Re: always_direct not working

Yes, but once the squid has looked at it, the mime type could be part of
the cache record; the cache may store that it's seen this request before
but, since it's mime type is text/html and we're configured to pass
through a la

acl is_html mimetype_regex text/html

the requests are passed through not bassed on what the request _looks_
like but what it's proven to _be_ once it's been examined. Seems
reasonable to me, though admittedly I haven't the wherewithall now to code
a squid patch to enable this behavior :)
-Ian

Meanwhile, back at the ranch...
> > acl is_html urlpath_regex /$ \.html \? cgi-bin
> > no_cache deny is_html
> > always_direct allow is_html
> >
> > but that seems weak, is there a way to to acl requests on based on mime
> > type that I've overlooked?
>
> No, there's not a way, except guessing by the file name like you are
> doing. You can't tell the MIME type of an object until you get it --
> by then it's too late...

--
Salon Internet 				http://www.salon.com/
  HTTP mechanic, Perl diver, Mebwaster, Some of the above
Ian Kallen <idk@salon.com> / AIM: iankallen / Fax: (561) 619-0995