[Raw Msg Headers][Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: route to address



...
> >   The routes processing does not do it properly.  The result
> >   will not be processed thru routing analysis again, and it
> >   really works only for relaying entire domains thru named
> >   channels and gateways.
> > 
> >   Could you send me the  rrouter.cf  code lines you refer to above ?
> 
> Matti,  it seems to me that doing this thru fqdnaliases would be most
> convenient to administer.  Is it hard to allow substring or regex
> matching to fqdnaliases?

	It is hard, very hard!

	Now it does full string lookups from a database, and from directory
	( MAILSHARE/fqlists/ ).  The address is the lookup key.

	Doing regex matching would mean the database should be read
	sequentially thru for every address, and possible regexp in
	the first field would be matched against the address -- "database"
	would become sequence of matching rules.  I am not even sure
	it could be coded in ZMSH.   Perhaps something like:

	while read fqregexp fqdata < $MAILVAR/db/fqregexps
	do
		ssift "$address" in
		"$fqregexp"
			# we have a match, do the thing with $fqdata
			return ...
			;;
		tfiss
	done

	I am not even sure the file be closed in the case of the return
	in sample code.

	So far all of the patterns have been static strings, so I have
	no experience on a possibility of dynamic regexp stuff.

	OTOH, the  fqdnaliases.cf DOES separate localpart, and domain,
	but uses them together for all lookups.
	Therefore it could be feasible to write aliasing to allow
	super-domain ("pathalias") lookups too.  It just needs some
	updates at the 'relation' entry.

	Oh yes, reading  fqdnalias.cf   I did notice that it would treat
	addresses like   "user@domain.com"@domain.com  a bit poorly;
	namely at the moment  user='"user', host='domain.com"@domain.com'

> Eugene

/Matti Aarnio