law@ioe.lon.ac.uk (Lindsay Wakeman) (04/25/91)
(Apologies to those who have seen this request - I got my distribution wrong the first time around) We have a problem with relaynews - it reports errors (many!) about an article number being too large (100000) for the max field of 5 digits. This has affected the junk newsgroup (possible overflow of articles with bad newsgroups?) Anyway, it wouldn't go beyond 99999. All the articles have now been expired but /usr/lib/news/sys still has 99999 as the last article number. How do we reset this - isn't it normally done automatically? I don't have my original sources online so it isn't easy to check. I'd be very grateful for any clues - Thanks. Lindsay PS we run Cnews 6.4 but not very up to date with patches. Could this be the problem? -- JANET: law@uk.ac.lon.ioe | Lindsay Wakeman EARN/BITNET: law%ioe.lon.ac.uk@ukacrl.bitnet | Institute of Education INTERNET:law%ioe.lon.ac.uk@nsfnet-relay.ac.uk | 20 Bedford Way London WC1H OAL UUCP: !mcvax!ukc!educ-isis!law | VOICE +44 71 636 1500 ext.512
geoff@world.std.com (Geoff Collyer) (04/26/91)
Lindsay Wakeman: > Anyway, it wouldn't go beyond 99999. All the articles have now been expired > but /usr/lib/news/sys still has 99999 as the last article number. > How do we reset this - isn't it normally done automatically? Assuming that you mean /usr/lib/news/active, no, the fields are not widened automatically. The active file is (and always has been, I believe) updated in place, so the fields must be made wide enough in advance (that's why there are all those leading zeros). To quote news(5): Both article-number fields are at least five digits wide. To fix this problem forever (or until we need article numbers larger than 32 bits anyway), lock the news system with locknews (see newsmaint(8)), add 5 zeroes onto the left of the second field of each line of the active file, and unlock the news system. > PS we run Cnews 6.4 but not very up to date with patches. Could this be the > problem? Where did you get your C News distribution from? We don't number our releases. -- Geoff Collyer world.std.com!geoff, uunet.uu.net!geoff
henry@zoo.toronto.edu (Henry Spencer) (04/26/91)
In article <1991Apr25.122242.1179@ioe.lon.ac.uk> law@ioe.lon.ac.uk (Lindsay Wakeman) writes: >We have a problem with relaynews - it reports errors (many!) about >an article number being too large (100000) for the max field of 5 digits... >How do we reset this - isn't it normally done automatically? No, it never has been, not by B News or C News. There have to be enough digis in the active-file field to begin with -- hence the leading zeros -- because it gets updated in place. The fix is to add some more zeros on the front manually. -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry
law@ioe.lon.ac.uk (Lindsay Wakeman) (04/26/91)
OK - fixed now. You have to manually edit the article numbers in /usr/lib/news/active (having first run /usr/lib/newsbin/maint/locknews) and make them large enough fields (add a few zeros on the front, better still zero altogether if no articles present - ie. 000000000 00000000 y Thanks to Jerry Aguirre of Olivetti ATC, USA and Icarus Sparry of Bath University, UK. :-) Lindsay -- JANET: law@uk.ac.lon.ioe | Lindsay Wakeman EARN/BITNET: law%ioe.lon.ac.uk@ukacrl.bitnet | Institute of Education INTERNET:law%ioe.lon.ac.uk@nsfnet-relay.ac.uk | 20 Bedford Way London WC1H OAL UUCP: !mcvax!ukc!educ-isis!law | VOICE +44 71 636 1500 ext.512
mrl@uai.com (Mark R. Ludwig) (04/27/91)
In article <1991Apr25.215042.25734@world.std.com>, geoff@world (Geoff Collyer) writes: >To fix this problem forever (or until we need article numbers larger >than 32 bits anyway), lock the news system with locknews (see >newsmaint(8)), add 5 zeroes onto the left of the second field of each >line of the active file, and unlock the news system. The version of NEWSBIN/expire/upact I have (circa December, 1990) explicitly rewrites the third field of NEWSCTL/active with 5 digits...if this hasn't been changed yet...$$ -- INET: mrl@uai.com UUCP: uunet!uaisun4!mrl PSTN: +1 213 822 4422 USPS: 7740 West Manchester Boulevard, Suite 208, Playa del Rey, CA 90293 WANT: Succinct, insightful statement to occupy this space. Inquire within.
