- From: Henrik Nordstrom <henrik@henriknordstrom.net>
- Date: Tue, 13 Feb 2007 22:41:33 +0100
- To: HTTP Working Group <ietf-http-wg@w3.org>
- Message-Id: <1171402893.16208.59.camel@henriknordstrom.net>
[reposted to the mailinglist] tis 2007-02-13 klockan 12:16 +1300 skrev Adrien de Croy: > Actually it's not intended to abort it but make the client wait for a > 100 before sending, so it's a "please wait for 100 before continuing" > response rather than an abort. Ah, right. I misunderstood you completely. Sorry. Been working a bit too much with the problems of NTLM lately... So we have been speaking of two completely different problems. The flaws I pointed out in my last message is obviously false and from my own misunderstanding of the draft. Must have been very tired when I first read it. Thinking. Makes sense in the origin server case. But for an intermediary to send this it must also implement the timer and fake a 100 Continue which it is not allowed to do, or know already that the path is HTTP/1.1 before it responds with 102. Also if clients follows the specs correctly then in many cases the client already should wait indefinitely without an additional flow control. The timer in theory only applies when the client hasn't already seen a 100 Continue from the same server in response to another request. Regards Henrik
Received on Tuesday, 13 February 2007 21:41:48 UTC