Hm, that's an interesting error. Judging by the undefined indices, I'd be inclined to think that it's possible you forgot to select a file, and/or something glitched with the captcha.
Either way, I like the facebook spider test idea. Though, I would think that their spider shouldn't trigger the download, because the way it's set up is each "download link" is just a page with the real download link on it, which submits a form that does all the processing to figure out if the file should be deleted.
I've also been having problems with the files actually staying until the set number of downloads, but it's something I'm working on. Thanks!
