examples not working on mac..

[willwade]

I may be being stupid..

from pymouse import PyMouse
from pykeyboard import PyKeyboard

m = PyMouse()
k = PyKeyboard()

dir(k)  #Use dir() on a PyKeyboard instance
k.numpad_keys.viewkeys()
k.function_keys

gives:
AttributeError: 'PyKeyboard' object has no attribute 'numpad_keys'
(and if I comment the numpad line)
AttributeError: 'PyKeyboard' object has no attribute 'function_keys'

(running latest zip on mac)

The initial readme could do with a one liner on how to send a normal key as well as a function key..

[willwade]

Ok so I'm getting this to work - kinda..

k.press_key('command')
k.press_key('shift')
k.press_key('3')
k.release_key('command')
k.release_key('shift')
k.release_key('3')

that actually just sends shift 3 (£ on my layout) - not creating a screenshot (apple shift 3 is a system wide shortcut for creating a screenshot). I'm guessing what its doing is sending command key and not holding it down with the other two.. (i.e. its just tapping it).

I'm grasping at straws here (i.e. I can't figure this out) but is that because there is only a sticky state for shift - and not the other command keys?

(UPDATE: I see that I should have read the previous issue #38 . Apologies for adding to the list..)

[SavinaRoja]

Actually I think you are demonstrating some facets of how the Mac implementation is incomplete. I am not so knowledgeable when it comes to Mac and don't have one to test on. It looks like it could use some fleshing out.

[willwade]

Im wondering if its something to do with it needing kCGEventFlagMaskCommand (I'm still getting my head around this but it looks like the modifier keys need a different approach: http://stackoverflow.com/questions/15206825/cmdoptiond-simulation-in-cocoa/15207244#15207244) -

I've done some test code and this works..

from Quartz import *
from AppKit import NSEvent

keyCode = 8; # C - i.e. we are going to do a a command(apple)+c

source = CGEventSourceCreate(kCGEventSourceStateCombinedSessionState);
eventDown = CGEventCreateKeyboardEvent(source, keyCode, YES);
CGEventSetFlags(eventDown, kCGEventFlagMaskCommand);
eventUp = CGEventCreateKeyboardEvent(source, keyCode, NO);
CGEventPost(kCGSessionEventTap, eventDown);
CGEventPost(kCGSessionEventTap, eventUp);
CFRelease(eventUp);
CFRelease(eventDown);
CFRelease(source);

NB: It does work but it also leads to a segmentation fault. I'm not sure why. It does though tend to suggest you can save pretending to press all the modifier keys - just the keys that you need with the kCGEventFlagMaskCommand set on them..

(also are you aware of this @SavinaRoja : https://github.com/abarnert/pykeycode ? - it may be more reliable than the lookup table - particularly when detecting keystrokes?)

Happy to pay for a remote mac hosting plan for a little while if it helps ;)

[willwade]

I've written an adaption to the current code which seems to be more reliable.. https://github.com/willwade/PyUserInput/blob/master/pykeyboard/mac.py - I would do a pull request but I'm not 100% confident its that much better - the eval is horrible and there is a shift problem (see this gist for an example: https://gist.github.com/willwade/52ccf1455772d6597c30)

Note that for the examples to work, run them in the terminal then switch to the finder.. more example shortcuts to test: http://support.apple.com/kb/HT1343

UPDATE: It does now work - I realised it was only sending one modifier so now it can send multiple including a shift modifier. I'm still a little concerned this may mess up the standard shift configuration - let me know if you want me to do a pull request.