Re: problems with 1.1.1

G'day!

> From the Release notes:
> =============
> Note that the Max-Age in a client request takes the highest precedence.
> The 'MIN' value should normally be set to zero since it has higher
> precedence than the server's Expires: value. But if you wish to
> override the Expires: headers, you may use the MIN value.
> ==============
>
> Ok - I have done this, and it is now caching these pages, but doesn't this
> mean that pages with an "Expires" header are now also cached for this
> "minimum" time?

Correct. This is to allow the cache admin to override Expires settings; ie
sites with Expires == Date

> Also - from the documentation:
>
> if (CLIENT_MAX_AGE)
> if (AGE > CLIENT_MAX_AGE)
> return STALE
> if (AGE <= MIN_AGE)
> return FRESH
> if (EXPIRES) {
> if (EXPIRES <= NOW)
> return STALE
> else //See note below
> return FRESH
> }
> if (AGE > MAX_AGE)
> return STALE
> if (LM_FACTOR > PERCENT)
> return STALE
> return FRESH
>
> If I set "MAX_AGE" shouldn't it apply to objects with "EXPIRES" headers?
>
> ie:
> if (EXPIRES) {
> if (EXPIRES <= NOW)
> return STALE
> }
> if (AGE > MAX_AGE)
> return STALE
> if (LM_FACTOR > PERCENT)

The idea is to honour the Expires time, except for when it is set for less than
the min value of a refresh_pattern. If you have a max-age of a week, but a
server says the object expires in a month, well it's still fresh for a month,
why mark it as stale? I don't think there is much chance of a cache being
filled with expires objects greater than max-age.

Of course the algorithm is open to discussion :))

> I think that you need two seperate "minimum ttl" settings - one for objects
> with expires headers, one without. This way I can say "I don't want to
> cache objects without expires headers for a minimum time", but I do want to
> cache all other objects for at least 3 days.

You need to think in terms of 'fresh vs stale' not 'ttl.' If Expires times
where always valid, then this wouldn't be a problem. Sites tend to send real
short expires times as opposed to real long expires times.

> Thus a group that hosts a "sports" page with realtime updates does get
> retrieved every few minutes, even if I set "min" to 10 minutes.

Good point. Will your users be checking more often then every 10 minutes? Maybe
a seperate refresh_pattern for sports pages? Don't most of these pages use
Netscape's refresh meta tag or a cgi script that isn't cached anyways, as
opposed to expires headers? Is 10 minutes that bad ... listen to the radio if
they want real time :))

Welcome to the world of trying to please everyone all of the time!

Later
Ed

-- 
Ed Knowles aka Jasper				   Phone : +61 2 9385 4962
E-mail: ed@fatboy.geog.unsw.edu.au	           Fax   : +61 2 9313 7878
            What I lack in morals I make up for in principles.