These programs keep the netnews overview database for newsreaders
up to date.
Initov creates the database initially by reading all the articles under
a . This takes a long time
(on the order of 12 hours,
in late 1992),
and can be avoided by letting
domkov incrementally generate a database for new articles as they arrive;
after a week or so,
most articles should be in the database.
The news system should not be locked when
initov is run;
initov takes care of the necessary locking internally,
and deals with articles that arrive while it is running
(primarily by deferring their processing to
domkov ).
Domkov is run from
newsrun (8CN). It updates the database files
from the header stream produced by
relaynews (8CN) for new articles.
Mkov normally reads a stream of RFC 1036 headers,
with those of each article separated from those of its neighbours
by an empty line,
from
file s (standard input if none)
and writes a summary in the format of
newsoverview(5) , but with a field containing the article's
news spool directory name relative to
/usr/spool/news prepended,
on standard output.
-a reads articles
instead of a header stream
(but ignores
files s which are directories or have
names which do not resemble article names,
i.e. do not have a leading digit on their last component).
and omits the initial directory name field;
-n requests the initial directory name
despite -a ; -o causes only articles older than the file
stamp to be summarised.
Expov should be run after
expire (8CN) (and is, by
doexpire (8CN)) to delete database entries for missing articles.
