sofia-sip packaging considerations

[George Danchev]

On Friday 16 June 2006 00:24, Mark Purcell wrote:
> On Friday 16 June 2006 05:43, Kai.Vehmanen at nokia.com wrote:
> > Hello Mark and others,
>
> Hi Kai,  Welcome to the list :-)
>
> [...]
>
> > >Also I wonder if we need some sort of upstream OpenSSL
> > >exception, such as
> > >http://en.wikipedia.org/wiki/OpenSSL_exception as sofia-sip
> > >links with SSL?
> >
> > Yes, we can take a closer look at this if needed. Hmm, we are
> > using LGPL, not GPL, does this problem apply to us?
>
> I know CUPS for one is licenced under LGPL and links with openssl and their
> upstream have provided the exception:
> http://packages.debian.org/changelogs/pool/main/c/cupsys/cupsys_1.2.1-2/cup
>sys.copyright
>
> > These references seem to indicate that LGPL+openssl would be ok:
> > http://lists.debian.org/debian-legal/2002/04/msg00024.html
> > http://lists.debian.org/debian-wnpp/2002/07/msg00154.html
>
> 2002 is a long way back for precedent on debian-legal. Can anyone comment
> on the contemporary view from debian-legal over linking LGPL & OpenSSL
> licences.

It seems to be fine to link LGPL code with OpenSSL, as suggested at:
http://lists.debian.org/debian-legal/2006/06/msg00289.html

> > But let me know if there is something we should do.
>
> It might be safer to get the exemption from upstream anyway, provided you
> are in a position to grant it..

Agreed. Kai, propbably it is a good idea to provide such an OpenSSL exception 
in your COPYRIGHTS file with your next release, e.g. 'that you explicitely 
allow linking your LGPL code with the OpenSSL library.'

Also reimplementing libsofia-sip-ua/ipt/rc4.c and libsofia-sip-ua/su/getopt.c 
is probably a good idea, as suggested at:
http://lists.debian.org/debian-legal/2006/06/msg00290.html

Here goes a little doxygen question, for sofia-sip upstream: do you consider 
having separately generated documentation for libsofia-sip-ua/docs/ and 
libsofia-sip-ua-glib/docs/ in your next release, since I have arranged two 
-doc package for these, but we can easily fallback to one of course. I think 
separation is a more fine grained approach though, since we can have more 
than -glib bindings in the future.

-- 
pub 4096R/0E4BD0AB 2003-03-18 <people.fccf.net/danchev/key pgp.mit.edu>
fingerprint 1AE7 7C66 0A26 5BFF DF22 5D55 1C57 0C89 0E4B D0AB