' => [
'expected' => 'content',
'body' => 'content [url=https://example.com/page]example.com…[/url]',
'url' => 'https://example.com/page',
],
'task-8797-shortened-link-label-dots' => [
'expected' => 'content',
'body' => 'content [url=https://example.com/page]example.com...[/url]',
'url' => 'https://example.com/page',
],
];
}