hi henrik, 1) refresh_pattern . 0 20% 4320 if i understand your note correctly, this is not correct for php-sites??? but it should, since the dot finds any object, doesn't it? nevertheless, i also tried refresh_pattern \.php 0 20% 4320 with the same effect (see 2)
2) if i do the request for test.php i see: HTTP/1.0 200 OK Daten: Sat, 15 Feb 2003 08:00:00 GMT Server: Apache/1.3.27 PHP/4.1.2 X-Powered-By: PHP/4.1.2 Content-Type: text/html X-Cache: MISS from test.de Connection: close hi-php (the last line is the content of the file) and still it is not in the cache does this help in finding an explanation? ----- Original Message ----- From: "Henrik Nordstrom" <[EMAIL PROTECTED]> To: "alp" <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]> Sent: Friday, February 14, 2003 8:04 PM Subject: Re: [squid-users] squid and php-sites > Two questions: > > 1. What is your refresh_pattern settings? > > 2. What is the full headers returend by your server? > > Just tested this with Squid-2.5 and a reply with only a Date header and > some content is cached if your refresh_pattern says it should be. > > > Note: The default refresh_pattern settings does not cache such replies > for the reasons indicated before. > > Regards > Henrik > > > alp wrote: > > > > sorry, i misunderstood your first reply. > > BUT: > > i have a site test.php (without any php-code, just for testing the suffix) > > on an apache server. > > it sends this site only with the DATE-header. no lastmod, no expires. it > > also does not mark the object as not cacheable. > > so the refresh-pattern IS used, as you say. > > > > so, first call: > > echo -e "GET /test.php HTTP/1.0\nHost:myhost\n\n" | netcat squidserver 80 > > gives the file together with the above header (date) > > second call: > > echo -e "GET /test.php > > HTTP/1.0\nHost:myhost\ncache-control:only-if-cached\n\n" | netcat > > squidserver 80 > > it says: object is not in cache. > > > > ??? > > doing the same with a file test.html i see the lastmod header and it is of > > course cached. > > > > i still seem to miss some important point in understanding this, i guess. > > but for me it seems as if the refresh-pattern is not used. >