5 Jun
2012
5 Jun
'12
9:47 a.m.
On Tue, Jun 05, 2012 at 06:45:58AM +0200, Harald Welte wrote:
I wish to create a thin wrapper that will allow me to use Freeswitch
as a MSC. Namely, I wish to perform subscriber management, call
routing, and SMS routing within Freeswitch. I also would like to have
visibility into the RR and MM aspects of the system, but I don't care
so much about controlling those.
Unless I'm mistaken, the MNCC socket system presently only allows for
call control.
Do you think there's a better way to accomplish this than working with
osmo-bsc and implementing the A-interface?
I'm
presently attemping to work with the code here:
<http://cgit.osmocom.org/cgit/erlang>.
What exactly are you trying to achieve? The code is currently not in a state that makes it
easy to use for
anyone not intimately familiar with it. It's mainly developed to
fulfill whatever current (customer) requirement I have, and has many
undocumented shortcomings and pitfalls.
Aye, that's about the impression I got as well. For now, my goal is
something robust enough to deal with a few hundred handsets, several
BTSes, one instance of osmo-bsc, and one instance of Freeswitch.
I hope it eventually will emerge into something more
generalized, but
until I get around writing some MAP applications (startin with a HLR),
it will probably remain in its current state.
Some of the code is used in production 24/7, a lot of other code has
never been running outside of my lab.
I made it happy to some extent this afternoon - I symlinked
~/osmo-erlang/signerl/* into ~/osmo-erlang/ and set
"ERL_LIBS=~/osmo-erlang", and most of the build system (erl -make,
rebar, etc) was able to find what it was looking for. I wasn't able
to replicate my success on my laptop, however, so I suspect my tree
wasn't actually clean.
--
.oO0Oo.
O _-_ O _____
O | | O /_,-,_\ Duncan Smith
O /__ __\ O / @ \ http://xrtc.net/f/
O - O +-----+
°°OOO°°
Is there something I'm missing? (I just
started working with
Erlang this week, so I may be missing something completely
obvious.) I think the main issue is the makefiles aren't putting
generated asn.1 code in places that Erlang expects to see headers
and code, and the makefiles don't chain together. :)
Anything but the signerl code is built with rebar, not with make. So I
guess if there are still any makefiles, they should be deleted from the
tree.