(Fwd) Re: [Mailman-Users] Re: [Mailman-Developers] How to rem

[Bryan Carbonnell]

On 3 Apr 2004 at 21:51, Brad Knowles wrote:

> At 9:43 AM -0500 2004/04/03, Bryan Carbonnell wrote:
> 
> >  Why not just have a list of UserDefined RegEx's of headers to
> >  strip?
> 
>  Because most people don't properly understand regular 
> expressions.  Since you can do the same thing in a different way and
> have Mailman build the proper regular expressions based on user input,
> it seems pretty obvious what the correct solution is.

Yes, it may be true that RegExs are foreign to most people (myself 
included to a certain degree), but wouldn't it make MM more flexible? 
And wouldn't it be easier for Barry to implement?

>  If you have a "keep-these-headers/strip-these-headers" option, 
> then there shouldn't be any headers that either can't be stripped or
> kept, depending on your particular type of choice.

Well, would the X-Habeas headers be added as an option to strip, or 
the next X-Habeas type header that comes along? Or...

Do you see my point? If there are an absolutely fixed number of types 
of headers that could across, then I could see that going your way 
would work better, but since I can add any header to an outgoing mail 
that I want, with my e-mail client (like I did with this e-mail), 
then should the MM admins be given the opportunity to strip them with 
a RegEx?

> >  If you go with keep these/strip these, I'd vote for keeping
> >  everything and let the admin HAVE to make the change for
> >  themselves.
> 
>  IMO, we should default to safety.

Wouldn't safety be to not mess with the mail while passing through 
MM?

> >  Don't the RFCs say not to mess with the headers if possible? Or
> >  something like that?
> 
>  The RFCs regarding mail are intended primarily for MTAs and MUAs, not
> mailing list management systems.  Moreover, they make it clear that
> you can add headers, but you can't change them, and under certain
> circumstances you can remove them.  Since the MLM is a gateway system
> of sorts, there is more freedom with regards to what the authors are
> allowed to do.

I understand that the RFC are for MTA and MUAs, but shouldn't MM 
follow them as well? As well as we can anyway?

I'm not trying to argue, just trying to get thing straighened out in 
my mind.

-- 
Bryan Carbonnell - carbonnb at sympatico.ca
It was difficult to code. So it damn well better be difficult to use.