[Nice] New API and refactor

Rémi Denis-Courmont remi.denis-courmont at nokia.com
Wed May 14 23:55:56 PDT 2008

Le Wednesday 14 May 2008 21:30:03 ext Youness Alaoui, vous avez écrit :
> My first step now in making libnice 'nicer' will be to fix the
> repository by removing all those files that have nothing to do there...
> example :
> conncheck.h shouldn't be there
> there are some vi/emacs temporary files in the repository
> there are some Makefile files in the repository (not .am)
> I will then rename/create some files with more meaningful names and move
> functions around so it will become easier to know where to look for
> specific functions.

> Remi, if you have anything to suggest, please do so now.

I am not if stun-ice.* is a good name.

> I also wanted 
> to ask you what the stunbdc program was for ?

It's a command line client for STUNbis.
It's not really a "test" per se, more of a manual tool. But of course, it can 
be used for automated testing too (and it is). The name of the tool could be 

There is also stund, a small STUNbis binding discovery server there.

> is it just another misplaced test ?
> or does it actually need to be installed as a 'utility' to the library ?

It does not "need" to be, but it would nicely fit a "nice-utils" in a package, 
alongside say libnice0, libnice-dev(el), libnice-d(e)b(u)g...

> And how would you suggest we clean that file from the library files?
> would stun/lib and stun/apps directories be ok ? 

I don't mind. There are only two files (stund.c and stunbdc.c) for 
applications at the moment.

Another possible split is to separate the generic STUN core (message 
formatting, timers and the transaction layer) from the STUN usage binding 
discovery and connectivity checks usages (bind.* and stun-ice.*). It may be a 
bit premature though.

Rémi Denis-Courmont

More information about the Nice mailing list