dww@stl.stc.co.uk (David Wright) (08/09/90)
In article <25157.26b95615@kuhub.cc.ukans.edu> sloane@kuhub.cc.ukans.edu (Bob Sloane) writes:
#Well, I really screwed up this time!
#
#Unfortunatly, I spelled "cancel" as "Cancel" with the result that
#about 70 invalid control messages were inflicted on the net.
and all the news admins on all the news hosts on all the net got 70 mail
messages each about them. But thanks for owning up!
More seriously - its an easy mistake to make. And why should control
messages be case-sensitive anyway? I suggest that the maintainers of B
and C news add a patch to the next set thereof which makes the check for
control command case-insensitive. Only a small thing, but it will prevent
re-occurance of this particular problem, at least for up to date sites,
which after a suitable number of years will be most of them.
Regards, "None shall be enslaved by poverty, ignorance or conformity"
David Wright STL, London Road, Harlow, Essex CM17 9NA, UK
dww@stl.stc.co.uk <or> ...uunet!mcsun!ukc!stl!dww <or> PSI%234237100122::DWW
<or> /g=David/s=Wright/org=STC Technology Ltd/prmd=STC plc/admd=Gold 400/co=GBjbuck@galileo.berkeley.edu (Joe Buck) (08/09/90)
In article <3295@stl.stc.co.uk>, dww@stl.stc.co.uk (David Wright) writes: [ Someone mis-spelled cancel as Cancel in a control message, generating tons of mail ] |> More seriously - its an easy mistake to make. And why should control |> messages be case-sensitive anyway? I suggest that the maintainers of B |> and C news add a patch to the next set thereof which makes the check for |> control command case-insensitive. Only a small thing, but it will prevent |> re-occurance of this particular problem, at least for up to date sites, |> which after a suitable number of years will be most of them. No, it won't help much, David. When someone writes "cancell" in a control message, will you then ask for the Interlisp DWIM package (do what I mean) to be added to news? The bug happened because the current software propogates bogus control messages. If it threw them away instead with a message on stderr when you tried to post one, the original sender would have gotten an error message and no cruft would have been sent out to other systems. If an article is defective in some way, it should not be forwarded to other sites. -- Joe Buck jbuck@galileo.berkeley.edu {uunet,ucbvax}!galileo.berkeley.edu!jbuck
henry@zoo.toronto.edu (Henry Spencer) (08/09/90)
In article <3295@stl.stc.co.uk> dww@stl.stc.co.uk (David Wright) writes: >and all the news admins on all the news hosts on all the net got 70 mail >messages each about them. But thanks for owning up! Actually, C News sites only get one per batch; here, for example, it was only one (rather long) mail message. >More seriously - its an easy mistake to make. And why should control >messages be case-sensitive anyway? I suggest that the maintainers of B >and C news add a patch to the next set thereof which makes the check for >control command case-insensitive... I'm lukewarm about this; on the one hand the case distinction has no particular value in this context, on the other hand control messages do have to be carefully spelled anyway, and it's one more nuisance in the code. This one's ultimately Geoff's decision, since it's his code, and I suspect he may not want to bother. A quick look at RFC1036 finds no imdication that control messages are supposed to be case-insensitive. -- The 486 is to a modern CPU as a Jules | Henry Spencer at U of Toronto Zoology Verne reprint is to a modern SF novel. | henry@zoo.toronto.edu utzoo!henry
karish@mindcrf.UUCP (08/10/90)
In article <38053@ucbvax.BERKELEY.EDU> jbuck@galileo.berkeley.edu (Joe Buck) writes: >In article <3295@stl.stc.co.uk>, dww@stl.stc.co.uk (David Wright) writes: >[ Someone mis-spelled cancel as Cancel in a control message, generating > tons of mail ] >The bug happened because the current software propogates bogus control >messages. If it threw them away instead with a message on stderr when >you tried to post one, the original sender would have gotten an error >message and no cruft would have been sent out to other systems. And it would be practically impossible to add a new control message to the currently-recognized set. Usenet would become even more unwieldy than it now is if every site were to start enforcing all sorts of policies on other sites' articles. -- Chuck Karish karish@mindcraft.com Mindcraft, Inc. (415) 323-9000
emv@math.lsa.umich.edu (Edward Vielmetti) (08/10/90)
In article <9008100218.AA23766@mindcrf.mindcraft.com> karish@mindcrf.UUCP writes:
And it would be practically impossible to add a new control message
to the currently-recognized set.
B News, perhaps. I think in C News it's just a matter of dropping a shell
script in the right place, I think....
--Ed
Edward Vielmetti, U of Michigan math dept <emv@math.lsa.umich.edu>darcy@druid.uucp (D'Arcy J.M. Cain) (08/10/90)
In article <38053@ucbvax.BERKELEY.EDU> jbuck@galileo.berkeley.edu (Joe Buck) writes: >In article <3295@stl.stc.co.uk>, dww@stl.stc.co.uk (David Wright) writes: >[ Someone mis-spelled cancel as Cancel in a control message, generating > tons of mail ] > >|> More seriously - its an easy mistake to make. And why should control >|> messages be case-sensitive anyway? I suggest that the maintainers of B > [...] >The bug happened because the current software propogates bogus control >messages. If it threw them away instead with a message on stderr when >you tried to post one, the original sender would have gotten an error >message and no cruft would have been sent out to other systems. > >If an article is defective in some way, it should not be forwarded to other >sites. > I think that sites should pass on control messages it doesn't understand. If in the future new control types are added we don't want old news software to throw it away because it doesn't understand it. We want it passed on in case another site can understand it. The bogusity (?) should be caught at the site that posts the control message. If their software doesn't recognize the control type then it shouldn't allow it to be generated or at least give an "Are you sure?" prompt. -- D'Arcy J.M. Cain (darcy@druid) | D'Arcy Cain Consulting | MS-DOS: The Andrew Dice Clay West Hill, Ontario, Canada | of operating systems. + 416 281 6094 |
smb@ulysses.att.com (Steven Bellovin) (08/11/90)
In article <EMV.90Aug10112634@urania.math.lsa.umich.edu>, emv@math.lsa.umich.edu (Edward Vielmetti) writes: > In article <9008100218.AA23766@mindcrf.mindcraft.com> karish@mindcrf.UUCP writes: > > And it would be practically impossible to add a new control message > to the currently-recognized set. > > B News, perhaps. I think in C News it's just a matter of dropping a shell > script in the right place, I think.... You misunderstand. If netnews dropped control messages it didn't understand, how could anyone invent a new one and expect it to be propagated? Wait for the whole net to upgrade its code...?