In article <004b01c19e15$607ea960$1d87bdac@ibmk62> you write:

> The backslash is a common problem ! Under the simple mail transfer protocol
> , it is supposed to be an 'erase preceding character" when found inside a
> header field! That dates back from the times when the authors of those
> protocols or RFCs had in mind slow , teletype-like , character oriented
> terminals , possibly operated by a human typist who would make typing
> mistakes and needed an erase character : the backslash AFA Smtp was
> concerned ( RFC 821 ) . Of course nowadays hardly one does type their emails
> on a teletype , but still the old RFC requirements ( and several generation
> of "sendmail" agents ) take the backslashes off . In principle , a single
> backslash should be represented as 2 backslashes , but wait,
> 
> Worse even , not all MTAs and trasmission systems act equally in presence of
> backslashes . I own one mailbox somewhere defined with a name in which two
> backslashes are included , as an anti-spam measure - and nobody (including
> me!)
> ever succeded in sending one email in there , either spam or not :) by
> whatever means , and I tried all the coding tricks I could think of based on
> the RFC !
> 
> Well sorry I must repeat  : backslashes in a header are a no-no , unless you
> really want trouble ...

I thought that the current SMTP RFC was rfc2821 and that makes no such 
claims. According to that, a backslash indicates a quoted character 
which is in line with all of the current RFCs I'm aware of.
You may well be right that in practice there are some non-compliant MTAs 
around, but is that any reason not to adhere to a standard ?

Alex.
-- 
   ____________________________    _______________________________
  (    Alex Venn               )  (   Success has many fathers,   )  
 (_)   [EMAIL PROTECTED]     (____)  but failure is an orphan.  (_)

Reply via email to