[asterisk-users] Limitations of IAX

[Andrew Kohlsmith]

On Wednesday 02 August 2006 10:29, Joshua Colp wrote:
> Writing it directly into the protocol is the dangerous part. You would need
> some control over who can set it and under what circumstances it is
> allowed. By encoding it into the dialed number only the users you control
> will be able to do it and the protocol doesn't need to be altered.

Nonsense; treat it like you'd treat Caller*ID; if you don't want to pass it 
over, explicitly zero it.   In fact, this is just plain old good programming 
common sense; do not assume anything implicitly; if you want a variable at a 
certain value (or not to have a certain value) then explicitly set it to that 
value.

This isn't hyperoptimized SSE2 does-a-billion-cycles superfast tight loop 
stuff; if a var gets set twice... oh well!  The peace of mind it gives is 
worth its weight in L2 cache transistors, and if you actually NEED this data 
transfered over... well it's there for you to use in all its splendor... no 
need to resort to ass-backward hacks like encoding things in the Dial() 
string and trying to parse it out on the other side, all the while worrying 
about escaping awkward values and trying to do so in a programming language 
slightly less functional than BASIC.

... I'm not attacking you, Joshua... and honestly if a patch giving this kind 
of function (with the ability to turn it off) will be accepted into trunk...  
I'll be all over this...  but generally what ends up happening is that no 
solution is accepted and the status quo is left in place, along with all of 
its problems.  I'm trying to help out, just as you are.

-A.