[squid-users] Re: squid cache prob: won't cache a 'pdf'

Eliezer Croitoru wrote:
> well i managed to make it being cached using specific rule.
> and your rule should do the trick
> but look at the difference between our rules:
> refresh_pattern -i ^http://www\.lsi\.com/.*AssetMgr\.aspx\?asset.* 4320
> 70% 10080
> leave the address ^^ alone
> refresh_pattern . 0 20% 4320 ignore-no-store ignore-no-cache
> ignore-private ignore-auth override-expire reload-into-ims
> ^
> your minimum time that you are using is 0 so you can try it for 2
> minutes also in the case you are breaking the http protocols.

---
Yeah, I could go for a minimum, but as you note, that pattern is a 'general
pattern' and I don't want to go breaking things I don't have to.
> i must tell you that a proxy with this kind of settings on the "." 
> pattern can lead to a lot of troubles for the users.
> so for for problematic sites that do not allow or want to be cached you 
> dont need  to make your whole server a mess of wrong refresh patterns.
---
Could -- but haven't in 10 years using that pattern...
> it's my line of thinking and it can also be a bug in the squid server 
> but i did mange to cache the file using the 3.2.0.5.
> and i think that also the older versions will do the trick on this 
> specific case.
----
	Well, that's what I was wondering -- I'll try a more specific
pattern, but the point was that something like those 'pdfs', I thought,
should just be cached!
	There's nothing special about them other than I happened to load
them more than once and wondered why it took so long for static content that
I thought, should have been cached.
	That's what got me to running my 'squidlog' monitoring script that produces the short output in the basenote.
	From there, I started massaging options -- trying to figure out why
it wasn't caching...
	Now, I guess -- it's down to 'must be some bug...'...   
	Been a while since I recompiled off the latest bzr, so maybe that's
the next step...