Okay, I see. Will take a look at this. And as you rightfully guessed, this is due to an optimization I did in rev 893 (I removed an sql join.) On 9/25/05, Julien Mudry <julien.mudry at gmail.com> wrote: > Hi > > All my deprecated feeds had unread items... and those deprecated > unread items are now counted in the unread total, which wasn't the > case before. This is the first case you describre. > > Best wishes. > > julien > > On 25/09/05, Marco Bonetti <mbonetti at gmail.com> wrote: > > Julien, > > > > I'm not quite sure I understand what is going on here. Do you mean all > > those feeds which have been deprecated have unread items which are > > counted as unread, and which weren't counted before? > > > > Or are those read items which are now counted in the total count? > > > > -m > > > > > > > > On 9/25/05, Julien Mudry <julien.mudry at gmail.com> wrote: > > > Hello > > > > > > I just upgraded to the latest svn version 904 (yeah, I like living on > > > the edge ;-) ). And I nearly got an heart attack when I got back to my > > > web aggregator: the total of unread items was five times higher than > > > the second before, totalling more than 10 thousand! > > > > > > Then I realised that I deprecated quite a few feeds lately to get down > > > to a more bearable level of unread items. And that the unread total > > > was around that level before deprecation. So, my guess is that the > > > changeset 893 (described as "Optimization: optimized the sql query in > > > unreadCount, cached the result") may well be the culprit for this > > > change of behaviour. > > > > > > Marco: is this the new wanted way of counting unread items (ie, > > > including unread but deprecated items in the total)? It will seem > > > strange to users, as the total could be way higher than the items they > > > can see in the page. > > > > > > Regards, > > > > > > julien > > > > > > _______________________________________________ > > > gregarius-dev mailing list > > > gregarius-dev at sinless.org > > > http://sinless.org/mailman/listinfo/gregarius-dev > > > > > >