RE: Re-checking disk cache

> what about a new -k option?
> would Duane be ok with that?

And how would you think to deliver the message to the running
process? AFAICS each -k option is mapped to a different
signal right now, and we'll soon (if we haven't already)
run out of signals.
This is why I suggested shmem+signal. Squid sets up a shmem
segment and leaves the ID somewhere (a file on disk or
something). The process running the -h fills it in and delivers
one signal asking the running server to check the message.
Or something similar.

-- 
	/kinkie