When hardware cursor is not available, mouse cursor should be disabled in BDirectWindow windows.

[bga]

In BeOS, the lack of a hardware mouse cursor would prevent using BDirectWindow in Window mode completely. One of the reasons for this is the fact that the drawing thread will draw over the mouse cursor and this looks very weird (check GLTeapot for example).

As we do not have hardware cursor support at all yet, this problem happens even with accelerated drivers.

I would like to propose that we keep the window support even without hardware cursor support but I would like to suggest that, in this case, we disable the mouse cursor automatically in such window.

If the code could be smart enough automatically reenable if it is over a interface kit widget inside the Window, even better.

of course this would have to be documented so developers know that, in this case, they would have to draw their own cursor (we could even provide a method to do exactly that that the developer could call from inside his drawing thread.

[axeld]

IMO weird looks don't matter much in that regard, and since you cannot know which parts of a direct window actually needs a mouse cursor, it doesn't make sense to disable it.

One day, we should have hardware cursor support anyway, but even now, I find the status quo acceptable.