> On Jan 16, 2008 12:14 PM, Adrian Chadd <[EMAIL PROTECTED]> wrote: >> Yeah, Pawel and i noticed this a few days ago. >> >> Could you look at the headers that bugzilla is returning and make >> sure its properly setting invalidation stuff? > > Here are the request: > GET /bugs/show_bug.cgi?id=2193 HTTP/1.1 > Host: www.squid-cache.org > > The response when first accessing the bug (with no attachments). > HTTP/1.1 200 OK > Date: Wed, 16 Jan 2008 14:08:21 GMT > Server: Apache/2.2.6 (FreeBSD) mod_ssl/2.2.6 OpenSSL/0.9.7e-p1 DAV/2 > PHP/5.2.5 with Suhosin-Patch > Cache-Control: max-age=86400 > Expires: Thu, 17 Jan 2008 14:08:21 GMT > Keep-Alive: timeout=5, max=100 > Connection: Keep-Alive > Transfer-Encoding: chunked > Content-Type: text/html > > The second response (when accessing the bug that has an attachment) > > HTTP/1.1 200 OK > Date: Wed, 16 Jan 2008 14:09:19 GMT > Server: Apache/2.2.6 (FreeBSD) mod_ssl/2.2.6 OpenSSL/0.9.7e-p1 DAV/2 > PHP/5.2.5 with Suhosin-Patch > Cache-Control: max-age=86400 > Expires: Thu, 17 Jan 2008 14:09:19 GMT > Keep-Alive: timeout=5, max=100 > Connection: Keep-Alive > Transfer-Encoding: chunked > Content-Type: text/html > > wireshark showed that the second response do list the attachment, but > Firefox did not show it. > > Seems that firefox shows an expired entity when it's age is less than > 'max-age'. > > Am I right? >
It seemed to me that it started happening here when I moved from the old "cache deny QUERY" config to the new refresh_patterns ones Adrian brought in for better dealing with dynamic content. If thats right, its an old bug just uncovered by the new handling method. Amos
