Xchars can make a all to xcb and find the latest track from the latest set cursor position. We can set a watch and wake IO loop when cursor seems headed this way. I think they are set up for this. This simplifies the sleeping fd problem, how long to sleep before looking at cursor input.
But the problem remains, the master screen can get lost. And it recoveries because users thrash out to tune into the desired cursor. In all cases, the equilibrium is when and one of the window cursors is in tune with window handler.
The muse tuning process
Take your mouse to the border of you screen and beyond. The muse is stuck at the screen edge. It can only go up. But your mouse is gone off, a foot else the screen, near the edge of he table. X has lost you, and X will find you when to the extent you move your mouse is a dy, up from the bottom edge, for example. You will jerk it once, maybe twice, and the cursor suddenly follows your velocity. The X filter trying to correlate with your dy coming from the mouse. Unknowingly, you are more like a morse code operator, giving it dots and dashes then a passive user of a non-adapive machine. Search engine, same trick. They become good because the user becomes good.
No comments:
Post a Comment