On 29/03/2008, Juliusz Chroboczek <[EMAIL PROTECTED]> wrote: > Polipo will always use a fresh connection for a POST or PUT request; > this is to avoid a race condition that is intrinsic in the HTTP protocol.
I can see that that makes sense for situations where the server may unexpectedly close the connection, and without any knowledge of the side effects of a POST request. > An HTTP server is allowed to shut down a persistent connection at any > time. This means in particular that it may shut down the connection > between the time a request is sent and the time it is received. Got it. > The solution is to use a reliable protocol, not HTTP, for RPC. I'm trying to make a bad system (which already uses HTTP for RPC) better here. I have complete control over the HTTP server and know what its behaviour is likely to be, moreover I know that resubmitting the POST request would not be very bad anyway in this particular case. Can you see anything that may go wrong. In any case this race will be very rare and I'm not at all concerned (in the context of our application) about it happening occasionally. I intend to try to hack it anyway. Mark ------------------------------------------------------------------------- Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://ad.doubleclick.net/clk;164216239;13503038;w?http://sf.net/marketplace _______________________________________________ Polipo-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/polipo-users
