I just want to clarify that the download url's that you submit when publishing content need to be links to the actual .rfcmp file and not to an html page. For example, dropbox urls need "?dl=1" appended to them, or in most browsers you can right click on the blue download button and select copy link address. I went ahead and fixed all of the invalid dropbox urls, but there are still some bad links from other hosts still up there.
With my GetMod service I do some string work for Dropbox URLs. So a standard link like this: https://www.dropbox.com/s/61utzv46hllrx4o/GetModMgr.exe?dl=0 Would turn into this: https://dl.dropbox.com/s/61utzv46hllrx4o/GetModMgr.exe Then it becomes a direct link. That way the end user doesn't need to know how to make a truly direct link. The problem is from time to time Dropbox changes their URL scheme and I have to update my code again. And it isn't practical to code around every file hosting service out there. Luckily Dropbox hasn't change in a while and it seems to be the most popular service being used on my service.
I was unaware of dropbox's dl subdomain. I have been thinking about doing some heuristic suggestions based on the url. I've also considered making a head request to the url and validating based on the content-type header of the response. I'll probably implement the latter before the former, since as you said, there are many hosts out there and the urls are likely to change.
How would a mediafire link be written then example this is a link http://www.mediafire.com/download/enu9dmwx2bwxtbg/HiddenValley2015v1.10.rfcmp
has nothing to do with getmod, getmod works perfect, don't need a pro acc for it to work,......... this is about publish content URL for mediafire