[erlang-questions] Memcached client implementation

[Christian S]

> 1) I'm using gen_server so that I can start it up and have a socket
> that remains open with the memcached server.  Does that seem like a
> good approach?  It's been several years since I've worked with Erlang
> and I'm a bit rusty.

It is not bad. What you get:

* A proven rpc mechanism, doing registered service lookup for you if needed.
* A simple callback for code-upgrade
* A suitable component to put into a supervisor for restart on failure,
you can handle memcached disconnects this way.

Also, it does not exclude having a pool of memcached connections.
However that is not something OTP helps you with that much unless you
go for random pick of a process in a pg process-group.

> 2) We would like to open source this code - what might be the best
> place for it?  I still have commit access to jungerl (isn't it time to
> switch to subversion, though?), so that's one possibility - anything
> else?

Jungerl is suitable.  If you would like to put some time into a svn
transfer then it would be great. Personally I don't see any aching
problems with CVS that SVN solves, I'm more appealed by using a
distributed version control, but that takes at least one person
actively pulling blessed patches into their source tree. In short, we
need a Linus Torvalds.

If you expose a git-repo you can lower the entry to people sending you
patches. Anyone can clone and improve locally using version control.

Of course, I recognize that there is an "emacs vs. vi"-like war
between git and mercury and other distributed version control systems.
_I_ am partial toward git, but not against using another distributed
version control. So we dont get into a dead-end discussion about
those.