🚀 go-pugleaf

In <01bddbde$26463580$0301a8c0@mackrory>, "Col Mackrory" <mackrory@acay.com.au> writes:

>Netscape does this to leafnode - and I am never sure why.
>
>When you subscribe to groups in Netscape it seems to tell leafnode that it
>is interested in all the newsgroups.
>
>What you need to do - is after you have subscribed to the news groups that
>you are interested in - go to /var/spool/news/interesting.groups and delete
>any files that are not of interest. It is these files that leafnode
>references to work out what groups to fetch. i.e. you should have a file
>for each news group you want leafnode to fetch.
>
>Leafnode will automatically expire these groups after 7 days if you have
>not read anything from that group.
>
>Once you get Netscape settled in - you will find that leafnode will happily
>collect your news if you place your fetch command in your ip-up.local.

I've been trying off and on for seveal weeks to get leafnode to 
work for me.  Originally, I couldn't even get it to start: fetch 
would die with a seg fault whenever it ran and any attempt to 
connect to leafnode, ie through the loopback either with a news 
reader client or simply "telnet 127.0.0.1 119" would cause 
leafnode itself to die.  I solved this last problem (seems to 
have been a problem with how my hostname was listed in 
/etc/hosts) but I still can't get fetch to run for me.  Usually 
it dies immediately with a seg fault.  

Occassinally I have been able to get fetch to start up and begin
to download an active list from my ISP, but I've never managed to
go beyond this.  I've never gotten a complete active list yet,
and certainly no articles.  And God, is it slow!  I suspect fetch
only runs a single thread to connect to the remote host.  The
longest it ran was for almost an hour and it still only managed
to get part of the active list.  My OS/2 server package (changi)
uses a multi-threaded "fetch" utility that can download my ISP's
active list in less than ten minutes using eight simultaneous
connections.  I tried downloading the active list with changi in
OS/2 and copying the file to where leafnode could use it, but
leafnode didn't like something about this and went into a
prolonged pout again.

I was hoping I'd be able to move my news server from OS/2 to 
linux, but things just aren't looking too promising at the 
moment.   At this point I am considering either finding an 
alternative to fetch to feed leafnode (since leafnode itself 
appears to work now; I just don't have any articles to use with 
it), maybe suck or slurp or whatever somebody can suggest that 
works, or running a full-blown nntp server like INN, which 
strikes me as overkill for the three or four users here at my 
house. 

Looking for ideas...

-John (John.Thompson@ibm.net)