🚀 go-pugleaf

In article <705236596wnr@roborat.demon.co.uk>
           dave@roborat.demon.co.uk "Dave Scott" writes:

> 3) The question then is, would it be correct or desirable to apply the
> same modifications to an *outgoing* message - i.e. destined for another
> host whose storage conventions are unknown.
> 
> My sudden interest in this was sparked by a newly discovered bug in
> WinDIS. Basically it does *not* prefix any incoming message with a line
> beginning 'From ' *when* the line is actually part of the RFC822 headers
> - within the body it does prefix correctly.
> 
> This triggers in at least the following case; when forwarding a message
> between two users using the following software setup, and electing to
> *not* edit the message before forwarding :-
>      WinPmail <--> PM-Demon <--> WinDIS
> 
> Upon receipt of the forwarded message, it has an extra 'From ' embedded
> in it's headers, which cause it to be broken into 2 separate posts - the
> 1st part consisting of a few "Resent-from:" type headers, the 2nd being
> the original message without any forwarding information.

Hmm; that second "From " ought not to have been included in the forwarded
message AT ALL; if the forwarding mechanism merely prefixes Resent-From
type headers, then it has a DUTY to ensure that the remainder of the
headers are RFC-822 conformant: a line starting "From " is not, and
should either have been discarded, or rewritten as Return-Path in RFC-822
syntax.

-- 
Brian {Hamilton Kelly}                                         bhk@dsl.co.uk
        "The fundamental design flaws [of the products of the Micros^WSirius
        Cybernetics Corporation]  are completely hidden by their superficial
        design flaws."  SLATFATF 1983 --- wasn't Douglas Adams prescient?:-)

In article <705236596wnr@roborat.demon.co.uk>, Dave Scott
<dave@roborat.demon.co.uk> writes

[of >From]

>3) The question then is, would it be correct or desirable to apply the
>same modifications to an *outgoing* message - i.e. destined for another
>host whose storage conventions are unknown.

Yes it is desirable. Most MIME messages sent in quoted-printable will be
like this - the From will be sent as =46rom because otherwise there is a
risk of damage by intermediate systems

-- 
richard                      richard.clayton    @    T U R N P I K E .com
                                                     tel: +44 1306 732300
"Assembly of Japanese bicycle require great peace of mind" quoted in ZAMM

In article: <885046994snz@dsl.co.uk>  bhk@dsl.co.uk (Brian {Hamilton Kelly}) 
writes:
> 
> In article <705236596wnr@roborat.demon.co.uk>
>            dave@roborat.demon.co.uk "Dave Scott" writes:
> 
> > Upon receipt of the forwarded message, it has an extra 'From ' embedded
> > in it's headers, which cause it to be broken into 2 separate posts - the
> > 1st part consisting of a few "Resent-from:" type headers, the 2nd being
> > the original message without any forwarding information.
> 
> Hmm; that second "From " ought not to have been included in the forwarded
> message AT ALL; if the forwarding mechanism merely prefixes Resent-From
> type headers, then it has a DUTY to ensure that the remainder of the
> headers are RFC-822 conformant: a line starting "From " is not, and
> should either have been discarded, or rewritten as Return-Path in RFC-822
> syntax.

Thank you Brian! I was thinking about this from the wrong end, and the
real problem I need to fix is in the *incoming*, not outgoing messages.
What happens is this...

WinDIS downloads into the mbox file, each message prefixed with a 'From '
header. PM-Demon's mailin.exe program then copies each message with this
line intact, and it's this that get's preserved in forwarded messages.

OTOH, a quick test with the Pegasus built-in POP3 transport shows
messages received that way don't have this header. The solution
therefore seems to be to fix mailin to translate the 'From ' into a
Return-Path: as you suggest.

(I might still also perform the original modification of 'From ' to
'>From ' on *outgoing* messages I was asking about, *unless* anyone
knows of a reason not to?)

Dave.
-- 
    ()---()
     (0 0)    Dave Scott   e-mail: dave@roborat.demon.co.uk
----m-( )-m-------------------------------------------------------------
      =o=

In article <885046994snz@dsl.co.uk>, Brian {Hamilton Kelly}
<bhk@dsl.co.uk> writes

>Hmm; that second "From " ought not to have been included in the forwarded
>message AT ALL; if the forwarding mechanism merely prefixes Resent-From
>type headers, then it has a DUTY to ensure that the remainder of the
>headers are RFC-822 conformant: a line starting "From " is not, and
>should either have been discarded, 

agree 100% up to here!

>or rewritten as Return-Path in RFC-822
>syntax.

no! this would be totally wrong - you should NOT have Return-Path
headers on the wire in SMTP messages [ but you should in POP3 :) ]

basically the forwarder has to undo the "final delivery" before sending
the message onward. The old "From " should end up in the SMTP envelope
if anywhere at all.

