> -----Original Message----- > From: Ceki Gulcu [mailto:[EMAIL PROTECTED]] > Sent: Sunday, 24 March 2002 00:38 > To: Log4J Developers List > Subject: RE: An alternative JTable > > At 21:52 23.03.2002 +1100, Oliver Burn wrote: > > >Yes, but for one limitation. With the current approach it is possible to > >pause the collection of events and to change the filter > >criteria on the currently > >collected events. With the technique you are proposing, this would not be > >possible. This is because when the filter critera changes, the filtered > >events buffer would need to be repopulated from the other buffer which > >contains new events. > > Hi Oliver, > > I don't understand this last comment. When the filtering criteria > change (new filters are added, removed etc.), you need to repopulate > the (mFilteredEvents) last buffer from the (mAllEvents) buffer > containing all the events. Right? > > Why would it be any different with two buffers? Buffer A accumulates > *all* the events and buffer B contains the events after filtering. If the > filter rules change, then B can be rebuilt from A. Where is the difference?
The difference is that with the current approach, when Chainsaw is paused, collected events are put into the pending buffer and not mAllEvents. So when the filter criteria is changed and mFilteredEvents is rebuilt from mAllEvents, it does not get the events collected whilst paused. When Chainsaw is resumed (un-paused), all the events in the pending buffer get added to mAllEvents and mFilteredEvents (as appropriate). I hope that makes sense - it is quite subtle. Regards, Oliver -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
