If-None-Match and IMS (new Issue IMS_INM_MISMATCH)

This documents discussions on the implementation list that shows
a new protocol issue (IMS_INM_MISMATCH).

Now added to the issues list.
				- Jim Gettys

Forwarded message 1

  • From: Henry Sanders (Exchange) <henrysa@EXCHANGE.MICROSOFT.com>
  • Date: Wed, 5 Nov 1997 16:26:22 -0800
  • Subject: If-None-Match and IMS
  • To: w3c-http@w3.org
  • Message-Id: <2FBF98FC7852CF11912A00000000000107196961@DINO>
Is there a consensus on the correct behavior for a server when it
receives a request with conflicting If-None-Match and If-Modified-Since
headers, for instance where the I-N-M ETag is outdated but the I-M-S
date is good?

For example, take the following sequence

GET /foo.txt HTTP/1.1
Host: server.company.com

HTTP/1.1 200 OK
Last-Modified: Wed, 05 Nov 1997 22:10:48 GMT
ETag: "12345"
Content-Length: ...

<data>

GET /foo.txt HTTP/1.1
Host: server.company.com
If-None-Match: "12344"
If-Modified-Since: Wed, 05 Nov 1997 22:10:48 GMT

Assuming the Last-Modified date hasn't changed, what should the server
send back as a response to the 2nd request, 304 or 200? I would have
thought 200 is correct, since the ETag is invalid or out of date.
However the spec indicates that the I-N-M header is to be treated as if
it isn't present if the ETag doesn't match, and then the I-M-S would
lead to a 304 response. Also the paragraph about I-N-M being ignored if
the request would otherwise generate a non-2xx response would indicate
that a 304 is correct. But there would seem to be a window here, where
/foo.txt is modified twice in less than a second, the ETag is updated
but the L-M date isn't, and a client using an out of date ETag gets a
304 anyway. Or am I missing something here?

Henry

Forwarded message 2

  • From: Roy T. Fielding <fielding@kiwi.ics.uci.edu>
  • Date: Tue, 11 Nov 1997 07:56:17 -0800
  • Subject: Re: If-None-Match and IMS
  • To: "Henry Sanders (Exchange)" <henrysa@exchange.microsoft.com>
  • Cc: w3c-http@w3.org
  • Message-Id: <9711110807.aa04829@paris.ics.uci.edu>
Henry Sanders writes:
>Is there a consensus on the correct behavior for a server when it
>receives a request with conflicting If-None-Match and If-Modified-Since
>headers, for instance where the I-N-M ETag is outdated but the I-M-S
>date is good?

Hmmm, I don't know about consensus, but I implemented it using the
principle that an ETag test always overrides IMS.  I mentioned it as
an issue on http-wg a long time ago, since the RFC fails to specify
the ordering.  The Apache code (and comments) are below.

>For example, take the following sequence
>
>GET /foo.txt HTTP/1.1
>Host: server.company.com
>
>HTTP/1.1 200 OK
>Last-Modified: Wed, 05 Nov 1997 22:10:48 GMT
>ETag: "12345"
>Content-Length: ...
>
><data>
>
>GET /foo.txt HTTP/1.1
>Host: server.company.com
>If-None-Match: "12344"
>If-Modified-Since: Wed, 05 Nov 1997 22:10:48 GMT
>
>Assuming the Last-Modified date hasn't changed, what should the server
>send back as a response to the 2nd request, 304 or 200?

200.

>I would have
>thought 200 is correct, since the ETag is invalid or out of date.
>However the spec indicates that the I-N-M header is to be treated as if
>it isn't present if the ETag doesn't match, and then the I-M-S would
>lead to a 304 response. Also the paragraph about I-N-M being ignored if
>the request would otherwise generate a non-2xx response would indicate
>that a 304 is correct. But there would seem to be a window here, where
>/foo.txt is modified twice in less than a second, the ETag is updated
>but the L-M date isn't, and a client using an out of date ETag gets a
>304 anyway. Or am I missing something here?

Nope, that about sums it up.  200 is what the response should be, but
the RFC is less than clear.

....Roy
=====================================================================

API_EXPORT(int) meets_conditions(request_rec *r)
{
    char *etag = table_get(r->headers_out, "ETag");
    char *if_match, *if_modified_since, *if_unmodified, *if_nonematch;
    time_t mtime;

    /* Check for conditional requests --- note that we only want to do
     * this if we are successful so far and we are not processing a
     * subrequest or an ErrorDocument.
     *
     * The order of the checks is important, since ETag checks are supposed
     * to be more accurate than checks relative to the modification time.
     * However, not all documents are guaranteed to *have* ETags, and some
     * might have Last-Modified values w/o ETags, so this gets a little
     * complicated.
     */

    if (!is_HTTP_SUCCESS(r->status) || r->no_local_copy) {
        return OK;
    }

    mtime = (r->mtime != 0) ? r->mtime : time(NULL);

    /* If an If-Match request-header field was given
     * AND if our ETag does not match any of the entity tags in that field
     * AND the field value is not "*" (meaning match anything), then
     *     respond with a status of 412 (Precondition Failed).
     */
    if ((if_match = table_get(r->headers_in, "If-Match")) != NULL) {
        if ((etag == NULL) ||
            ((if_match[0] != '*') && !find_token(r->pool, if_match, etag))) {
            return HTTP_PRECONDITION_FAILED;
        }
    }
    else {
        /* Else if a valid If-Unmodified-Since request-header field was given
         * AND the requested resource has been modified since the time
         * specified in this field, then the server MUST
         *     respond with a status of 412 (Precondition Failed).
         */
        if_unmodified = table_get(r->headers_in, "If-Unmodified-Since");
        if (if_unmodified != NULL) {
            time_t ius = parseHTTPdate(if_unmodified);

            if ((ius != BAD_DATE) && (mtime > ius)) {
                return HTTP_PRECONDITION_FAILED;
            }
        }
    }

