Re: Interacting with cnfs

Hi Nigel,

I was thinking about a report I did for my older server, which used the
tradspool storage method, and that was to count how many articles are
in a group and list how many new articles have been added since last
run.
 I can't see a way to make this happen using cnfs without pulling the
storage token using a similar method as 6.4 in the FAQ and then
manually parsing the newsgroups line from each article.

You can use overview information to get the number of articles in a newsgroup.
% getlist -R counts
news.software.nntp 17619 1 13223 y
gives you a list of newsgroups containing the information (13223 articles in news.software.nntp, between article numbers 1 and 17619).
For accurate results, see the groupexactcount setting in readers.conf. You may want to set it to 0 for connections coming from localhost or whichever server running your command.
How many articles have been added since the last run is not straight-forward though...  You have to somehow record the latest high water marks for each newsgroup, and see how many articles have a greater article number than these recorded high water marks.
Use tdx-util or ovsqlite-util for that, assuming you have a tradindexed or ovsqlite overview.
For instance, assuming I have recorded that the latest highest article number for news.software.nntp was 17619, I get the number of new articles with the following command:
% ovsqlite-util -g -n news.software.nntp -a 17620- | wc -l
0
Incidentally, you may also use tdx-util or ovsqlite-util instead of "getlist -R counts" for the counts:
% ovsqlite-util -i -n news.software.nntp
news.software.nntp 17619 1 13223 y 1720923358 0
I hope this could help.

Do we have any better tools for interacting with cnfs...or if not,
should we?

Naturally, every tool is perfectible and can be enhanced.  It just needs someone to write it and share it with the community.

A similar thought that when looking for text, I could grep through a
group or set of groups. Not sure this would even be possible with cnfs
either?

You have to force a text search with "grep -a".
--
Julien ÉLIE
« Pour aller plus vite, j'additionne toujours de bas en haut : je fais
   du même coup l'addition et la preuve. » (Aurélien Scholl)