[quote='PyrostasisTDK',index.php?page=Thread&postID =32901#post32901]Nice catch, no it only loses stick when Im walking forward with W pressed and hit assist key.[/quote]
Thanks. Let me explain what's happening and what I'm doing about it. (By the way, your description was terrific. Absolutely clear.)

The toon's stick is breaking because his copy of DAOC thinks the person who's playing him hit the "go forward" key on his keyboard. Why does DAOC think that? Because the keypresses you're generating by holding down the W key, which are supposed to go only to the foreground window, are "leaking" into the background window.

Why are they leaking? Because HotkeyNet is doing unnatural things with keystrokes in an effort to manipulate DAOC into accepting keystrokes while its in the background, which is something it was deliberately designed not to do. It's not surprising that this doesn't work perfectly.

What can I do to stop the leaking? Well, for the past week I've had two ideas about this. One idea was to insert some delays in the SendWinMF code. The second idea was to add a buffer for keyboard keystrokes so they don't get sent until after HotkeyNet's unnatural tricks are finished. I decided to implement these ideas one at a time so we can test them and see how well they work.

I started implementing the first idea (the delays) last week. I've changed the implementation of them several times now. In the current build (55) you adjust them with this command:

[url='http://www.hotkeynet.com/ref/setfocusdelay.html']<SetFocusDelay 0 15>[/url]

If you make the second number lower, you'll get less leakage. Unfortunately, there's a tradeoff. As the second number gets lower, the desirable keystrokes (the ones from the macro) are also less likely to get through.

In other words, as you make that number lower, the W's are less likely to get through to the background window, but the keystrokes in the word "/assist" are also less likely to get through.

The question I've been waiting for the testers to answer is: Can you make that second number low enough that leakage stops 100%, but the desirable keystrokes stlil get through?

As soon as I've received a few answers to this question from testers, I'll move on to the second part of the solution, the keystroke buffer, which will (I hope) make these functions work even better.