karl@ficc.uu.net (karl lehenbauer#) (09/08/88)
If one's mail had its route autogenerated by pathalias/smail, then one shouldn't mind if another site reroutes it, right? (assuming their maps are at least as up to date as yours) I mean, you've already implicitly said that you trust smail to do the routing for you. So shouldn't there be a way to differentiate between hand-routed (hard-routed) mail and machine-routed mail (a Routed-By line?), and machine-routed mail could be rerouted by anyone who thought they had a better path, while hard-routed stuff would be hands off, unless undeliverable as addressed. This way, one could still hammer out a route that works to a site that had a screwed-up map or path, or no map, in order to send mail to flame them into fixing it or whatever, but still get near-optimal routing (by pathalias' criteria) most of the time. On the other hand, if the maps are up to date, sites "down the line" are going to generate the same path smail did originally. Disclaimer: I'm new to this stuff. -- -- +1 713 274 5184, uunet!ficc!karl -- Ferranti International Controls, 12808 W. Airport Blvd., Sugar Land, TX 77478
vixie@decwrl.dec.com (Paul Vixie) (09/09/88)
In article <1426@ficc.uu.net> karl@ficc.uu.net (karl lehenbauer#) writes:
# If one's mail had its route autogenerated by pathalias/smail, then one
# shouldn't mind if another site reroutes it, right? (assuming their maps are
# at least as up to date as yours) I mean, you've already implicitly said that
# you trust smail to do the routing for you.
Nope. Sorry, I know it sounds like a good idea but there's a missing link:
my paths database was generated with some extra glue and real data that is
not part of the published maps (nor should it be!). I exert certain prefer-
ences and exploit certain private information when I build my paths database,
and I do not at all expect that anyone else will have the same preferences
or have access to the same information.
I tell pathalias to stay away from sites I know are having trouble this week.
I tell it to prefer a link between two remotes if I control both remotes and
know the reliability (and want to pay for that hop). I tell it about links
between remotes that most people won't know about because they are for private
use only.
When my mail gets rerouted, all that is lost.
So, while it sounds like a good idea to be able to tell when a path was
generated by a machine and reroute it, in fact _any_ path that comes out
of a machine you don't control may (will!) have been generated from data
which you don't possess. Or so you ought to assume.
I know you're new at this, and I hope you'll keep at it. We need to get and
keep lots of good people thinking about problems like these and looking for
and finding better solutions than the ones presently in use.
--
Paul Vixie
Work: vixie@decwrl.dec.com decwrl!vixie +1 415 853 6600
Play: paul@vixie.sf.ca.us vixie!paul +1 415 864 7013les@chinet.UUCP (Leslie Mikesell) (09/10/88)
In article <1426@ficc.uu.net> karl@ficc.uu.net (karl lehenbauer#) writes: >... So shouldn't there be a way to differentiate >between hand-routed (hard-routed) mail and machine-routed mail (a Routed-By >line?), and machine-routed mail could be rerouted by anyone who thought they >had a better path, while hard-routed stuff would be hands off, unless >undeliverable as addressed. I would rather see a standard way to explicitly request a site to re-route and everything else should be left alone if deliverable. Then news software running on machines that do not have the map data could pass replies off to the nearest site that knows how to route it. There are currently ways of making this happen, but they are not standardized and may depend on interpretation by intermediate sites. Les Mikesell
phil@amdcad.AMD.COM (Phil Ngai) (09/14/88)
In article <1426@ficc.uu.net> karl@ficc.uu.net (karl lehenbauer#) writes: >So shouldn't there be a way to differentiate >between hand-routed (hard-routed) mail and machine-routed mail (a Routed-By Why not route joe@site.uucp type addresses and leave sitea!siteb!sitec!joe address alone? -- I speak for myself, not the company. Phil Ngai, {ucbvax,decwrl,allegra}!amdcad!phil or phil@amd.com
dave@onfcanim.UUCP (Dave Martindale) (09/14/88)
In article <6548@chinet.UUCP> les@chinet.UUCP (Leslie Mikesell) writes: > >Then news software running on machines that do not have the map data could >pass replies off to the nearest site that knows how to route it. There are >currently ways of making this happen, but they are not standardized and may >depend on interpretation by intermediate sites. This is exactly what happens if you define INTERNET when you compile B news, and then set up "internet" in /usr/lib/news/mailpaths to point to a machine that will handle Internet-style addresses and is willing to do this for you. This works pretty well - we used it for a year here until I finally got around to installing smail.