    /* If an If-None-Match request-header field was given
     * AND if our ETag matches any of the entity tags in that field
     * OR if the field value is "*" (meaning match anything), then
     *    if the request method was GET or HEAD, the server SHOULD
     *       respond with a 304 (Not Modified) response.
     *    For all other request methods, the server MUST
     *       respond with a status of 412 (Precondition Failed).
     */
    if_nonematch = table_get(r->headers_in, "If-None-Match");
    if (if_nonematch != NULL) {
        int rstatus;

        if ((if_nonematch[0] == '*')
            || ((etag != NULL) && find_token(r->pool, if_nonematch, etag))) {
            rstatus = (r->method_number == M_GET) ? HTTP_NOT_MODIFIED
                                                  : HTTP_PRECONDITION_FAILED;
            return rstatus;
        }
    }
    /* Else if a valid If-Modified-Since request-header field was given
     * AND it is a GET or HEAD request
     * AND the requested resource has not been modified since the time
     * specified in this field, then the server MUST
     *    respond with a status of 304 (Not Modified).
     * A date later than the server's current request time is invalid.
     */
    else if ((r->method_number == M_GET)
             && ((if_modified_since =
                  table_get(r->headers_in, "If-Modified-Since")) != NULL)) {
        time_t ims = parseHTTPdate(if_modified_since);

        if ((ims >= mtime) && (ims <= r->request_time)) {
            return HTTP_NOT_MODIFIED;
        }
    }
    return OK;
}

Forwarded message 3

  • From: Henry Sanders (Exchange) <henrysa@EXCHANGE.MICROSOFT.com>
  • Date: Tue, 11 Nov 1997 10:04:50 -0800
  • Subject: FW: If-None-Match and IMS
  • To: "'jg@pa.dec.com'" <jg@PA.DEC.COM>, "'Larry Masinter'" <masinter@parc.xerox.com>
  • Message-Id: <2FBF98FC7852CF11912A00000000000107196985@DINO>
Just what you wanted to see, another HTTP/1.1 issue....

This is something I sent out to the HTTP implementors list about a week
ago. Looks like the spec says to send a 304 when a 200 response would be
better. Any comments?

Henry

> -----Original Message-----
> From:	Henry Sanders (Exchange) 
> Sent:	Wednesday, November 05, 1997 4:26 PM
> To:	w3c-http@w3.org
> Subject:	If-None-Match and IMS
> 
> Is there a consensus on the correct behavior for a server when it
> receives a request with conflicting If-None-Match and
> If-Modified-Since headers, for instance where the I-N-M ETag is
> outdated but the I-M-S date is good?
> 
> For example, take the following sequence
> 
> GET /foo.txt HTTP/1.1
> Host: server.company.com
> 
> HTTP/1.1 200 OK
> Last-Modified: Wed, 05 Nov 1997 22:10:48 GMT
> ETag: "12345"
> Content-Length: ...
> 
> <data>
> 
> GET /foo.txt HTTP/1.1
> Host: server.company.com
> If-None-Match: "12344"
> If-Modified-Since: Wed, 05 Nov 1997 22:10:48 GMT
> 
> Assuming the Last-Modified date hasn't changed, what should the server
> send back as a response to the 2nd request, 304 or 200? I would have
> thought 200 is correct, since the ETag is invalid or out of date.
> However the spec indicates that the I-N-M header is to be treated as
> if it isn't present if the ETag doesn't match, and then the I-M-S
> would lead to a 304 response. Also the paragraph about I-N-M being
> ignored if the request would otherwise generate a non-2xx response
> would indicate that a 304 is correct. But there would seem to be a
> window here, where /foo.txt is modified twice in less than a second,
> the ETag is updated but the L-M date isn't, and a client using an out
> of date ETag gets a 304 anyway. Or am I missing something here?
> 
> Henry
> 

Forwarded message 4

  • From: Henry Sanders (Exchange) <henrysa@exchange.microsoft.com>
  • Date: Tue, 11 Nov 1997 10:00:12 -0800
  • Subject: RE: If-None-Match and IMS
  • To: "'Roy T. Fielding'" <fielding@kiwi.ics.uci.edu>
  • Cc: w3c-http@w3.org
  • Message-Id: <2FBF98FC7852CF11912A00000000000107196984@DINO>
> >that a 304 is correct. But there would seem to be a window here,
> where
> >/foo.txt is modified twice in less than a second, the ETag is updated
> >but the L-M date isn't, and a client using an out of date ETag gets a
> >304 anyway. Or am I missing something here?
> 
> Nope, that about sums it up.  200 is what the response should be, but
> the RFC is less than clear.
> 
	I agree that 200 is what the response should be, but saying the
RFC is unclear is an understatement. I think the RFC actually
contradicts this view, twice. I'll pass this on to the HTTP group, maybe
we can get it fixed in the update.

	Henry

Received on Tuesday, 11 November 1997 13:04:19 UTC