From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Fri Aug 03 2001 - 09:04:38 CDT
In <yllmleknfp.fsf@windlord.stanford.edu> Russ Allbery <rra@stanford.edu> writes:
>Charles Lindsey <chl@clw.cs.man.ac.uk> writes:
>> So perhaps something like
>> History -> API -> Inode-for-article (retains original Xref)
>> \-> List-of-overview-records (or list of group names/whatever)
>That would be a useful future enhancement, I think. Somewhat low on the
>priority list since it's not needed for anything really important, but it
>makes the database generally more consistent and lets one move articles
>between storage backends more easily (which I find much more interesting
>than moving articles between newsgroups).
So would a fair summary be to say that the present INN mechanisms are
"Ugly" (I think that was your word) and need to be fixed when tuits are
available? And that when it gets fixed, making it cope sensibly with
aliasing and mvgroup could be incorporated as part of the fix without
undue extra difficulty? But that this particular fix (though desirable) is
low on the priority list at the moment?
It would be useful at this point to hear from the implementors of some of
the other serving agents.
>> Now, after a mvgroup (or other re-aliasing), you leave the article
>> (inode) strictly alone, but change the List-of-overview-records in the
>> API (or accessed via the API/whatever).
>That leaves incorrect information in the Xref header of the article.
Yes, but in a revised system you might not need to change it at all.
>> BTW, you include the complete Xref in your .overview. Is there some
>> benefit in this for newsreaders out there, or is it primarily to assist
>> the inner workings of INN?
>Newsreaders need it for killfiling based on crossposts. Pretty much all
>news servers provide it these days because killfiling based on crossposts
>is just too useful not to support.
OK.
But the interesting question, assuming the problems of the server are taken
care of somehow, is whether the newsreaders care whether the Xref of
existing articles gets updated. A little miscounting of cross-post limits
during the overlap period can probably be tolerated.
As for the other main use of Xref (avoiding seeing crossposts twice) the
evidence from my experiments was that nn didn't care whether Xref was
'corrected' or not, and Netscape got some things wrong if you corrected it,
and others if you didn't (maybe it would have worked if you included both
groups in the Xref). Do we have information on other readers?
-- Charles H. Lindsey ---------At Home, doing my own thing------------------------ Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl Email: chl@clw.cs.man.ac.uk 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