2011/8/4 Peter Krebs <[email protected]>: > I have a question regarding the ABNF of the "header" component of a > SIP/SIP-URI as defined in RFC 3261, page 223. It seems from the rule that it > is possible for a header to not have a value (more precisely, to have a > value of length 0, as there is no numeric value preceding the asterisk in > the rule for "hvalue" - RFC 2234 states 0 as the default value in this > case). However, the rule for "header" states that the "=" character is not > optional which would make a construct like "someheader=" possible (which is > rather unusual in my eyes). > Is my interpretation of the ABNF correct in this case and a SIP parser must > accept a header without a value while still checking for the "=" or did they > just forgot the "1" in the "hvalue" rule and a header must always have a > nonzero-length value (or did they forgot about square brackets around " "=" > hvalue") ? Or is it still too early and I'm unable to read ABNF this > morning? :D
Good point. I confirm that "=" after hvalue is mandatory and as per RFC 3261 BNF, the following SIP URI is valid: sip:qwe.com?qwe=qwe&asd= while this one is not valid: sip:qwe.com?qwe=qwe&asd I've confirmed it using my SIP parser which is 100% strict according to RFC 3261 BNF grammar. Indeed strange and a bit ugly but.... who is using URI headers? :) -- Iñaki Baz Castillo <[email protected]> _______________________________________________ Sip-implementors mailing list [email protected] https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
