[Top] [All Lists]

Re: ambiguities in RFC 2821 regarding end of mail data

2003-06-21 04:42:41

On Fri, 20 Jun 2003 18:10:22 -0400
"Richard O. Hammer" <ROHammer(_at_)earthlink(_dot_)net> wrote:

If Vladis and Keith are right (that the entire CRLF.CRLF should be 
dropped), then the final "line" of data would typically have no CRLF. 
  It would consist of a set of bytes ending without a CRLF.

Do you think that is right?

what you might see is ... cr lr cr lf . cr lf

in other words, the last line of the message really does end in cr lf,
and is followed by cr lf . cr lf.

if I'm writing code to send a message, I certanly don't want to trust
the message to end in cr lf, and I don't want to check either.  
I want to say

send ("data\r\n")
get (response)
send (message)
send ("\r\n.\r\n")
get (response)

though I'd be curious if anyone has tried to measure how many messages
received don't end in CR LF.