nelson@IMAGE.SOE.CLARKSON.EDU (Russ Nelson) (10/12/89)
Now that rpi.edu is running C news, we are now in the position of being able to receive news without a Lines: header. I just read my first such article, with a Path: like so: image.soe.clarkson.edu!rpi!brutus.cs.uiuc.edu!psuvax1!psuvm!cunyvm!maine.bitnet The implication is that all these sites are running C news, or some other news software that doesn't write Lines: headers. I wonder what is the longest path for an article without a Lines: header? Or rather, what is the longest path in terms of sites rather than characters? -- --russ (nelson@clutx [.bitnet | .clarkson.edu]) Live up to the light thou hast, and more will be granted thee. A recession now appears more than 2 years away -- John D. Mathon, 4 Oct 1989.
coolidge@brutus.cs.uiuc.edu (John Coolidge) (10/12/89)
nelson@IMAGE.SOE.CLARKSON.EDU (Russ Nelson) writes: >[article without Lines: header, path:] >image.soe.clarkson.edu!rpi!brutus.cs.uiuc.edu!psuvax1!psuvm!cunyvm!maine.bitnet >The implication is that all these sites are running C news, or some other >news software that doesn't write Lines: headers. Actually, that's not right. Neither B nor C news actually adds on Lines: headers as far as I know. Lines: is intended to be generated at posting time by inews & friends. C news, by default, doesn't generate Lines: at inews-time (of course, I and many others have patched C news' inews to generate Lines:, and I wish creation was the default. I also wish news clients and nntp were smarter and didn't need it at all, but...). Once the article is created, though, both B and C will pass it on as written (wrt Lines:). Neither one will either delete or add Lines: headers. I'm not sure what rpi is running (they haven't responded to my 'version' control yet). I (brutus.cs.uiuc.edu) and psuvax1 are running C. I really doubt either psuvm or cunyvm is running C, since they're VM machines and I don't think C has been ported to those yet :-). I'm not sure what maine.bitnet is, but it's probably either VM (most likely) or VMS (second most likely), since there are precious few unix machines on BITNet. --John -------------------------------------------------------------------------- John L. Coolidge Internet:coolidge@cs.uiuc.edu UUCP:uiucdcs!coolidge Of course I don't speak for the U of I (or anyone else except myself) Copyright 1989 John L. Coolidge. Copying allowed if (and only if) attributed. You may redistribute this article if and only if your recipients may as well.
flee@shire.cs.psu.edu (Felix Lee) (10/13/89)
PSUVM runs homegrown netnews software that several other VM machines also run, and the current version doesn't generate "Lines:". There's at least one other netnews system for VM machines, and it might not generate "Lines:" either. -- Felix Lee flee@shire.cs.psu.edu *!psuvax1!flee
tale@pawl.rpi.edu (David C Lawrence) (10/14/89)
In <8910121231.AA17253@image.soe.clarkson.edu.> nelson@IMAGE.SOE.CLARKSON.EDU (Russ Nelson) writes: Russ> [article without Lines: header, path:] Russ> image.soe.clarkson.edu!rpi!brutus.cs.uiuc.edu!psuvax1![..BITNET..] Russ> The implication is that all these sites are running C news, or Russ> some other news software that doesn't write Lines: headers. Right; as John later points out the BITNET nodes are running software which doesn't add Lines:. In <1989Oct12.165913.18364@brutus.cs.uiuc.edu> coolidge@brutus.cs.uiuc.edu (John Coolidge) writes: John> Actually, that's not right. Neither B nor C news actually adds John> on Lines: headers as far as I know. B News does add it; if it isn't there, it will be when it hits a B News site. Until we set up the brutus connexion, we had not really gotten any postings without Lines:. Now the only ones I get without it are all from BITNET sites downstream from John. John> I'm not sure what rpi is running (they haven't responded to my 'version' John> control yet). C, as of mid-September when I finally found out that all my grief was being caused by hard-coded, incorrect pathnames in the bottom of batch.c, provided by Jean-Francois Lamy. (_Why_ did you commit such a sin, Jean-Francois? I was about ready to kill when I found that. :-/) I never saw your version message, and I "read" control. Odd. One other interesting thing in all of this is that I think we are seeing a case of the non-dying groups here. John will apparently carry any group any of his feeds carry, and I'll carry nearly anything in our regional distributions and alternate hierarchies, but I delgroup'ed alt.sex.beastiality as soon as it showed up. It will not be carried here, but since the brutus connexion was made I've got it heading for junk. Russ probably has it at his site too (whether he wants it or not) because C News insists on propagating it. I put !alt.sex.beastiality in my sys file, yet it still shows up. I did this in accordance with Geoff's article: In <1989Oct4.234355.3680@utstat.uucp> geoff@utstat.uucp (Geoff Collyer): Geoff> If you really want to drop articles in certain groups, and ignoring Geoff> cross-posting for now, you must either write "!unwanted.group" in your Geoff> sys file subscription list, or you must change the flags field of Geoff> "unwanted.group" to "x" (see news(5)). So now am I required to also add this group to my active file and give it an 'x' flag? Blech. Dave -- (setq mail '("tale@pawl.rpi.edu" "tale@itsgw.rpi.edu" "tale@rpitsmts.bitnet"))
geoff@utstat.uucp (Geoff Collyer) (10/14/89)
I put the hardcoded paths in NNTP's batch.c, not Jean-Francois. It
seemed to me that there was no existing mechanism in the NNTP source to
allow all the directory names that we use to take non-standard values
and there were only a few #defines for the directory names, and since we
don't own NNTP and didn't want to produce "C NNTP", I didn't retrofit
the C news configuration mechanisms into NNTP, I just hardcoded the
pathnames. It didn't seem to matter given the state of the NNTP
configuration mechanisms at the time. I may have missed something,
though. With any luck, Stan Barber has integrated our changes smoothly
into NNTP; I haven't checked yet.