les@chinet.UUCP (Leslie Mikesell) (09/14/88)
In article <16129@onfcanim.UUCP> dave@onfcanim.UUCP (Dave Martindale) writes: >>Then news software running on machines that do not have the map data could >>pass replies off to the nearest site that knows how to route it. There are >>currently ways of making this happen, but they are not standardized and may >>depend on interpretation by intermediate sites. >This is exactly what happens if you define INTERNET when you compile >B news, and then set up "internet" in /usr/lib/news/mailpaths to >point to a machine that will handle Internet-style addresses >and is willing to do this for you. > >This works pretty well - we used it for a year here until I finally got around to >installing smail. Yes, it can be done now, but as I understand things, the kludge that you use to pass the combination route/address depends on the parsing by the intermediate sites. RFC976 states that uucp gateways are supposed to take domain.dom!user and treat it like user@domain.dom, but if the site that you want to do the routing is not your uucp neighbor, then you need to know how to pass the request through the subsequent sites. Suppose you want siteC to route siteA!siteB!siteC!domain.dom!user or siteA!siteB!siteC!user@domain.dom. SiteA is your uucp neighbor which you would obviously know about. Doesn't this depend on siteB also using uucp style forwarding which could easily change without your knowledge since it is not a direct contact? Les Mikesell
chip@ateng.uucp (Chip Salzenberg) (09/15/88)
According to karl@ficc.uu.net (karl lehenbauer#): >If one's mail had its route autogenerated by pathalias/smail, then one >shouldn't mind if another site reroutes it, right? (assuming their maps are >at least as up to date as yours.) No one's maps are "up to date". The only reason I trust my smail database as far as I do -- which isn't far -- is because I've manually marked as DEAD some published links which don't work. I cannot assume that all other sites have done likewise. -- Chip Salzenberg <chip@ateng.uu.net> or <uunet!ateng!chip> A T Engineering My employer may or may not agree with me. The urgent leaves no time for the important.
roberts@edsews.EDS.COM (Ted Roberts) (09/16/88)
In article <6581@chinet.UUCP>, les@chinet.UUCP (Leslie Mikesell) writes: > > Suppose > you want siteC to route siteA!siteB!siteC!domain.dom!user or > siteA!siteB!siteC!user@domain.dom. These are not the same thing. The first route would send to siteA, then siteB, then siteC and siteC would route to user@domain.dom which they would know how to do since they know how to resolve domain addresses (you hope:-). The second would attempt to send to domain.dom, then to siteA, then to siteB, then to site C, then to user. This is because the "@" syntax takes precedence over the "!" syntax. > SiteA is your uucp neighbor which > you would obviously know about. Doesn't this depend on siteB also > using uucp style forwarding which could easily change without your > knowledge since it is not a direct contact? Maybe I don't understand what you mean here, but if you're saying that they will all of a sudden and without warning start trashing routes that use the "!" syntax then the answer is no. I'm not sure what you mean by "uucp style forwarding", do you mean versus being directly on the internet? If so, I would assume that they would also be able to handle that syntax. -- Ted Roberts | My opinions are not necessarily those EDS Technical Services Division | of my employer. Does that mean I'm UUCP: roberts@edsews.EDS.COM | wrong?
