Subj : dosxnt To : Paul Williams From : Jasen Betts Date : Sat Sep 28 2002 06:54 am Hi Paul. 26-Sep-02 01:34:34, Paul Williams wrote to Rob Swindell PW> Hi Rob Swindell, hope you are having a nice day >>> The msgid I see above has extra stuff in the field that should >>> only PW> >>have the originating address and not the echotag and... whatever PW> the >>123. is for. RS>The 123 is a unique message number. It helps PW> to generate a guaranteed RS>unique msgid. This message (for PW> example) will have a different message RS>number. PW> I see. Reasonable thing to do. :> >>> It should look just like the reply kludge. PW> RS>Not *all* MSGIDs look exactly like what is described in PW> FTP-0009. Here RS>are some examples (parsed from backbone echoes): PW> 88 PW> Those however are all from internet based msgs and are kinda right PW> and kinda wrong at the same time. They're right in that the spec PW> says that the format is MSGID: and in the PW> case of the net the originating addy is quite a bit different than PW> a fido addy. PW> Oddly enough you might want to add and rescan the FTSC_PUBLIC echo PW> as msgid generation is a current topic. PW> One of the things brought up is that if a msgid is generated by a PW> fido system then the msgid needs to use a valid ftn address so PW> z:n/f.p is proper while email@domain.etc isn't quite right. PW> Another thing brought up which I didn't know abt is that the PW> uniqueness is applied on a per *echo* basis and not per system. PW> So it's possible for you to have a msg in different echos which PW> have the same msgid. The reason being that a dupe checker isn't PW> going to compare msgs that are in different echos, which is why PW> most spamming runs go thru. If the msg is repeated in a single PW> echo then all the dupes will be caught, but if you spread the same PW> msg w/ the same msgid across several echos then the dupes are PW> never seen by the dupe checker. Which is good because they're not dupes. sometimes it's desirable to put a message in more than one echo. spam needs a different detection mechanism. PW> Nothing at all wrong w/ that. You're trying to avoid the very PW> thing I had problems w/ several years ago when I was gating mail PW> between 2 echos and their respective mailinglists. Though my PW> problem was a system upstream ignoring the msgid's and only PW> looking at the msg headers which meant I had at times over half of PW> them being killed as dupes. :< Another feature to add to pktfix? (it'd be real easy to bump the timestamp up by one minute when same-looking headers are found) PW> Pick up ftsc_public and try to get a rescan of the last month or PW> so of msgs, as that's also part of the msgid discussion. The msgid PW> kludge is used by some programs to allow you to move a thread over PW> to another echo and have those following the thread be able to PW> find it. PW> Just how *that* works I haven't a clue, but that seems to be one PW> thing you can do w/ them. You can use it and reply like index keys to a database, to follow threads, some messagebase formats like Hudson (pity about the size limit) let you follow threads between different echoes. -=> Bye <=- --- * Origin: Drive defensively. Buy a tank. (3:640/531.42) .