- From: William A. Rowe, Jr. <wrowe@rowe-clan.net>
- Date: Fri, 05 Jan 2007 16:32:25 -0600
- To: Scott Lawrence <scott@skrb.org>
- CC: "Travis Snoozy (Volt)" <a-travis@microsoft.com>, Larry Masinter <lmm@acm.org>, "ietf-http-wg@w3.org" <ietf-http-wg@w3.org>
Scott Lawrence wrote: > > Yes we have. For a large body in a request, the server can return an > error; the client is required to abort sending the body if an error is > received. For a large body in a response, it's not lovely, but it's well > understood - the client just closes the connection. Since there is no > 3-way handshake, that's the best that can be done. +1, this is well understood and reasonably well documented. There is no need, and significant arguments against defining a range for DIGIT*1 as anything less than infinity, and it is up to the implementor to decide what it can parse, what it cannot parse yet still work around, or what it simply cannot work around. Disconnecting, in the third case, is the only remaining alternative... <snip> 8.1.4 Practical Considerations [...] A client, server, or proxy MAY close the transport connection at any time. For example, a client might have started to send a new request at the same time that the server has decided to close the "idle" connection. From the server's point of view, the connection is being closed while it was idle, but from the client's point of view, a request is in progress. This means that clients, servers, and proxies MUST be able to recover from asynchronous close events. Client software SHOULD reopen the transport connection and retransmit the aborted sequence of requests without user interaction so long as the request sequence is idempotent (see section 9.1.2). Non-idempotent methods or sequences MUST NOT be automatically retried, although user agents MAY offer a human operator the choice of retrying the request(s). Confirmation by user-agent software with semantic understanding of the application MAY substitute for user confirmation. The automatic retry SHOULD NOT be repeated if the second sequence of requests fails. Servers SHOULD always respond to at least one request per connection, if at all possible. Servers SHOULD NOT close a connection in the middle of transmitting a response, unless a network or client failure is suspected. Clients that use persistent connections SHOULD limit the number of simultaneous connections that they maintain to a given server. A single-user client SHOULD NOT maintain more than 2 connections with any server or proxy. A proxy SHOULD use up to 2*N connections to another server or proxy, where N is the number of simultaneously active users. These guidelines are intended to improve HTTP response times and avoid congestion. [...] 8.2.2 Monitoring Connections for Error Status Messages An HTTP/1.1 (or later) client sending a message-body SHOULD monitor the network connection for an error status while it is transmitting the request. If the client sees an error status, it SHOULD immediately cease transmitting the body. If the body is being sent using a "chunked" encoding (section 3.6), a zero length chunk and empty trailer MAY be used to prematurely mark the end of the message. If the body was preceded by a Content-Length header, the client MUST close the connection. [...] 8.2.4 Client Behavior if Server Prematurely Closes Connection </snip> (Note: no corresponding description of "Server Behavior if Client Prematurely Closes Connection"). I'm searching for any evidence of a behavioral requirement for Server Response that the Client "does not understand" and I'm coming up empty. So (s. 8.1.4 Practical Considerations) "A client, server, or proxy MAY close the transport connection at any time" is the final word. Unless I'm mistaken, if the client is unwilling to handle the server's response, the client MUST (gracefully? abruptly?) close the transport connection. And that's the beginning and end of the story for ALL RESPONSES AND RESPONSE HEADERS THAT THE CLIENT IS UNWILLING/UNABLE TO HANDLE. (*) (*) distinct from s7.1 "Unrecognized header fields SHOULD be ignored by the recipient" Of course, other means to avoid the condition in the first place; initially issuing a HEAD request against the desired resource can avoid this quandary but that's the choice of the user-agent's implementor, and doesn't avoid the situation for C-E: chunked. There is no need to proceed with this discussion. There's an opportunity to expand upon 8.1.4 offering that the client may disconnect with example reasons; or an opportunity to expand upon 8.2 for the converse of 8.2.4. The participants here, and the host of implementors to date, seem to have no problem understanding this from the existing spec. As 8.1.4 says, for whatever reason. You don't need one. If you aren't going to use the response, disconnect already. It's up to the implementor not to ignore overflows of ANY response header, and that includes DIGIT*1. Can we please move on?
Received on Friday, 5 January 2007 22:32:49 UTC