les@chinet.UUCP (Leslie Mikesell) (09/16/88)
In article <2105@edsews.EDS.COM> roberts@edsews.EDS.COM (Ted Roberts) writes: >> Suppose >> you want siteC to rout siteA!siteB!siteC!domain.dom!user or e >> siteA!siteB!siteC!user@domain.dom. >These are not the same thing. The first route would send to siteA, then >siteB, then siteC and siteC would route to user@domain.dom which they >would know how to do since they know how to resolve domain addresses >(you hope:-). The second would attempt to send to domain.dom, then to >siteA, then to siteB, then to site C, then to user. This is because the >"@" syntax takes precedence over the "!" syntax. If siteA and siteB are vanilla uucp mailers they will forward both forms to siteC because they won't know anything about the "@" syntax. >> SiteA is your uucp neighbor which >> you would obviously know about. Doesn't this depend on siteB also >> using uucp style forwarding which could easily change without your >> knowledge since it is not a direct contact? >Maybe I don't understand what you mean here, but if you're saying that >they will all of a sudden and without warning start trashing routes that >use the "!" syntax then the answer is no. I'm not sure what you mean by >"uucp style forwarding", do you mean versus being directly on the internet? >If so, I would assume that they would also be able to handle that syntax. Uucp style forwarding looks only at the next host, left to right and expects "!" as the separator. As you noted, if siteB changes from vanilla uucp to running smail, sendmail, or becomes an internet site, the interpretation of the "@" notation will change. Will the notation: siteA!siteB!siteC!domain.dom!user continue to work if siteB becomes an internet site and uses that instead of the uucp connection to siteC? Les Mikesell
wcf@psuhcx.psu.edu (Bill Fenner) (09/18/88)
In article <2105@edsews.EDS.COM> roberts@edsews.EDS.COM (Ted Roberts) writes: |In article <6581@chinet.UUCP>, les@chinet.UUCP (Leslie Mikesell) writes: |> |> Suppose |> you want siteC to route siteA!siteB!siteC!domain.dom!user or |> siteA!siteB!siteC!user@domain.dom. | |These are not the same thing. The first route would send to siteA, then |siteB, then siteC and siteC would route to user@domain.dom which they |would know how to do since they know how to resolve domain addresses |(you hope:-). The second would attempt to send to domain.dom, then to |siteA, then to siteB, then to site C, then to user. This is because the |"@" syntax takes precedence over the "!" syntax. Who says? I don't think any of the RFC's mention ! syntax at all. If your site is a smart site, running sendmail or smail, then @ will usually have precedence over !. but UUCP-only sites only understand !'s, and only look for !'s. To answer Leslie's question in the first place - inews will change siteA!siteB!siteC!user@domain.dom to ..siteC!domain.dom!user, to get around this exact problem... it changes the address to all !'s, if it needs to. | |> SiteA is your uucp neighbor which |> you would obviously know about. Doesn't this depend on siteB also |> using uucp style forwarding which could easily change without your |> knowledge since it is not a direct contact? | |Maybe I don't understand what you mean here, but if you're saying that |they will all of a sudden and without warning start trashing routes that |use the "!" syntax then the answer is no. I'm not sure what you mean by |"uucp style forwarding", do you mean versus being directly on the internet? |If so, I would assume that they would also be able to handle that syntax. What (I think) Leslie means is, are all the sites on the way going to handle this address the same way I expect them to (I.E. will they all send to the next host on the bangpath, or will one of them see the @ and send to domain.dom??) -- Bitnet: wcf@psuhcx.bitnet Bill Fenner | "Ain't got no cash, Internet: wcf@hcx.psu.edu | Ain't got no style UUCP: {gatech,rutgers}!psuvax1!psuhcx!wcf | Ain't got no girls Fido: Sysop at 263/42 (814/238 9633) \hogbbs!wcf| To make me smile"
dave@onfcanim.UUCP (Dave Martindale) (09/19/88)
In article <6581@chinet.UUCP> les@chinet.UUCP (Leslie Mikesell) writes: > >Yes, it can be done now, but as I understand things, the kludge that you >use to pass the combination route/address depends on the parsing by the >intermediate sites. RFC976 states that uucp gateways are supposed to >take domain.dom!user and treat it like user@domain.dom, but if the site >that you want to do the routing is not your uucp neighbor, then you need >to know how to pass the request through the subsequent sites. Ok, let me be more complete. The distributed B news software provides the INTERNET flag plus the internet line in the mailpaths file to help you. To make use of it, you have to find a site that handles Internet addressing and is willing to perform this service for you. To do this, you have to talk to the system administrator there. While talking, you can probably either determine a route from you to them that will not corrupt adresses, or you can establish a direct connection. Then your problems are solved. (Note that a site which can handle Internet addresses is not necessarily on the Internet itself - it's probably your nearest well-managed major UUCP site.) I realize that sometimes you may not be able to find someone to handle your mail this way, but given that you have found such a machine, finding a "safe" path to it or just setting up a direct connection should seldom pose a problem.
