Hi Brett, it is not as naive as you think. Although it looks like we are blocked at the stdin (prompt) we actually are waiting at a "select()" which monitors activity on: the X11 socket, stdin (keyboard), timers and networks sockets (if any). No threads are needed to monitor all these events. Concerning X11 exposure events in TCanvases, they used to be handled correctly. I just investigated and it turned out that due to some internal API change the request for exposure events was left out from the event mask. My fault. It is now fixed in the CVS head. Backing store is not needed anymore for ROOT! Cheers, Fons. Brett Viren wrote: > > Rene Brun writes: > > This has nothing to do with Root. > > This is not entirely true. ``Well behaved'' X apps are expected to > redraw their windows when an expose event occurs. Relying on: > > > You must activate the backing store option of your window manager. > > is technically a crutch. However, since the flow of control is > usually in the root prompt and there is no way for it to respond to X > events when idle, it is an understandable one. Putting the root > prompt in one thread and all GUI code in another would let root be > ``well behaved'', but at the expense of all the complications threads > bring. > > Regards, > -Brett. -- Org: CERN, European Laboratory for Particle Physics. Mail: 1211 Geneve 23, Switzerland E-Mail: Fons.Rademakers@cern.ch Phone: +41 22 7679248 WWW: http://root.cern.ch/~rdm/ Fax: +41 22 7677910
This archive was generated by hypermail 2b29 : Tue Jan 01 2002 - 17:50:36 MET