Alexander Gottwald wrote: >I've wrapped the ShowCursor handling with if (g_fSoftwareCursor).
Yea, that's what I was wondering. You might consider removing the ShowCursor from wincursor.c altogether. I don't think it hurts but it's not really necessary as long as other bits of the system aren't going to go hiding it. Oh, and you might want to update the bit in the docs to include the differences between hardware and software cursor handling. >Can you please test it with the Terminal Service? The cursor works fine. There are a few other miscellaneous problems in multiwindow mode though. Window activation doesn't appear to affect the X window stacking order, window titles are all "Cygwin/X X." I hope those aren't unexpected. Uh, never mind, seems to be an artifact of Spy++ running when starting XWin. Uggh. And while on the subject of cursors, have there been any discussions about xeyes support in multiwindow mode and the problems caused because X clients don't realize that they're not visible? Try this: run emacs in (default) x-mode. Minimize all other windows besides emacs. Take note of where any of the icon bar buttons are. Minimize the emacs window. Hover the mouse over where one of the buttons was. Up pops the emacs tooltip/balloon help window. You get the same behavior if it was another windows app obscuring the emacs window. I was thinking about a -noxeyes flag that just turned off the WIN_POLLING_MOUSE_TIMER_ID timer, or maybe just not calling miPointerAbsoluteCursor when the cursor is over an non-X window. Suboptimal but less annoying. Thanks, Dan
