2010/5/31 Derrell Lipman <derrell.lip...@unwireduniverse.com>: > On Mon, May 31, 2010 at 13:36, Qoo Goo <qoo...@gmail.com> wrote: >> >> 2010/5/31 Derrell Lipman <derrell.lip...@unwireduniverse.com>: >> >> Yes, I agree. It's not the same calling a Response.End() than calling >> Response.Close(), for sure (I must admit that I haven't never used >> this method, I have to catch up who was!!). In fact, quoting VS >> built-in methods' descriptions, the first >> >> "Sends all currently buffered output to the client, stops execution of >> the page, and raises the System.Web.HttpApplication.EndRequest event" >> >> while Response.Close() simply >> >> "Closes the socket connection to a client." > > Ok, so that's telling you what the problem is. The latter does not send all > currently buffered output to the client, so the client just gets a > connection closed notification without receiving the data it expects to > receive.
Haha, interesting discussion. Yes. If you do not flushes data before.. I've seen these past days people out there saying (it's easy to find it googling) that the correct way to do things is: Response.Flush() Response.Close() Response.End() This would break ajax connections with Chrome 5. (This not means that I agree with this way of coding...) >> >> Anyway, I think this is a Chrome's bug because if you flushes >> explicitly the contents before closing the connection, every bit of >> data should arrive at client side before the connection is closed and >> in this situation I think the browser should definitely process data >> before considering that the connection is finished (because it was not >> closed when content was sent). > > Ah, but if the data is never flushed before closing the connection, it will > never arrive, right? :-) Yes, but we were flushing it! > > Here is the relevant section from RFC793 (TCP) pertaining to what happens on > the server and client sides of a close: > Yes, but is generally the OS who should handle all this stuff. The browser should only deal with higher layers. Just "give me data, pls". And it happens only with specific versions... Yes, I know: I am trying to make our bug smaller by trying to demonstrate it was caused in part by a hyper-giant Google's bug :-) ------------------------------------------------------------------------------ _______________________________________________ qooxdoo-devel mailing list qooxdoo-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/qooxdoo-devel
