Hi Wietse,

Thanks for your reply:

>> I am not sure whether this is specified in an RFC, but I understand
>> that for every part of the remote MTA's response which starts with
>> "544-" there should be a new line in the message to the sender.
> 
> Error reports are standardized in 346[1-4]. I invite you to point
> out the requirement that ASCII art shall be preserved.

I understand that the requirements for the section of the error
message I am concerned with is the "Diagnostic code field":

 http://tools.ietf.org/html/rfc3464#section-2.3.6

= = = =

    . . .

    diagnostic-code-field =
          "Diagnostic-Code" ":" diagnostic-type ";" *text


   . . .

   Since the Status code
   will sometimes be less precise than the actual transport diagnostic
   code, the Diagnostic-Code field is provided to retain the latter
   information.  Such information may be useful in a trouble ticket sent
   to the administrator of the Reporting MTA, or when tunneling foreign
   non-delivery reports through DSNs.

   . . .

   In addition to the Diagnostic-Code itself, additional textual
   description of the diagnostic, MAY appear in a comment enclosed in
   parentheses.

   This field is optional, because some mail systems supply no
   additional information beyond that which is returned in the 'action'
   and 'status' fields.  However, this field SHOULD be included if
   transport-specific diagnostic information is available.

= = = = =

My understanding is that the MTA returns multiple lines, each starting
with the required diagnostic code.

Neither qmail or Postfix follows the RFC in terms of using brackets to
enclose any text following the diagnostic code.  Including such text is
optional anyway (MAY).

I agree there is no requirement about preserving formatting.  It seems
there is no requirement to handle multiple diagnostic codes following a
CR/LF.

It is reasonable to assume that these diagnostic codes and any text
which follows them is the MTA supplying information which is intended to
help a human.

I find qmail's approach of not altering the formatting of the MTA's
output to be more helpful than Postfix's approach of arbitrarily
reformatting this text, by ignoring newlines, by adding newlines
according to an 80 column wrapping limit and by inserting four
blank spaces at the start of every such reformatted line.

What benefits could there be to Postfix's reformatting of the
diagnostic information?

 - Robin

Reply via email to