I'm working with a small number of hearty souls on getting River2 to install smoothly on their machines. As always, I thought it should be a slam-dunk, because I am able to install it in seconds without problems, but a couple of users are having an amazing series of problems.
I asked one of them, Fred Zelders, to send me a copy of his OPML folder so I could see if I could reproduce the problems here. And they did show up here, so I could step through the process in the debugger, and see where it fails.
The problem appears to be in some code to prevent re-entrant calls to the scanner. No point in having two scans running concurrently. And it's getting false positives, that is, it thinks a scan is running when it completed long-ago.
I added code to make it more clear when a scan terminates, and set the test so it doesn't believe a positive result 20 minutes after the last scan started.
I have tested this code on two machines: 1. The one that's running Fred's installation. 2. My production server that maintains a set of rivers used at NYU and my own personal river. The scanning problem was solved by these fixes on the "Fred simulator" and continued to work on the production server.