On Thu, 10 Jul 1997, Marc van Selm wrote:
> Guess what I've found: an expires header on many of their pages with the
> expires time almost the same as "now"....
>
> So, MS doesn't want us to cache their pages and in this case squid does what
> it should (don't cache). Now you also know why Bill G. is talking about the
> electronic super highway because you are going to need it if all follow this
> lead....
>
> A good discussion though: What would be nice is some settings in the config
> to force a minimal caching-time and/or IMS based on url/site (also when the
> server tries to prevent caching via expires headers and users try to force a
> reload). This way we could improve the hit rate on low importance /
> frequently used sites.
This prob doesn't work, but you can get the idea:
refresh_pattern/i ^http://.*microsoft\.com/.*\.jpg 86400 50% 432000
refresh_pattern/i ^http://.*microsoft\.com/.*\.html 7200 50% 432000
refresh_pattern/i ^http://.*microsoft\.com/.* 3600 90% 172800
You get the idea.. You might also want to make it cache thing on msoft.com
that look like cgis.. As they use things like that for alot of regular
content.. Also look out for espn..
>
> Another thing I would like in squid is a setting to make squid cache files
> with cookie headers (also in those cases the files was downloaded via a
> parent/sibling without this setting and thus telling my squid to make it
> private) This also defeats the caching of squid. I'm now using a dirty hack
> in http.c to do this but its very rude...
> My reason for this is that cookies get more popular but squid doesn't cache
> these files. (I don't care that this defeats the cookie mechanism...)
>
> Marc
Also, we should make clients requesting Pragma: Nocache, IMS-Get but not
through the hierarchy.. (Or if we have to go through the hierarchy, then
we Pragma: Nocache)..
Received on Thu Jul 10 1997 - 06:19:24 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:35:43 MST