I think there is a bug in aps_crawler.inc.php.
In line 223:
preg_match_all("/\<a href=\"(.+)\/\" class=\"vendor\"/i", $vendor_page, $matches);
You take the href from the result page, which was http://appscatalog.com/all-app/
. Something probably changed with listing, because that page returns the links as relative, so taking out the relative part of the url is not enough.
I made a fast dirty hack on my side, just modify the formatVendorCallback to return '/all-app/'.$array_item.'.atom?vendor='.$array_item.'&pageSi ze=100'; and it is almost ok.
But something still buggy, because the download of some packages simply fail with this error:
"[INTERFACE]: APS crawler: String could not be parsed as XML"
Maybe the package's atom feed is missing or something, but you can easily avoid that by check the returned page for xml validity, by adding this line
if (!simplexml_load_string($xml)) continue;
to the inner "for" loop where you fetch the packages. (Probably would be better to exception check with simpleXMLElement, but I didnt want to mess so much with that code, just want a working aps installer
Now, I can list and install packages without an issue