> Interesting test...Milt, what's your opinion on this. Good effort, in any > event!
I'm not Milt but I have an opinion. This most likely won't work in all circumstances. i) Once you write data and flush it, you can't write any headers back. ii) It may be that the container is using a Content-Length header (in which case it will have to buffer data anyway, whether you flush or not, although I admit this is unlikely) iii) Another problem is filters. If a filter is trapping upstream data writes, then no matter how many times you call flush the data won't get written back to the client (ditto for any wrapped response, it doesn't have to be a filter). You may be better sending '100 Continue' headers, although I've never tried this, and you may still not get any information back from the server if the client has gone away. In general, I would say there is no way of detecting the user stopping the request in all servers under all circumstances. IMHO it's one of those things that you just have to 'suck up' if you live in the world of HTTP, Kevin Jones Developmentor ___________________________________________________________________________ To unsubscribe, send email to [EMAIL PROTECTED] and include in the body of the message "signoff SERVLET-INTEREST". Archives: http://archives.java.sun.com/archives/servlet-interest.html Resources: http://java.sun.com/products/servlet/external-resources.html LISTSERV Help: http://www.lsoft.com/manuals/user/user.html