>!alt.sex.beastiality in my sys file, yet it still shows up.
There is only one "a" in bestiality. If you spell it right, you can
exclude it either in your sys file or your active file.
--
Geoff Collyer utzoo!utstat!geoff, geoff@utstat.toronto.edulamy@ai.utoronto.ca (Jean-Francois Lamy) (10/14/89)
tale@pawl.rpi.edu (David C Lawrence) writes: >being caused by hard-coded, incorrect pathnames in the bottom of >batch.c, provided by Jean-Francois Lamy. (_Why_ did you commit such a >sin, Jean-Francois? I was about ready to kill when I found that. :-/) Having comitted no such sin (go ask my ex-students about how much of a fanatic I can be on such matters) I was just about ready to call you up for a duel at dawn to clear up my name, when I found this article in which geoff@utstat.uucp (Geoff Collyer) confesses: >I put the hardcoded paths in NNTP's batch.c, not Jean-Francois. So there. I'm willing to settle out of court :-) Jean-Francois Lamy lamy@ai.utoronto.ca, uunet!ai.utoronto.ca!lamy AI Group, Department of Computer Science, University of Toronto, Canada M5S 1A4
smaug@eng.umd.edu (Kurt Lidl) (10/14/89)
In article <1989Oct13.225725.15574@utstat.uucp> geoff@utstat.uucp (Geoff Collyer) writes: >I put the hardcoded paths in NNTP's batch.c, not Jean-Francois. That was an evil deed :-( >It seemed to me that there was no existing mechanism in the NNTP source to >allow all the directory names that we use to take non-standard values >and there were only a few #defines for the directory names, and since we >don't own NNTP and didn't want to produce "C NNTP", I didn't retrofit >the C news configuration mechanisms into NNTP, I just hardcoded the >pathnames. It didn't seem to matter given the state of the NNTP >configuration mechanisms at the time. I may have missed something, >though. With any luck, Stan Barber has integrated our changes smoothly >into NNTP; I haven't checked yet. I would just like to add this plea for future releases of the code, especially for the documentation. Have a minimal release note on using nntp with C-News. While I think that C-News is the greatest thing since sliced bread (after all, I run it here) -- I was rather frustrated with the documentation. No mention of how nntp shows up in any really helpful manner. I was sadly deluded into thinking that the ihave/sendme stuff was relevant to nntp transfer... A couple of questions to other sites on campus that had news running with nntp shed enough light on the matter so I could be up and running with nntp. While maybe even most of the people responsible for installing the C-News systems across the nation were veterens, *I* certainly was not. I *wish* that I had known lots of little things that I know now. I *wish* that they had been documented someplace. I have read lots of newsgroups for the last 2 years, and have followed C-News from the first "real" (non-alpha) release of it. Still I learned quite a bit from actually playing with the installation. I hope that at least something gets into the docs for nntp for your next "patch" or "upgrade/suppliment" or whatever. Also, a little bit of knowledge about the common flags that you might want in your sys files (eg "F" for nntp...) would have been greatly appreciated. If anybody is interested in starting a guide to UseNet management, let me know. I would like to be in on the project. Conversely, if there is already an existing guide, please tell me where I can find TFM! Thanks. >Geoff Collyer utzoo!utstat!geoff, geoff@utstat.toronto.edu I really admire the C-News system -- it runs extremely well on mojo.eng.umd.edu -- a Sun 3/60 with 12 megs and a 300 meg CDC for /var/spool/news! -- ========================================================== == Kurt J. Lidl (smaug@eng.umd.edu) (301) 454-3184 == == UUCP: uunet!eng.umd.edu!smaug (301) 454-1526 == ========="X Windows: Power Tools for Power Fools"=========
lamy@ai.utoronto.ca (Jean-Francois Lamy) (10/14/89)
geoff@utstat.uucp (Geoff Collyer) writes: >though. With any luck, Stan Barber has integrated our changes smoothly >into NNTP; I haven't checked yet. NNTP 1.5.6 does have some C news support, and no hardwired path names. There were however a couple of typos introduced in the process and a couple of glitches due to the absence of h_addr_list on 4.2-ish systems. Patches for these things were posted, and I posted a version of nntpsend with an example sys file entry to this newsgroup. Hopefully all these will make it in 1.5.7 Jean-Francois Lamy lamy@ai.utoronto.ca, uunet!ai.utoronto.ca!lamy AI Group, Department of Computer Science, University of Toronto, Canada M5S 1A4