david@ms.uky.edu (David Herron -- One of the vertebrae) (09/19/88)
In article <915@psuhcx.psu.edu> wcf@psuhcx (Bill Fenner) writes: >In article <2105@edsews.EDS.COM> roberts@edsews.EDS.COM (Ted Roberts) writes: >|In article <6581@chinet.UUCP>, les@chinet.UUCP (Leslie Mikesell) writes: >|> Suppose >|> you want siteC to route siteA!siteB!siteC!domain.dom!user or >|> siteA!siteB!siteC!user@domain.dom. >|These are not the same thing. The first route would send to siteA, then >|siteB, then siteC and siteC would route to user@domain.dom which they >|would know how to do since they know how to resolve domain addresses >|(you hope:-). The second would attempt to send to domain.dom, then to >|siteA, then to siteB, then to site C, then to user. This is because the >|"@" syntax takes precedence over the "!" syntax. >Who says? I don't think any of the RFC's mention ! syntax at all. Ever hear of RFC-976? It talks about how the UUCP network works and a couple of ways of inter-operating with the Internet. It's all FULL of !'s. >If your >site is a smart site, running sendmail or smail, then @ will usually have >precedence over !. but UUCP-only sites only understand !'s, and only look >for !'s. To answer Leslie's question in the first place - inews will change >siteA!siteB!siteC!user@domain.dom to ..siteC!domain.dom!user, to get around >this exact problem... it changes the address to all !'s, if it needs to. You're partly right. I think there are sites out there which are sensitive to the route the mail came through. Such a site would give precedence to ! on mail coming over a UUCP link and to @ on mail coming over other sorts of links. Personally I don't understand why they would want to make it different like that, but the RFC doesn't specify a precedence. Instead the RFC says to avoid the issue altogether by using either pure user@domain forms or host!host!domain!user forms, and not to mix. -- <---- David Herron -- One of the MMDF guys <david@ms.uky.edu> <---- ska: David le casse\*' {rutgers,uunet}!ukma!david, david@UKMA.BITNET <---- What does the phrase "Don't work too hard" <---- have to do with the decline of the american 'work ethic'?
allbery@ncoast.UUCP (Brandon S. Allbery) (09/24/88)
As quoted from <2105@edsews.EDS.COM> by roberts@edsews.EDS.COM (Ted Roberts): +--------------- | In article <6581@chinet.UUCP>, les@chinet.UUCP (Leslie Mikesell) writes: | > | > Suppose | > you want siteC to route siteA!siteB!siteC!domain.dom!user or | > siteA!siteB!siteC!user@domain.dom. | | These are not the same thing. The first route would send to siteA, then | siteB, then siteC and siteC would route to user@domain.dom which they | would know how to do since they know how to resolve domain addresses | (you hope:-). The second would attempt to send to domain.dom, then to | siteA, then to siteB, then to site C, then to user. This is because the | "@" syntax takes precedence over the "!" syntax. +--------------- WRONG!!! "@" and "!" are used by different networks -- there is NO defined precedence between them! (You're comparing apples to oranges.) UUCP sites give "!" precedence, Internet domain mailers give "@" precedence. On ncoast (a UUCP system) "!" has precedence. Our neighbor "hal" gives "@" precedence because they are on the Internet. This can be quite useful; if I want to send mail across the Internet I can mail to hal!foo@bar.COM, thus overriding the UUCP route (which may well be slower) that smail would give me from a straight "foo@bar.COM". On your system, you would probably want to switch to UUCP only after you got as "close" to the recipient's system as possible, so you would want "@" to have higher precedence. Thus, the current system is useful for both of us. Just don't assume that everyone's mailer handles things the same way, as we *are* on different networks. Things get even more interesting when you try to apply your rules to a DecNet network or to the Bitnet as well (on Decnet, does "::" have precedence above or below "@"? How about "!"? Does Decnet even *care* what the relative precedence of "!" and "@" is?) ++Brandon -- Brandon S. Allbery, uunet!marque!ncoast!allbery DELPHI: ALLBERY For comp.sources.misc send mail to ncoast!sources-misc "Don't discount flying pigs before you have good air defense." -- jvh@clinet.FI
vixie@decwrl.dec.com (Paul Vixie) (09/25/88)
