AutoHotkey Community

[Robert Carnegie]

In the context of of discussion http://www.tabletpcbuzz.com/forum/topic.asp?TOPIC_ID=32151
I want to detect mouse or stylus pen-down in more-or-less any application, without interfering with the normal effect of the action - including double-click, drag-and-drop, etc.

There are hints that I can use OnMessage() to detect activity, possibly using AHk's GUI, but I don't understand GUI.

The actual goal is to implement a new behaviour with existing on-screen keyboard software for Tablet PC - the new behaviour being that the keyboard window is to pop up next to the text cursor when you left-click in a text field, i.e. when you click with an I-beam cursor.

This is in the context of scripting for general useful context-sensitive behaviour of an on-screen keyboard program - I'm using Fitaly but others are popular - appearing when it's wanted, not when it isn't; and if possible making constructive suggestions for a script such as WinWarden, which I found on a link from here. That is to say, I want to find out how to do this and then I want to suggest to Skrommel to put it into WinWarden. Then whenever someone says "I want to make my on-screen keyboard behave this way" - fade away, fold up, bounce away from the stylus - other users can say "Do it with WinWarden".

Executing an arbitrary script definition from a WinWarden "mode" (more or less an event) could be useful, I think. It currently can execute a command line, but that is not a completely satisfactory solution because an AHk script run that way takes noticeable time to get started and run. An alternative that might be good is to send a message from WinWarden to another AHk script already loaded - waiting to be woken up. The message should identify the window and/or the WW "mode" that generated the message; in principle, the other script can read the settings WinWarden.ini if need be.

I have a simple and apparently effective prototype script at http://fitaly.com/board/fsbetaforum/posts/372.html that does not catch clicks (and another which does) but merely loops with short sleeps for the system to be momentarily idle and then tests for I-beam cursor, which usually implies that focus is on a text field.

I also welcome suggestions of other approaches.

Incidentally, is a sleep-loop that repeatedly tests the idle time efficient? An AHk timer that runs whenever idle would be arguably more elegant, but that doesn't seem to be an option - and if it's okay to do it this way, it may not be necessary. However, would it be better to use a timer instead of a loop?

[Chris]

Your approaches seem reasonable. You probably already know that you can catch mouse clicks (perhaps even those generated with a tablet pen) by creating pass-through hotkeys for the mouse buttons. For example:

~LButton::
~RButton::

As for timers vs. sleep-loops, a timer is usually more convenient but a sleep loop might perform slightly better (I'm not sure).

[Robert Carnegie]