On Thursday 18 July 96, at 1 h 5, the keyboard of Henrik Nordstrom 
<henrik.nordstrom@ida.his.se> wrote:
> Both should bypass the hierarchy since Pragma: no-cache is set. 
They do not. A Lynx reload does this on the Squid cache:
[17/Jul/1996:17:33:53 +0200] icp.c:1583:         parseHttpRequest: 
Complete request received
[17/Jul/1996:17:33:53 +0200] icp.c:651:  icp_hit_or_miss: GET 
<URL:http://www.eu.org/>
[17/Jul/1996:17:33:53 +0200] icp.c:668:  icp_hit_or_miss: REQ_NOCACHE = 
SET
[17/Jul/1996:17:33:53 +0200] icp.c:670:  icp_hit_or_miss: REQ_CACHABLE = 
SET
[17/Jul/1996:17:33:53 +0200] icp.c:672:  icp_hit_or_miss: 
REQ_HIERARCHICAL = SET
[17/Jul/1996:17:33:53 +0200] icp.c:716:  icp_hit_or_miss: TCP_REFRESH for 
'http://www.eu.org/'
[17/Jul/1996:17:33:53 +0200] icp.c:781:  icpProcessMISS: 'GET 
http://www.eu.org/'
REQ_HIERARCHICAL is SET which is not the case with Netscape, because of 
the IMS.
The routine which determines if the request can use hierarchy does not 
use "Pragma: No-Cache". A bug? IMHO, no. A reload should be transmitted 
and update the whole hierarchy. 
static int icpHierarchical(icpState)
     icpStateData *icpState;
{
    char *request = icpState->url;
    request_t *req = icpState->request;
    method_t method = req->method;
    wordlist *p = NULL;
    if (BIT_TEST(icpState->flags, REQ_IMS))
        return 0;
    if (BIT_TEST(icpState->flags, REQ_AUTH))
        return 0;
    if (method != METHOD_GET)
        return 0;
    /* scan hierarchy_stoplist */
    for (p = getHierarchyStoplist(); p; p = p->next)
        if (strstr(request, p->key))
            return 0;
    if (BIT_TEST(icpState->flags, REQ_LOOPDETECT))
        return 0;
    if (req->protocol == PROTO_HTTP)
        return httpCachable(request, method);
    if (req->protocol == PROTO_FTP)
        return ftpCachable(request);
    if (req->protocol == PROTO_GOPHER)
        return gopherCachable(request);
    if (req->protocol == PROTO_WAIS)
        return 0;
    if (req->protocol == PROTO_CACHEOBJ)
        return 0;
    return 1;
}
> Are you inside_firewall with more than one parent? If you are,
No. No firewalls at all. And just one parent (no neighbour).
Received on Thu Jul 18 1996 - 02:47:04 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:32:36 MST