Branden, you and I almost always agree in this group. So consider this a small nit brought up by a copatriot: # UUCP sites give "!" precedence, Internet domain mailers give "@" precedence. The terminology is confusing. Vixie.sf.ca.us is a "UUCP site" if you mean that all my connections to the world are over UUCP links. However, I run smail and sendmail there, and I definitely give "@" a higher precedence than "!". In fact, by virtue of my dotted name, I can be considered to be "on the internet", though I am not "directly connected to the internet". So, okay, this is a minor nit. But terminology is often learned in context, and apparently a large number of novices read this group. I'd like to get them started out on the right foot. Cf. my recent "crucial distinction" article, wherein I lament that people defend "rerouting" without knowing what parts of it they are actually depending on. -- Paul Vixie Work: vixie@decwrl.dec.com decwrl!vixie +1 415 853 6600 Play: paul@vixie.sf.ca.us vixie!paul +1 415 864 7013
moore@utkcs2.cs.utk.edu (Keith Moore) (09/28/88)
In article <12604@ncoast.UUCP> allbery@ncoast.UUCP (Brandon S. Allbery) writes: >As quoted from <2105@edsews.EDS.COM> by roberts@edsews.EDS.COM (Ted Roberts): >+--------------- >| In article <6581@chinet.UUCP>, les@chinet.UUCP (Leslie Mikesell) writes: >| > >| > Suppose >| > you want siteC to route siteA!siteB!siteC!domain.dom!user or >| > siteA!siteB!siteC!user@domain.dom. >| >| These are not the same thing. The first route would send to siteA, then >| siteB, then siteC and siteC would route to user@domain.dom which they >| would know how to do since they know how to resolve domain addresses >| (you hope:-). The second would attempt to send to domain.dom, then to >| siteA, then to siteB, then to site C, then to user. This is because the >| "@" syntax takes precedence over the "!" syntax. >+--------------- > >WRONG!!! "@" and "!" are used by different networks -- there is NO defined >precedence between them! (You're comparing apples to oranges.) UUCP sites >give "!" precedence, Internet domain mailers give "@" precedence. The example quoted assumes an RFC 976 compliant mailer. One must recognize that not all uucp-based mail systems are RFC 976 compliant (this is probably the exception rather than the rule), which is (I think) the point Brandon was trying to make. >On ncoast (a UUCP system) "!" has precedence. Our neighbor "hal" gives "@" >precedence because they are on the Internet. This can be quite useful; if I >want to send mail across the Internet I can mail to hal!foo@bar.COM, thus >overriding the UUCP route (which may well be slower) that smail would give >me from a straight "foo@bar.COM". On your system, you would probably want >to switch to UUCP only after you got as "close" to the recipient's system as >possible, so you would want "@" to have higher precedence. Thus, the >current system is useful for both of us. Just don't assume that everyone's >mailer handles things the same way, as we *are* on different networks. The caveat here is that "hal" must be careful to ensure that any mixed bang/@ paths that escape into Internet mail conform to RFC 976. Also, if "hal" accepts mixed bang/@ paths at all, it should interpret them according to RFC 976. Although you can get away with using hal!foo@bar.com, we should probably be encouraging people to use hal!bar.com!foo instead, which *should* work also and had the advantage of being unambiguous. The ambiguity of the mixed syntax really causes problems with situations like the following arise: I send mail to "node1!node2!node3!node4!user@node5.com" . node1 has an old-style uucp mailer, so it just strips off its own name, and sends the mail on to node2. node2, however, has an RFC 976 compliant mailer, and so sends the mail (perhaps via a gateway) to node5.com, rather than to node3. node5 will then try to send the mail to "node3!node4!user" . Surely this is not what I intended, and there is no guarantee that the mail will even arrive at node4, since node3 may not be visible to node5. (Can anyone say how often things like this happen?) >Things get even more interesting when you try to apply your rules to a >DecNet network or to the Bitnet as well (on Decnet, does "::" have >precedence above or below "@"? How about "!"? Does Decnet even *care* what >the relative precedence of "!" and "@" is?) I think the answer here is that it is undefined. The safest thing to do seems to be to rewrite all addresses so that the node::user syntax is never seen in either the Internet or uucp worlds. The *best* thing to do is to define an MX record for each DECnet node and translate addresses when crossing a gateway. The easy thing to do is to rewrite node::user into user%node.decnet@well.known.domain . It's ugly but it does work. (Of course, within the DECnet world, :: has precedence over anything else! ...and you usually have to put the rest of the address in double quotes.) -- Keith Moore UT Computer Science Dept. Internet/CSnet: moore@utkcs2.cs.utk.edu 107 Ayres Hall, UT Campus BITNET: moore@utkcs1 Knoxville Tennessee 37996-1301 Telephone: +1 615 974 0822
