--On Thursday, 08 September, 2005 12:24 -0400 John Leslie
<john(_at_)jlc(_dot_)net> wrote:
Alex van den Bogaerdt <alex(_at_)ergens(_dot_)op(_dot_)het(_dot_)net> wrote:
rfc2821bis, as does rfc2821, contains "<CRLF>.<CRLF> where
the first CRLF is actually the end of the previous line. It
needs a kludge to cover the case of an empty mail, CRLF is
then the end of the DATA command.
We should be very hesitant to change this.
I have to agree with John Leslie here. I know one can get a
long way these days pretending that there are only Apple,
Linux/Unix, and Windows (alphabetical order) systems out there
and that, except for internal line-endings, they are all the
same. John and I are lucky enough to remember more interesting
times. But, even today, there are a number of interesting edge
cases. In particular, there are still systems around which want
to send LF or CR around as line-ends, rather than CRLF, and
enough others who want to _use_ bare LF or CR to carry
information in CRLF-terminated lines. From a standards
standpoint, we've pushed back as hard as we reasonably could on
that sort of thing, but there is still a good deal of code
around that tries to adjust for receiving nonsense in order to
get the mail through if possible.
Seemingly subtle changes like this could have all sorts of nasty
side-effects in the real world and we should, IMO, be extremely
cautious about making such changes unless there is strong
evidence of serious need for them.
john