Re: accessing cached objects direct from the cache

I vote yes...if it's neatly wrapped up in optional #ifdefs anyway. It'd
make disaster-recovery ("...but I accidentally deleted all my homepage
files off system X...but they're all in the cache. There's only 4302 of
them. You can get them back, can't you?") a _lot_ easier.

D

George Michaelson wrote:
>
> We've come up with a scheme to re-use the objects in the cache explicitly
> from the cache. It depends on exposing the hash function from the URL
> so a client can put in a GET cache_object: request in some efficient
> manner to let the cache send back the specific object.
>
> This has some 'social engineering' consequences, like bypassing the
> cookie and other access control on the real object. None-the-less,
> the data is in the cache, its known to be in the cache, and it could
> be bloody useful to be able to 'see' it direct from the cache.
>
> So:
> Does anybody else think it would be 'useful' to have a way
> to request a cache_object: form of URL which allowed the cache
> to return the content of another specific URL: form known to be
> in the cache?
>
> the cachemgr.c code seems to have 3/4 of whats needed. the unimplemented
> REMOVE function for instance, takes a url as an argument. I can see that
> it would work to display that url if in the cache. The idea to use the
> hash on the name is to bypass having to compute that inside the cache but
> if the cost saving is too low to be worth it, then it would be simple to
> just get the object.
>
> Why not just GET <url> I hear you ask? Because specific remote sites are
> now tagging their data so that the IMS request to the thing fails because
> associated headers like referrer are missing. I'm split over the merits
> of bypassing this kind of thing, but rather than pre-empt the issue I'd
> like to see some feedback from other cache-droids.
>
> cheers
> -George

-- 
Did you read the documentation AND the FAQ?
If not, I'll probably still answer your question, but my patience will
be limited, and you take the risk of sarcasm and ridicule.