Awesome! These small tools built on top of standard GNU tools are so refreshing.
Would be good to see some benchmarks on this! :)
Sorry. Benchmarks is too wrong to use in production. In fact, it is x200~1000+ slower than real memcached.
They won't look very good. I'm pretty sure the hash map data type that bash provides is of a fixed size, so performance will degrade pretty quickly as the number of cached items increases past that.
Now, it uses socat instead of ncat because socat is more simple than ncat and it can cut off lsof.
So cool!
Cool. Why?
Why ask "why" when the delicious question is "why not?"
I do honestly think this is cool. The readme does not make it clear if this is a joke (which would be fine), or has any practical use whatsoever (which would be awesome).
Potentially as a dev-environment-only drop-in for an application that uses Memcached in production?
It's not like full-blown Memcached is hard to set up locally, so this might be a bit of a stretch. ;-)
Because bash.
I wonder how many security holes this thing has. can != should.
It's obviously more oriented towards testing and stuff (as something compatible with memcached, but lighter) more than actual production use. Security holes are a bit less urgent in that case.