The web programming at Sourceforge has always been difficult, but I've noticed today when trying to download another package, that it has been "enhanced" to become completely useless. I'm not talking about accessing the site from a web browser. Instead, I am referring to doing package source downloads by automated scripts. I am assuming that the SF people don't want that to happen. So what would be needed is for developers that host on SF (I used to, but abandoned it because of the mess) to be properly informed that the download hosting provided does not accommodate automated build scripts, and that if they wish for their source project to be supported that way, they need to find other download hosting.

What concerns me is that most of the stuff hosted there is NOT FLOSS. Can you believe it. What's the point of source forge without the source ?

Are you referring to hosting non-free source there, or hosting projects entirely without source?

Both.

If you're talking about using wget or curl to download files hosted on Sourceforge, they've blogged instructions on how to do it:

http://sourceforge.net/blog/download-ninja/

Hmmm. It used to work like that, but doesn't, anymore. The contents is now just the HTML page for getting users to do downloads. I was dabbling around some and found that if I change the URL to a specific download site, then it's back to working. So this fails at "sourceforge.net" but works at "surfnet.dl.sourceforge.net". Maybe their different sites are running different software. I wonder if this will keep working.

While dabbling, I did find a number of URLs ended up at the same redirect URL, but in some cases got 404. That is, the very same URL (ultimate destination) sometimes got 404 and sometimes not, depending on where it was redirected from/through. Yikes! They are either making this just way too overcomplicated, or, as I suspect, trying to phase out automated downloads (like used in build scripts from some system/distro projects).

If it is their intention to allow automated downloads, why not just make plain direct URLs that just get the file contents and don't bother with fancy redirects (I'm not impressed by their leet skillz at web programming to make these redirects work).

I also don't like the fact that they don't just have a direct download link, and instead have this page that makes you wait a few seconds and thenstarts it automatically. What's the point of doing it that way?

To make you look at the ads that are displayed during those seconds.

I had a need to use lynx on sf.net a day or 2 ago.
Similar frustrations.

I have been an increasingly infrequent user for some time. Seems they want to piss-off users.

I suspect that non-interactive "build the world" scripts aren't in the market to buy stuff those ads want to show. They need to decide if they want to support the bandwidth for all the people running such scripts, or not. Then they simply need to be transparent about what they are willing to support and what not. Their servers, their network, their rules. But they really need to be honest about their intentions.

I know I long since abandoned SourceForge. There were simply better solutions, like github.

Uhm, the following works just fine for me:

Using Slackware 13.37 here.

This is the download link that Sourceforge gives you, with the GET parameters stripped out.

Don't know what was going on with Sourceforge over the last few days, but they seem to have gotten their act together.