asterisk users - Sep 2008 - Channel variables "materializing" ...

I am trying to track a strange bug down, and need to ask a really stupid 
question, just so I can eliminate the possibility ..

When a SIP channel is hung up, I import a variable called "MEETMEROOM"
from the BRIDGEPEER channel, and if it is set, jump to another part of 
the dialplan.

[snip]
exten => h,1,ImportVar(PARKED=${BRIDGEPEER},MEETMEROOM)
exten => h,n,GotoIf($["${PARKED}" != ""]?end)
exten => h,n,goto(DialStatus,${DIALSTATUS},1)
exten => h,n(end),NoOp()
[snip]

There have been several occasions over the past couple of days where 
this variable has not executed the goto, and gone to the (end) label 
when I know for certain that the BRIDGEPEER channel does not have the 
variable set (I was able to duplicate the error once during a test phase 
when I was not setting the MEETMEROOM variable at all)

so, to the stupid question: If at some stage the BRIDGEPEER channel 
*has* had the MEETMEROOM variable declared, are there any circumstances 
at all where this variable may be transmitted to the next call that uses 
this channel.

There, I asked it. I don't believe that I just did. But there you have 
it. It's out in the open now ...

The only other thing that I was thinking of - if the PARKED variable was 
already set on the SIP channel, would an import of a non-existant 
variable from the BRIDGEPEER channel overwrite it, or keep it at the 
previous value ? Hmmm. Time to experiment.

Julian.

______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email 
______________________________________________________________________

Julian Lyndon-Smith wrote:
> I am trying to track a strange bug down, and need to ask a really stupid 
> question, just so I can eliminate the possibility ..
>
> When a SIP channel is hung up, I import a variable called
"MEETMEROOM"
> from the BRIDGEPEER channel, and if it is set, jump to another part of 
> the dialplan.
>
> [snip]
> exten => h,1,ImportVar(PARKED=${BRIDGEPEER},MEETMEROOM)
> exten => h,n,GotoIf($["${PARKED}" != ""]?end)
> exten => h,n,goto(DialStatus,${DIALSTATUS},1)
> exten => h,n(end),NoOp()
> [snip]
>
> There have been several occasions over the past couple of days where 
> this variable has not executed the goto, and gone to the (end) label 
> when I know for certain that the BRIDGEPEER channel does not have the 
> variable set (I was able to duplicate the error once during a test phase 
> when I was not setting the MEETMEROOM variable at all)
>
> so, to the stupid question: If at some stage the BRIDGEPEER channel 
> *has* had the MEETMEROOM variable declared, are there any circumstances 
> at all where this variable may be transmitted to the next call that uses 
> this channel.
>
> There, I asked it. I don't believe that I just did. But there you have 
> it. It's out in the open now ...
>
> The only other thing that I was thinking of - if the PARKED variable was 
> already set on the SIP channel, would an import of a non-existant 
> variable from the BRIDGEPEER channel overwrite it, or keep it at the 
> previous value ? Hmmm. Time to experiment.
>
> Julian.
>
> ______________________________________________________________________
>   
This may be a long shot but would it not be better to check to see 
whether or not the MEETMEROOM variable is defined before assigning it's 
value to another variable?  With just a cursory glance through the 
asterisk documentation I have available I don't see any indication of 
how asterisk variables behave if they are undefined. 

The other possibility I was considering is maybe BRIDGEPEER is not 
always being set to the correct channel?

Good luck,
-Brent