🚀 go-pugleaf

In article <869817068snz@dsl.co.uk>
           bhk@dsl.co.uk "Brian {Hamilton Kelly}" writes:

> Yesterday I discovered a thread (in alt.satellite.tv.europe) where
> *every* article appeared twice; to aid identification of the thread, the
> first article has Message-ID <33D61F24.13B1@interport.net>.

> Further analysis showed that the original poster had included
> alt.satellite.tv.europe TWICE in the list of newsgroups to which it was
> posted.  Seemingly, the result of NewsBorg's NEWNEWS command was to serve
> up the article's Message-ID twice (albeit that the newsgroup makes only
> one appearance in the Xref header line); this is OK, but why did neither
> KA9Q nor SNews' unbatch process notice the duplication?

From the information you gave, it's really not clear that ka9q did
receive two copies or that newsborg did anything wrong whatever. I
believe that what you have seen is solely the result of unbatch
taking the list of newsgroups from the header of one single copy of
the article and then acting literally and according to the stupid
instruction that it found there.

As snews history is a single file and not a series of single 
newsgroup files, the decision to insert an article can only be taken 
one time prior to starting to insert the article everywhere, and then 
the newsgroups line has to be followed, plus or minus whatever 
junking rules are used.  If the history was re-written after each 
newsgroup file was closed then no cross-posting would be possible.

If this was a serious problem, a special case test could be placed 
into unbatch to search out duplicate names prior to insertion.  But I 
think that any such effort would be disproportionate to this very 
rare problem.

Regards,
-- 
   Tony - G3SKR / W2TG                   email: tgold@panix.com