From: Thomas L. <ta...@ec...> - 2003-10-15 09:51:40
|
On Wed, Oct 15, 2003 at 11:06:18AM +0100, Geoff Youngs wrote: [...] > Which leads on to the topic of ROX's IPC as used for communicating > between processes. It works fine most of the time, but it is not > reliable, because it relies on an X property of an invisible window not > changing inbetween the time that X sends the PropertyNotify message to > ROX and the time that ROX reads the value of the property from the > invisible window. > > If you want an example (albeit unlikely in normal usage), then try > finding a large directory (e.g. /usr/lib) and running > for i in *; do (rox -s $i &); done > And see how long it is before your terminal window is filled with > messages telling you that the exisiting ROX-Filer is not responding. That's not unreliable, it's just slow. If you open a few hundred windows, it will slow the X server down so much that the 15 seconds (or whatever it is) timeout will trigger and the sender will give up. You'd see exactly the same with D-BUS, if you used the same timeout. -- Thomas Leonard http://rox.sourceforge.net tal00r at ecs.soton.ac.uk tal197 at users.sourceforge.net GPG: 9242 9807 C985 3C07 44A6 8B9A AE07 8280 59A5 3CC1 |