sample newsrc:
thufir@dur:~$ cat .newsrc
gwene.com.androidcentral: 1-99999999
gwene.com.blogspot.emacsworld: 1-99999999
gwene.com.blogspot.googlecode: 1-99999999
gwene.com.blogspot.googlereader: 1-99999999
gwene.com.economist: 1-99999999
gwene.com.googlereader: 1-99999999
thufir@dur:~$
Now, I have leafnode configured to save messages for 999 days. When GNU NNTP connects to leafnode on localhost there can be a big of lag while things load. Currently, what I’ve done is to configure Leafnode to just store 35 days for particular groups, so that things load more quickly.
However, when looking over the .newsrc again, it seems that I could probably let leafnode store a very large number of articles and then configure the .newsrc which GNU NNTP uses so that only a small portion of those messages are fetched. If so, this would allow for long retention in leafnode, 999 days, which alleviating the delay in connecting to leafnode and loading the articles from GNU NNTP.
However, will that work? Will only the newer articles get loaded, or will it load the old articles?
Unfortunately, RFC 977 doesn’t actually mention .newsrc files. Of course, whether GNU NNTP follows the RFC would be a separate question, but, at least according to the RFC, presumably newer questions are loaded and older questions left?
It looks like RFC 3977 clarifies RFC 977 a bit, so I’m reviewing that now.
You are actually asking about the behaviour of a particular NNTP client, not of the NNTP protocol per se. According to the spec, the only NNTP verb for retrieving articles is ARTICLE, and it fetches exactly one article. There is no constraint on the order in which a client makes ARTICLE “calls”.
So to answer your question you’d need to look at the GNU NNTP library’s documentation … or the source code. And I suspect it also depends on how your code uses the library methods.
If I was in your shoes, I’d try it out and see whether it works.