Re: refresh_pattern based on Content-type: ?

Hi,

Lars Marowsky-Bree wrote:
> while tuning our proxy for maximum performance, I noticed it would be good
> to be able to base the refresh_pattern on the content-type.
>
> image/* for example often changes more seldom than text/html. application/*
> is even more static (most of the time).

Right now the StoreEntry struct which is used for checking
refresh patterns in refreshCheck() doesn't contain the content-type.
It may be possible to copy it in there, but that would consume more
memory for the internal object table. On the other hand if the
MemObject structure were present, the entire mime header is in there.

I'll have to do some debugging to verify this, though.

> It would be even better to be able to base refresh_times on an arbitrary
> header, (like Cache-control: pragma: check for php or something) and
> being able to use a scoring refresh_pattern (so that multiple expressions
> can affect the stale/fresh state), but this might get too complex to
> implement I guess... But I would be happy to be proven wrong ;-)

What do you propose for the scoring mechanism? An expression parser?
(yikes!)

-- 
miguel a.l. paraz  <map@iphil.net>                              +63-2-893-0850
iphil communications, makati city, philippines          <http://www.iphil.net>