Figure out what's going on with PURGE

[GoogleCodeExporter]

The summary says it - figure out the code paths involved with PURGE.

There's three calls to clientPurgeRequest():

* one in clientCacheHit() after the store url and vary logic checks;
* one in clientProcessRequest();
* one in clientProcessMiss().

I think the multiple points for handling PURGE is because of the object 
processing pipeline with the lookups and processing handling occuring 
asynchronously.

But I think it needs to be properly understood and untangled. Eg - 
clientPurgeRequest() will do a cache lookup and swapin the metadata by 
calling clientCacheHit(). clientCacheHit() will then do whatever other 
checking is going on, then hopefully end up with the right object to feed 
clientPurgeRequest() again. That will hopefully end up completing the purge 
request. But there's a few places in the Vary handling where things fail or 
aren't in the cache - then it ends up calling clientProcessMiss() which 
will also call clientPurgeRequest() to complete the purge.

In any case, the logic is horrible, hairy, and deserves to be untangled as 
part of a client-side processing pipeline tidyup.

Original issue reported on code.google.com by adrian.c...@gmail.com on 14 Feb 2010 at 7:17