Thanks for the input, Steve! You know, I finally got the whole file yesterday once before I had to go somewhere, and it was after I had tried everything else. I ended up setting both the default timeout (normally 30) and the http timeout (normally none) to high values e.g. 9000. How long did it take you to download the file? It takes me around 30 minutes. I wrote a little loop that prints a period each time another 16k buffer full comes in and you can watch it march across the screen, then it will pause at random points for a couple of seconds or even for 15 or 20 seconds or longer, and then eventually it picks up again and goes for another while and etc. Also, I noticed yesterday that when I hit the end of the file it was just banging on returning empty buffer over and over, because I had never even reached the end of the file before because of the timeout problems I hadn't yet dealt with that happy condition of actually being done with the download. I don't know if the real connection inside rebol to the server has already been terminated at this point or what. I know that with a real tcp connection instead of the http connection you get a little closer to what's going on, and you can do wait [port] and test the status of the port to see if it's still open and so forth. I will just assume that the http is doing it's thing and does it's own wait automatically and returns an empty buffer when it's all done and the connection is closed. Is that right? Thanks for everybody's help with this! galt
