From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Fri Sep 18 1998 - 17:20:13 CDT
In <199809162210.SAA03236@cis.ohio-state.edu> Dave Barr <barr@cis.ohio-state.edu> writes:
>Don't say "The History file". Not only is it implementation-specific,
>but it's wrong for INN and C-news. I don't forsee the history file
>being modified to accomodate a P-ID, notwithstanding the fact that
>dbz precludes this from working.
Eh? CNews certainly uses a history file such as I mentioned, and I always
understood that INN (indeed most servers) were much the same. I agree that
there is a problem with dbz (see my comments elsewhere in this thread).
>INN would implement it with another database (not dbz) with P-ID->M-ID
>mappings.
That sounds much more complex than what I suggested (assuming my scheme
can be fixed to work).
>>4. Initial posting of Named article (i.e. one which contains a Named:
>>header, and provides both an M-ID and a P-ID, even if they are identical
>>on the first occasion).
>>
>> If P-ID already exists in the History file, then report Error and
>Shouldn't that be "M-ID"? Posting a message with an existing P-ID
>is not an error--it's by design.
No, an invariant of the whole scheme is that there must only ever exist
one article with a given P-ID (i.e. one P-ID/M-ID pair). So creating
another new article with the same P-ID is an error. If your intention was
to replace the old article with a new one, then you should be using a
Supersedes/Replaces header, not another initial posting.
>Delete (4) since it's redundant with existing practice.
No it isn't.
>> reject. Else create History file entries under both M-ID and P-ID
>> (or just the one if they are identical) and store article with new
>> Xref numbers for each group.
>>
>>5. Cancel of Named article (cancel command uses the M-ID)
>Same with (5).. there's no changes in cancels so there's no need to
>mention it.
Yes there are. There are two IDs to get rid of (well, if we give up
attempting to delete history file entries, then it may come back to the
existing cancel mechanism).
>>6. Supersedes of an ordinary article (i.e. replacing article has no Named:
>>header).
>Same again.
No, there are definitely differences needed here, because we need to make
the same P-ID point to a new Xref. I agree we have a problem here if this
means having two History entries with the same P-ID.
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Email: chl@clw.cs.man.ac.uk Web: http://www.cs.man.ac.uk/~chl Voice/Fax: +44 161 437 4506 Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K. PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5