[asterisk-dev] Time for a bug fix phase?

[John Lange]

Hey Murf, thanks for taking the time to do such a lengthy reply. I know
from researching this that you have done a lot of thinking about this
already so I'm very pleased to see you are still on top of it.

I've ready through your blog posting here a couple of times:

http://www.asterisk.org/node/48358

In the posting you outline some 21 scenarios that need to be tested.

Here in my simplistic view of the world (which means I don't have a deep
understanding of the inner workings of Asterisk) I think it boils down
to this; every time a channel is unbridged from something an event log
should be generated that shows (among other things) who the two
endpoints were and how long they were connected.

Looking through the scenarios I don't really see anything that leads me
to believe this wouldn't work.

To recreate the entire call you would have to piece together the log
entries.

Just to make sure I'm being clear, lets take a look at how this works
for an scenario with multiple attended transfers such as: A calls B, B
attended transfers to C, C Blind transfers to D. D talks to A, call
ends.

A calls B. A is effectively put on hold while B talks to C prior to the
transfer so the A-B bridge is torn down and a log event is generated for
the A-B "call".

The B-C also generates a log event when B stops talking to C when the
transfer happens AND the A-music also generates a log event when the
A-muisc is torn down to setup the A-C leg.

C then Blind transfers to D so a A-C log is written and then finally an
A-D log is written at the end of the call.

I have to go out so I didn't have time to recreate what this would look
like in the actual log entries but hopefully I'm making sense.

Anything wrong with this idea?

Regards,
-- 
John Lange
www.johnlange.ca