On Wed, Feb 11, 2009 at 11:13 AM, Jay McCarthy <jay.mccarthy at gmail.com> wrote: > On Wed, Feb 11, 2009 at 10:01 AM, Robby Findler > <robby at eecs.northwestern.edu> wrote: >>> I don't recognize the error immediately, but I think it might be that >>> the continuation wasn't found for some reason. Perhaps the file was so >>> big the continuation was gone before the upload was complete. >> >> That doesn't seem good! If that were the case, it would mean that a gc >> that happens during an upload could potentially cause the upload to >> fail? > > The continuations are reclaimed independently from the GC. They are > always reachable, because I never know when a new request will come in > for them. (For example, someone could go to a site, write down the > URL, store it in a safety deposit box with instructions for their > grand-child to go to the site on their 50th birthday, etc.) > > It would be cool to do something GC-like by inspecting all current > requests during a reclamation phase, but it's not on my radar. (And > actually wouldn't fix this problem, because until the upload is > complete there is no request to observe.)
Oh -- so you were saying that the timeout of the continuation expired before the upload completed, then? That makes much more sense. Robby
