On Sat, 2002-08-03 at 06:54, Tomas Frydrych wrote: > > Martin sent this to me privately, but I think this needs to go to the list, > it raises an important issue -- I hope Martin does not mind. > > > HI Tomas, > > Think this through very very carefully. I'm not saying it can't be > > done > > but it is not obvious to me right now. The issue is cursor dirt. If > > the cursor is on (black) and a clearScreen is called over the cursor > > then it gets set to white even though the internal state says it's > > black. Then if the blink thread XOR's it now it will turn to black > > even though the cursor's internal state says it should be white. > > > > What is even worse is if the clear screen stops after clear the first > > set of vertical pixels. Now the cursor is half white, half back and no > > matter what, it looks funny and your guarenteed to get pixel dirt. > > > > Getting rid of cursor dirt was a real pain, I worry that putting the > > blink into a seperate thread will mess it all up. > > > > I do not want the change the actual way the cursor behaves, only > put the timer which fires it into a separate thread so that it fires > every 500 ms and does not get blocked by some other timer. That's > all. >
I don't see how this would help. You'd have to turn the cursor off while doing screen updates anyway (like doing spell checking). I think the effect would be much the same. > I had a look at Mike's cursor class; it looks very good, but it would > mean to rip the whole cursor code up, and that's not something I > personnally want to do just now. > > Tomas
