[asterisk-dev] libpri

[Matthew Fredrickson]

Timo Teräs wrote:
> Hi,
> 
> Some months ago I filed bug 12357 about libpri enhancements for Q.SIG
> connected line updates. But nothing has happened there.

Hello.  First of all, I would like to apologize about the delay on 
review of your patch.  I have been spending a great deal of my time over 
the last few months getting BRI support into libpri, which has put a lot 
of additional feature review on hold.

I know that a lot of people would love to have more Q.SIG features 
working, so if you have any questions for me (I'm the maintainer of 
libpri), feel free to contact me about them.

> I'm also now planning to do more stuff: the
> rose_call_transfer_complete_decode is broken (I need that to receive
> connectedline updates during call), and of course writing the encode
> function is on list to send the updates during established call.
> 
> I might be also doing the so called Single Step Call Transfer (QSIG-SSCT as
> specified in ECMA-300). That is transferring of existing call to another
> number without additional calls (would make app Transfer work with QSIG).
> 
> For the above, I would need to do some ROSE / ASN.1 encoding and decoding
> functions. But the current ASN.1 stuff in pri_facility.[ch] is pretty
> awful. I'd like to clean up the macros and existing functions. So I'd like
> to talk about it in advance to know the new framework is good for all.

They're not great.  What we really need is a true ASN.1 parser/generator 
for all of that code.  The current scheme makes it non trivial to add 
new supplementary services, at least if you are not very familiar with 
ASN.1.

> And finally, there are two variants of how the ROSE operation ids are
> encoded: by local value (called sometimes ISO-QSIG) or by global OIDs
> (ECMA-QSIG). Currently libpri does only the local values. Having the global
> OIDs would increase interoperability in some cases. So I might be taking a
> look at that on some point.
> 
> So please, would the libpri people reply to me so I could coordinate my
> efforts with them.

Consider yourself replied to :-)

-- 
Matthew Fredrickson
Software/Firmware Engineer
Digium, Inc.