Linking to a page that has a large number of characters in its html source code causes the ajax request to fetch the metadata to fail, leaving the post status button disabled/grayed out and the user unable to post their status with that link.

An example problematic link is www.yahoo.com/food/buttermilk-fried-chicken-98838171056.html as well as many other articles on yahoo. The problem is caused by any site which exceeds the size limit set in the included library php simple html dom (included in components/com_community/libraries/vendor/simple_html_dom.php).

This problem can be addressed by overriding the value of the constant MAX_FILE_SIZE to something larger than the default of 600,000 characters. Ideally, the limit should be something harder to reach. In any case, Jomsocial should include error handling for failed parse attempts (specifically where used in the CParserMetas class (components/com_community/libraries/parsers/metas.php) in the _init method.

Hi, Greg.

I passed your report to our developers.
I'll update you with their response as ASAP.

Hi, Greg.

Please, make a safe copy of this file:

ROOT/components/com_community/libraries/parsers/metas.php

Then replace original file with patched one from package.

Please do a test and post about results on forum.

The patch does not work. I see it is designed to avoid the problem of exceeding the html parser's maximum page source length by stripping out all non-images from the page before trying to parse it. There is an error in the patch's call to preg_replace.

It should allow matching across multiple lines by using the /s modifier in the search pattern. So this line:

$content = preg_replace('/<body.+<\/body>/i', '<body>' . implode(' ', $matches[0]) . '</body>', $content);

$content = preg_replace('/<body.+<\/body>/is', '<body>' . implode(' ', $matches[0]) . '</body>', $content);

Hi, Greg.

Thnak you for reporting this. I passed it to our developer. It seems its just an omittance.