How does header retrieval work?

sturkimov · Post by **sturkimov** » Tue Sep 28, 2004 6:38 pm

If I state that I want to get all headers 1 to 3 days old (1-3 in configuration), then how does Newspro actually fetch those headers?

I'm having the suspicion that Newspro actually retrieves every single header in that group, then filtering those away that does not conform to the date-range. But I'm not sure whether or not that is the case always, or only when the newsgroup is empty in Newspro.

So what is it? :)

alex · Post by **alex** » Tue Sep 28, 2004 6:53 pm

it tries first to find the first matching header through retrieving only the date field, then retrieves the rest of headers, it is the fastest possible way, also newly incoming headers may have any date but previously scanned non-matching headers can be skipped, so after you've change the retention (or reset the newsgroup) the first time it will be slower but then it will be faster.

using 'last headers' instead of 'retention' will work faster since there are direct commands in the protocol to do that.