Sorry, meant to just bring up the NNTP support to the dev list. Anyway, here goes with the other questions...
David Smiley wrote: > Another few unrelated questions/comments: > 1: What's so wrong about sending out single innocent DHCP UDP packet to > query the DNS server on startup? Is that really so bad? Ok, but then what? > 2: James detects that my machine's name is "dsmiley" but it doesn't also > realize that "dsmiley.mitre.org" is also my machine's name. Is it possible > that the RecipientIsLocal matcher could do the DNS lookup of the machine > name if it appears that the DNS name might resolve to the local James? Or > maybe the initial detection could be smarter to know at that time. Really I think you should be configuring the domains you handle rather than relying on DNS. I can't think of a way for you to determine every possible name that could map to your server... reverse DNS will just give a single address, and there's nothing to stop someone managing otherdomain.com to point an MX record at the hostname for your server... no real way to query to determine that ahead of time, and you probably don't want to allow that anyway. We could also do a dynamic check so that as you receive emails you do a lookup to see if this is a local host. However, again you may end up thinking someone is a local recipient when they're really not, and this is using what is transient information to make what should be permanent configuration changes, IMHO. In any case, remote delivery does do this check so it doesn't try to deliver to itself, to avoid race conditions. Sorry, I'm rambling. Summary: not possible to make a smarter initial detection; you're pretty much stuck configuring this. > 3: The "Handling mail for" info log messages in James.log confused me at > first. I think they ought to be worded > differently to clarify that these are machine names, not user names. I'll look into that. seems to make sense. > 4: Where is the NotifySender mailet documented? I looked for it here: > http://jakarta.apache.org/james/configuration_v2_0.html but didn't find it. > Ultimately, I am just curious to make sure that it will try to detect that > messages between two mail servers won't bounce back and forth forever if > there is a configuration problem. Not really documented anywhere... anyone is free to contribute documentation though. :) More seriously, James has a check (RelayLimit) set by default to 30 hops before a message gets killed, in case you unintentionally relay between 2 servers (although true, if you're constructing a new notice message, you could see this loop). > ~ David Smiley -- Serge Knystautas Loki Technologies - Unstoppable Websites http://www.lokitech.com/ -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
