At 09:41 PM 10/6/2003, Samir Srivastava wrote:
> The liberalism you have added is gone with other. If nobody is
>going
> to add / having anything of this sort, then only I think it is
>okay
> to add the presence of lr parameter checking.
For sake of brevity, I skipped the arguments and agreed with the conclusion.
In general, I think that a robust parser doesn't care about whether
parameters show up as ';parameter' or ';parameter=' or ';parameter=value'.
The first two cases should be handled as "present, with empty value".
Not only with record-routing -- we have seen all of the variations
with other parameters such as rport and to-tag frequently.
The record-routing logic should evaluate only the parameter presence.
("The lr parameter, when present, indicates that the element responsible
for this resource implements the routing mechanisms specified in this
document." S. 19.1, RFC3261)
The eye-striking case is than record-route with lr=off (thanks to Paul
for fingerpointing it) -- my robustness logic processes it as ;lr. I can
live with this unesthetical side-case better than with other alternatives.
-jiri
_______________________________________________
Sip-implementors mailing list
[EMAIL PROTECTED]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
