[Devel] re: custom usrloc flags for registrar module

Bogdan-Andrei Iancu bogdan at voice-system.ro
Tue Dec 12 11:32:46 CET 2006


Hi Juha,

I would say the per-branch mask of flags will not affect at all the 
behaviour. Actually you will have 2 sets of flags:
    1) in message (saved by TM in transaction) - per-branch mask applies 
here
    2) in usrloc (saved by USRLOC module in location)

in my opinion there are two approaches for implementing the UL flags:
    A) use a new set of set/get/reset_flag() functions - in this case, 
the script writer will have to take care and map some msg flags to ul 
flags if he want to save them; the advantage is that the 2 sets of flags 
are independent and you do not need a msg flag for setting an UL flag
    B) have a mask over the msg flags telling which flags to be 
automatically saved in UL flags (maximum 16). it is more simple to use, 
but reduce the possibilities: 1) less flags (the UL flags will not be 
directly visible, but only as mapped on a msg flag) 2) permanent mapping 
between msg and ul flags - if you have a condition, you need to set the 
msg flags which will be automatically set the UL flag.....

now, going back to the nat_flag :
    1) if we let it as it is now, there will be problem as there is no 
mixture between the ul and msg flags.
    2) if you want to have it as dynamic flag (ul flag),  the nat_flag 
parameter will have to tell to reg/ul what UL flag should be used as nat 
flag - this is needed internally when fetching only natted records for 
pinging.

note that the UL flags will be visible only in the route type after 
location() and before save()

regards,
bogdan

Juha Heinanen wrote:

>i'm returning back to the usrloc flag issue.  i was proposing to add a
>couple of new flags that would behave the same way as nat_flag.  then
>bogdan proposed a more general solution with using new script functions
>set_reg_flag(x)/get_reg_flag(x).
>
>since nat_flag (and mediaproxy flag) needs to be from the range of flags
>that are branch specific, i started to wonder if
>set_reg_flag(x)/get_reg_flag(x) would work or if we need to go back my
>original proposal to have a couple of new nat_flag style flags that can
>(if necessary) be allocated from the branch flag space.
>
>-- juha
>
>_______________________________________________
>Devel mailing list
>Devel at openser.org
>http://openser.org/cgi-bin/mailman/listinfo/devel
>
>  
>




More information about the Devel mailing list