rickert@mp.cs.niu.edu (Neil Rickert) (05/04/91)
Enclosed are excerpts from article headers that were rejected yesterday. They were all forwarded to me by one of my feeds, which has itself disabled the Cnews rejection so it can see what is happening. I receive only a very limited feed (mostly comp and news). Yesterday there were 9 such articles. Of these, 7 were subsequently received from my other feed, having presumably been appropriately munged along the way. For each article, I include only the 'From:' and 'Message-ID:' headers, together with all invalid headers. The problem headers are either: Empty headers with no trailing blank. 'References:' with no blank after ':' In one case a user entered 'Sender:' The most common problem seems to be with the 'References:' header. It seems apparent that problems are occurring mostly with empty headers, or headers which were probably edited by humans. The machine generated headers are mostly correct. Clearly news posting software needs to be more careful to guard against these kinds of easy mistakes. (I agree with Henry and Geoff, though, that correcting entry mistakes is not the responsibility of the transport). From: dlarson@cbmvax.commodore.com (Dale Larson) Message-ID: <21154@cbmvax.commodore.com> Summary: Expires: Sender: Followup-To: Distribution: From: phillip@BARTAL.COM (Phillip M. Vogel) Message-ID: <451@bartal.BARTAL.COM> References: From: us264465@mmm.serc.3m.com (Rich A. Nazarian) Message-ID: <1991May2.124144.4686@mmm.serc.3m.com> Sender:Rich Nazarian From: jln@casbah.acns.nwu.edu (John Norstad) Message-ID: <1991May2.151351.14748@casbah.acns.nwu.edu> References:<1991Apr30.121736.10695@ux1.cso.uiuc.edu> <1991Apr30.172718.11744@dartvax.dartmouth.edu> <1991May1.153943.5808@engin.umich.edu> <1991May1.161827.23293@ux1.cso.uiuc.edu> <1991May1.215732.11589@engin.umich.edu> From: jln@casbah.acns.nwu.edu (John Norstad) Message-ID: <1991May2.153326.15711@casbah.acns.nwu.edu> References:<1991May1.161827.23293@ux1.cso.uiuc.edu> <13275@goofy.Apple.COM> From: kcv0@bunny.gte.com (Kurt Van Schalkwijk) Message-ID: <11122@bunny.GTE.COM> References:<1991Apr29.150159.12586@wam.umd.edu> <1991Apr30.150818.6304@ccad.uiowa.edu> <1991May1.045558.29716@leland.Stanford.EDU> From: william@kaula.keck.hawaii.edu (William Lupton) Message-ID: <12802@uhccux.uhcc.Hawaii.Edu> Expires: References: Sender: Followup-To: Keywords: From: robt@mummy.agsm.unsw.oz.au (Rob Trevor) Message-ID: <1406@usage.csd.unsw.oz.au> References:<1991Apr26.060223.9904@fs7.ece.cmu.edu> <1991Apr26.194705.4975@unx2.ucc.okstate.edu> <WVENABLE.91Apr29002559@spam.ua.oz.au> From: drg@mdaali.mda.uth.tmc.edu (David Gutierrez) Message-ID: <5026@lib.tmc.edu> References:<233@hermix.UUCP> <22556@natinst.natinst.com> -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
cathyf@lost.rice.edu (Catherine Anne Foulston) (05/10/91)
In article <1991May3.185024.26725@mp.cs.niu.edu> rickert@mp.cs.niu.edu (Neil Rickert) writes: >It seems apparent that problems are occurring mostly with empty headers, >or headers which were probably edited by humans. Or both at the same time. I was deleting a stupid Distribution: the other day, and I use vi. How easy it is to put the cursor one character to the right of the colon and press C, producing an article which C News (if I understand correctly) will reject. I remembered in time, but who would if they didn't read this group? Don't get me wrong, I agree that bogus headers should be bounced (if I had tons of free time I'd make mods to bounce the Distribution I was trying to fix -- I think it was "rec.sport.football.pro,news.admin,local") and that it's the responsibility of posting software to fix or block such articles. I'm not sure I believe that a space after the colon should be required in an otherwise empty header, though. If that's what the RFC says, well then I guess I disagree with the RFC. :-) Though I'm not a fan of implementing "improvements" to an RFC since generally all they do is produce incompatibility. Cathy -- Cathy Foulston + cathyf@rice.edu + Rice University, Network & Systems Support
rickert@mp.cs.niu.edu (Neil Rickert) (05/10/91)
In article <1991May9.220348.2906@rice.edu> cathyf@lost.rice.edu (Catherine Anne Foulston) writes: >In article <1991May3.185024.26725@mp.cs.niu.edu> rickert@mp.cs.niu.edu (Neil Rickert) writes: > >>It seems apparent that problems are occurring mostly with empty headers, >>or headers which were probably edited by humans. > >Or both at the same time. I was deleting a stupid Distribution: the other >day, and I use vi. How easy it is to put the cursor one character to the >right of the colon and press C, producing an article which C News (if I >understand correctly) will reject. I remembered in time, but who would if >they didn't read this group? I believe, at least with the current Cnews, this is not a problem. It will be corrected by inews and/or anne.jones. Or at least this should take care of the empty headers. I am not so sure of the 'Header:noblank' situation. It only is rejected if it gets past these and into relaynews. Thus the usual problem is with articles from another site with bad headers, which never see inews and anne.jones. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940