Re: [Sip-implementors] Call-ID header grammar changed

[Sachin Shenoy]

Hi, I got answer to (2) of the Call-Id problem. Keith Drage says   "Compatibility problems with anyone who implements an internet draft, i.e. work in progress, lie with whoever decided not to wait for the published RFC."   I checked up RFC 2543. RFC 2543 & 3261 are compable w.r.t. the spaces between tokens in call-id. But the number of character set that 3261 allows is more than that is allowed in 2543.   Thanks Sachin       ----- Original Message ----- From: Sachin Shenoy To: Sip-Implementors Sent: Monday, October 14, 2002 5:40 PM Subject: [Sip-implementors] Call-ID header grammar changed Hi, 1) New RFC allows a greater set of chracters to be present in the Call-Id tokens    than bis-05. So, should a new RFC compliant implementation take care not    to add these characters while generating call-id so that it can interop with    an earlier draft implementation? 2) Also, earlier call-id grammar allowed for LWS between the tokens, like    token LWS '@' LWS token. Now, according to the new RFC grammar this    LWS is not present.  Again over here, if an earlier implementation sends with    this, "Call-Id :  123  @ hostname", new implementation would fail to parse.      Looking at all these, should the new RFC clearly mention what an implementation    MUST be able to recieve & what it MUST be sending out?   Thanks Sachin   RFC 3261: Call-ID = ( "Call-ID" / "i" ) HCOLON callid callid = word [ "@" word ] word = 1*(alphanum / "-" / "." / "!" / "%" / "*" /      "_" / "+" / "`" / "'" / "~" / "(" / ")" / "<" / ">" /      ":" / "\" / DQUOTE / "/" / "[" / "]" / "?" / "{" / "}" ) Bis-05: Call-ID = ( "Call-ID" j "i" ) HCOLON callid 4523 callid = token [ ATSIGN token ] ATSIGN = LWS "@" LWS ; atsign