Previously I wrote a fairly elaborate shell script to send/receive
news to/from my PC. One day I tried to upload part of ba.jobs.offerred
to yarn on my PC for a friend. That group had about 5000 articles.
Uqwk refused to process that many articles -- even to produce a
summary file with just 5000 subject headers. At some point in its
processing uqwk reported ``bus error'' -- to the Unix machine did.
Now I no longer use uqwk, I use trn: I simply put the patterns I
am interested in one of trn's kill files. Only this time, trn selects
these articles, instead of junking them. I enter trn, save the selected
articles to a file, and voila -- I have my downloadable file!
Trn seems much faster, and much more able to handle large numbers of
articles than uqwk. I think this is because trn reads an overview file
to get article headers, whereas qwk and other programs (tin) actually
have to open each article (file) and grab its header.
So much for the tip for the day.