-- 
richard                      richard.clayton    @    T U R N P I K E .com
                                                     tel: +44 1306 732300
"Assembly of Japanese bicycle require great peace of mind" quoted in ZAMM

In article <772157067wnr@roborat.demon.co.uk>, Dave Scott
<dave@roborat.demon.co.uk> writes

>WinDIS downloads into the mbox file, each message prefixed with a 'From '
>header. PM-Demon's mailin.exe program then copies each message with this
>line intact, and it's this that get's preserved in forwarded messages.
>
>OTOH, a quick test with the Pegasus built-in POP3 transport shows
>messages received that way don't have this header.

The existence of "From " is a local convention. It's up to an individual
program what it does. Of course to interwork with other programs on the
same machine, or to create files for import into other systems (not
necessarily local) then it might be desirable.

> The solution
>therefore seems to be to fix mailin to translate the 'From ' into a
>Return-Path: as you suggest.

No! (see other article)

>(I might still also perform the original modification of 'From ' to
>'>From ' on *outgoing* messages I was asking about, *unless* anyone
>knows of a reason not to?)

Deeply undesirable :(
X-From or X-Old-Envelope would be OK

-- 
richard                      richard.clayton    @    T U R N P I K E .com
                                                     tel: +44 1306 732300
"Assembly of Japanese bicycle require great peace of mind" quoted in ZAMM

In article <ToBFchASHzw0EAIM@turnpike.com>
   Richard Clayton <richard@turnpike.com> writes:
>> The solution
>>therefore seems to be to fix mailin to translate the 'From ' into a
>>Return-Path: as you suggest.
>
>No! (see other article)
>
>>(I might still also perform the original modification of 'From ' to
>>'>From ' on *outgoing* messages I was asking about, *unless* anyone
>>knows of a reason not to?)
>
>Deeply undesirable :(
>X-From or X-Old-Envelope would be OK

Thanks Richard and Dave for your comments.

I've just uploaded pmdem407.zip with a modified mailin which changes
the incoming "From " line to "X-From: From ...".

Outgoing "From " lines remain unchanged, as I've yet to see a real
problem with them.

Phil
-------------------------------------------------------------------
Phil Randal                                philr@hwcces.demon.co.uk
Environmental Services Department
Hereford and Worcester County Council
Worcester, UK

In article: <885237407snx@hwcces.demon.co.uk>  Phil Randal 
<philr@hwcces.demon.co.uk> writes:
> 
> Thanks Richard and Dave for your comments.
> 
> I've just uploaded pmdem407.zip with a modified mailin which changes
> the incoming "From " line to "X-From: From ...".

And thank you for such a quick fix to a problem that in all probability 
only myself (and 1 other) were actually suffering from :)

> Outgoing "From " lines remain unchanged, as I've yet to see a real
> problem with them.

Having read RC's comments, I take back what I'd suggested re this. It
would only have been a problem for WinDIS users receiving a forwarded 
message, and with the removal of the original 'From ' message separator 
this should not now be an issue.

Urrrm <sheepish look>, there is maybe one other small point worth 
mentioning. (this is without 1st checking PMD 4.07 for any changes in 
this area). Such a forwarded message (via Pegasus) acquires additional 
headers in the form of "Resent-from:", "Resent-to:" and "Resent-date:"

PM-Demon then attaches a new Message-ID: and a new Path: header also. 
Ideally shouldn't that be a "Resent-Message-ID:", to distinguish it from
the existing Message-ID already present? (and I've no idea about the 
Path:, but again presume there shouldn't normally be 2 such headers
present?)

AFAIK this is not causing any problems for anyone, and would be 
considerably more tricky to fix. Just thought as we were on the 
subject.... Sorry, I'll get me coat! :)

All the best,
Dave.
-- 
    ()---()
     (0 0)    Dave Scott   e-mail: dave@roborat.demon.co.uk
----m-( )-m-------------------------------------------------------------
      =o=

In article: <ToBFchASHzw0EAIM@turnpike.com>  Richard Clayton 
<richard@turnpike.com> writes:
> 
> In article <772157067wnr@roborat.demon.co.uk>, Dave Scott
> <dave@roborat.demon.co.uk> writes
> 
> > The solution
> >therefore seems to be to fix mailin to translate the 'From ' into a
> >Return-Path: as you suggest.
> 
> No! (see other article)

Oh, ok then I won't! (re-reads RFC822, ahhhh, I see...)
Rather glad now I was too lazy to start work on this yesterday :)

> >(I might still also perform the original modification of 'From ' to
> >'>From ' on *outgoing* messages I was asking about, *unless* anyone
> >knows of a reason not to?)
> 
> Deeply undesirable :(

<puzzle> but previously you seemed to indicate it *would* be desirable, 
or was that only in relation to MIME? (a subject I wish to steer well
clear of for the time being...) My original gut feelings told me this 
was the wrong approach to the problem, so I'm happy to have your 
confirmation it's not necessary.

Dave.
-- 
    ()---()
     (0 0)    Dave Scott   e-mail: dave@roborat.demon.co.uk
----m-( )-m-------------------------------------------------------------
      =o=