[leafnode-list] feature request for fetchnews

[Server Admin]

Matthias Andree wrote:
> Roi Dayan <roi at roi.no-ip.org> wrote in June 2005:
> 
> 
>>can fetchnews have a dynamic option of max sizes of messages allowed for 
>>retrival ?
>>I want this so my crontab will first get all small messages quickly and 
>>then I'll run it again
>>to fetch all the rest of the messages.
>>this way i could read some normal text posts while it fetches the binaries.
> 
> 
> This feature request has riped a bit, but I must say it's a bit
> delicate: fetchnews relies on fetching articles in order, so we'd either
> have to concoct something around the -x option (which will become rather
> inefficient quickly) or use a separate file to track larger articles, or
> it could reuse the existing delaybody scheme and record all skipped
> articles to the interesting.groups files - which will also become
> inefficient.
> 
> I must say I am not particularly fond of leafnode being used as binary
> server, as Usenet is rather ill-suited for binaries - there are much
> better transports available, with varying degrees of anonymity.
> 

I have been digging into the debug outputs of v1 recently.

It appeared to me as though leafnode was checking all the messages it 
needed and compiling a list of 'interesting' ones that passed X-Over 
filters. It should be relatively easy at that point to sort by 
size/lines. That would go some way toward getting the smaller ones first 
and fastest. Even if its not a full feature.

AYJ