ane@hal.UUCP (Aydin "Bif" Edguer) (09/29/88)
In article <12604@ncoast.UUCP> allbery@ncoast.UUCP (Brandon S. Allbery) writes: >As quoted from <2105@edsews.EDS.COM> by roberts@edsews.EDS.COM (Ted Roberts): >+--------------- >| In article <6581@chinet.UUCP>, les@chinet.UUCP (Leslie Mikesell) writes: >| > >| > Suppose >| > you want siteC to route siteA!siteB!siteC!domain.dom!user or >| > siteA!siteB!siteC!user@domain.dom. >| >| These are not the same thing. The first route would send to siteA, then >| siteB, then siteC and siteC would route to user@domain.dom which they >| would know how to do since they know how to resolve domain addresses >| (you hope:-). The second would attempt to send to domain.dom, then to >| siteA, then to siteB, then to site C, then to user. This is because the >| "@" syntax takes precedence over the "!" syntax. >+--------------- >WRONG!!! "@" and "!" are used by different networks -- there is NO defined >precedence between them! (You're comparing apples to oranges.) UUCP sites >give "!" precedence, Internet domain mailers give "@" precedence. BZZZ! Wrong WRONG :-) All internet sites should give "@" precedence. All RFC976 compliant UUCP sites should give "@" precedence. All lowlife, noncompliant UUCP sites give ! precedence :-) Here is the relevant quote from RFC976: "UUCP Mail Interchange Format Standard 2.1 Hybrid Addresses In conformance with RFC-822 and the AT&T Message Transfer Architecture, we recommand that any host that accepts hybrid addresses apply the (a!b)@c.d interpretation." 5. Summary The originating site should ensure that the addresses conform to RFC-822, since no requirement is placed on forwarding sites or gateways to transform addresses into legal RFC-822 format. > On ncoast (a UUCP system) "!" has precedence. Our neighbor "hal" gives "@" > precedence because they are on the Internet. This can be quite useful; if I > want to send mail across the Internet I can mail to hal!foo@bar.COM, thus > overriding the UUCP route (which may well be slower) that smail would give > me from a straight "foo@bar.COM". Yes. But the proper way to do this is to send to hal!bar.COM!foo. WHENEVER POSSIBLE, use non-ambiguous paths (only "!"'s or "@"'s, not both). I promise to correct the problem if it does not attempt delivery to foo@bar.com when given the path hal!bar.com!foo. (note: I make no guarantees it will actually get there) I am not yet a class 3 site (I do not currently use the pathalias database) but I do try to conform to the rest of the RFC976 standard. > Things get even more interesting when you try to apply your rules to a > DecNet network or to the Bitnet as well (on Decnet, does "::" have > precedence above or below "@"? How about "!"? Does Decnet even *care* what > the relative precedence of "!" and "@" is?) If the site is DECnet \(tm ONLY then it is welcome to do anything it wants (including eating fudge brownies whenever it sees an "@") but if it is on the Internet, then it is supposed to be RFC822 compliant and RFC822 says that "@" has precedence! The information to the left of the "@" is local information. Aydin Edguer Case Western Reserve University
allbery@ncoast.UUCP (Brandon S. Allbery) (10/07/88)
As quoted from <290@hal.UUCP> by ane@hal.UUCP (Aydin "Bif" Edguer): +--------------- | Yes. But the proper way to do this is to send to hal!bar.COM!foo. | WHENEVER POSSIBLE, use non-ambiguous paths (only "!"'s or "@"'s, not both). | I promise to correct the problem if it does not attempt delivery to foo@bar.com | when given the path hal!bar.com!foo. (note: I make no guarantees it will | actually get there) +--------------- You will understand that we stopped using this when hal (or was it mandrill?) had a broken sendmail.cf that converted a path that on ncoast was something like "hal!baz.bar.com!foo" into something resembling "baz!foo!bar.com@decvax". (Actually, what was going on was even weirder, but I don't remember it exactly; we switched en masse to mailing through necntc after that.) I can spread the word that it works now, but that wasn't the first time that people here have had problems sending mail through the CWRU mail system; it is to be hoped that it was the last, but "once burned...". (Note: I have restored our sendmail to using "@" "properly", with the restoration of our pathalias database to [partial] functionality.) ++Brandon -- Brandon S. Allbery, uunet!marque!ncoast!allbery DELPHI: ALLBERY For comp.sources.misc send mail to <backbone>!sources-misc comp.sources.misc is moving off ncoast -- please do NOT send submissions direct "So many articles, so little time...." -- The Line-Eater