henry@zoo.toronto.edu (Henry Spencer) (04/27/91)
In article <1991Apr26.182609.23026@uai.com> mrl@uai.com (Mark R. Ludwig) writes: >>... add 5 zeroes onto the left of the second field of each >>line of the active file, and unlock the news system. > >The version of NEWSBIN/expire/upact I have (circa December, 1990) >explicitly rewrites the third field of NEWSCTL/active with 5 >digits... 2 != 3. If you read the upact coverage in expire(8) carefully, you will find a discussion of why the second and third fields get treated differently. -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry
jerry@olivey.ATC.Olivetti.Com (Jerry Aguirre) (04/27/91)
In article <1991Apr26.182609.23026@uai.com> mrl@uai.com (Mark R. Ludwig) writes: >The version of NEWSBIN/expire/upact I have (circa December, 1990) >explicitly rewrites the third field of NEWSCTL/active with 5 >digits...if this hasn't been changed yet...$$ The third field is not updated "in place" as the second field is. Besides I think you will find that it writes it with a minimum of 5 digits, more if needed. The update of the minart field involves re-writing the entire active file so it can expand the field when it is needed. On the other hand I don't know why the minart could not be updated in place just as the second field is. That would eliminate the need to lock out news processing while doing the update.
henry@zoo.toronto.edu (Henry Spencer) (04/27/91)
In article <50711@olivea.atc.olivetti.com> jerry@olivey.ATC.Olivetti.Com (Jerry Aguirre) writes: >On the other hand I don't know why the minart could not be updated in >place just as the second field is. That would eliminate the need to >lock out news processing while doing the update. Alas, not so. Apart from that being harder to do, the in-place updates are done *in memory* by relaynews on a large-memory machine, and it rewrites the entire active file in one shot when it's done, so you really have to lock it out. -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry
wisner@ims.alaska.edu (Bill Wisner) (04/28/91)
> The fix is to add some more zeros on >the front manually. Or do what I did a few days ago: % cat > widen < FnorD #!/usr/bin/perl while (<>) { ($ng, $low, $high, $flag) = split; printf("%s %010d %010d %s\n", $ng, $low, $high, $flag); } FnorD % chmod +x widen % locknews % mv active foo % ./widen foo > active % exit Bill Wisner <wisner@ims.alaska.edu> Gryphon Gang Fairbanks AK 99775 bnug, dude yeah .
gemini@geminix.in-berlin.de (Uwe Doering) (04/29/91)
geoff@world.std.com (Geoff Collyer) writes: >Lindsay Wakeman: >> PS we run Cnews 6.4 but not very up to date with patches. Could this be the >> problem? > >Where did you get your C News distribution from? We don't number our >releases. This number looks rather like the nn news reader release number. It's currently at release 6.4. Uwe -- Uwe Doering | INET : gemini@geminix.in-berlin.de Berlin |---------------------------------------------------------------- Germany | UUCP : ...!unido!fub!geminix.in-berlin.de!gemini
law@ioe.lon.ac.uk (Lindsay Wakeman) (04/30/91)
In <U7FQGOO@geminix.in-berlin.de> gemini@geminix.in-berlin.de (Uwe Doering) writes: >geoff@world.std.com (Geoff Collyer) writes: >>Lindsay Wakeman: >>> PS we run Cnews 6.4 but not very up to date with patches. Could this be the >>> problem? >> >>Where did you get your C News distribution from? We don't number our >>releases. >This number looks rather like the nn news reader release number. It's >currently at release 6.4. > Uwe >-- Oops, yes you are quite right, I confused our nn release number with this in my effort to give some pertinent info. Silly. It's all sorted now anyway. ..Lindsay -- JANET: law@uk.ac.lon.ioe | Lindsay Wakeman EARN/BITNET: law%ioe.lon.ac.uk@ukacrl.bitnet | Institute of Education INTERNET:law%ioe.lon.ac.uk@nsfnet-relay.ac.uk | 20 Bedford Way London WC1H OAL UUCP: !mcvax!ukc!educ-isis!law | VOICE +44 71 636 1500 ext.512