[code] Menu Validation (MF_DISABLED bugged on PC?)

[Dan Rathbun]

Re: Menu and Toolbar validation.

It seems on SU 8.0M1 the validation flag MF_DISABLED has no effect. The user can still select the item, and the comand is run.

Is this a new bug?

How does this compare to Mac ?

Code: Select all

#  Valid integer values for PC:
#
# ** MF_BITMAP not used by Sketchup
#
#  0 = MF_UNCHECKED | MF_ENABLED
#  1 = MF_UNCHECKED | MF_GRAYED
#  2 = MF_UNCHECKED | MF_DISABLED
#  3 = MF_UNCHECKED | MF_GRAYED | MF_DISABLED
#
#  4 = MF_UNCHECKED | MF_BITMAP | MF_ENABLED
#  5 = MF_UNCHECKED | MF_BITMAP | MF_GRAYED
#  6 = MF_UNCHECKED | MF_BITMAP | MF_DISABLED
#  7 = MF_UNCHECKED | MF_BITMAP | MF_GRAYED | MF_DISABLED
#
#  8 = MF_CHECKED   | MF_ENABLED
#  9 = MF_CHECKED   | MF_GRAYED
# 10 = MF_CHECKED   | MF_DISABLED
# 11 = MF_CHECKED   | MF_GRAYED | MF_DISABLED
#
# 12 = MF_CHECKED   | MF_BITMAP | MF_ENABLED
# 13 = MF_CHECKED   | MF_BITMAP | MF_GRAYED
# 14 = MF_CHECKED   | MF_BITMAP | MF_DISABLED
# 15 = MF_CHECKED   | MF_BITMAP | MF_GRAYED | MF_DISABLED
#
# note 16 = MF_POPUP


If @@validation is a class var holding the validation integer for a menu item or toolbar item. You could use an instance var @validation. Or you could pass the integer in, by putting a validation argument into the () on the def method line, and changing all the @@validation to just validation:

Produce it's Bitwise OR string:

Code: Select all

def valid_string()
  
  valid=''
  
  # Test the 4th bit: MF_CHECKED==1
  if @@validation[3]==1
    valid << 'MF_CHECKED | '
  else
    valid << 'MF_UNCHECKED | '
  end
  
  # Test the 3rd bit: MF_BITMAP==1
  #
  # valid << 'MF_BITMAP | ' if @@validation[2]==1 
  
  if @@validation[0]==1 or @@validation[1]==1
    # Test the 1st bit: MF_GRAYED==1
    if @@validation[0]==1
      valid << 'MF_GRAYED'    if @@validation[1]==0
      valid << 'MF_GRAYED | ' if @@validation[1]==1
    end
    # Test the 2nd bit: MF_DISABLED==1
    if @@validation[1]==1
      valid << 'MF_DISABLED'
    end
  else #
    valid << 'MF_ENABLED'
  end
  return valid
end # def
  

checked?() method

Code: Select all

#  checked?()
#
#  Returns Boolean state of MF_CHECKED.
#
def self.checked?()
  #@@validation >= 8 # on PC
  (@@validation & MF_CHECKED).nonzero? ? true : false
end


enabled?() method

Code: Select all

#  enabled?()
#
#  Returns Boolean state of MF_ENABLED.
#  !! NOT EXACT same as MF_UNCHECKED !!
#
def self.enabled?()
  # [0,4,8,12].include?(@@validation) # on PC
  # on PC MF_DISABLED has no effect. Use MF_GRAYED instead.
  ((@@validation & MF_GRAYED).nonzero? ||
    (@@validation & MF_DISABLED).nonzero?) ? false : true
end


disabled?() method

Code: Select all

#  disabled?()
#
#  Returns Boolean state of MF_DISABLED.
#
def self.disabled?()
  # [2,3,6,7,10,11,14,15].include?(@@validation) # PC
  (@@validation & MF_DISABLED).nonzero? ? true : false
end


grayed?() method

Code: Select all

#  grayed?()
#
#  Returns Boolean state of MF_GRAYED.
#
def self.grayed?()
  (@@validation & MF_GRAYED).nonzero? ? true : false
end


unchecked?() method

Code: Select all

#  unchecked?()
#
#  Returns Boolean state of MF_UNCHECKED.
#  !! NOT EXACT same as MF_ENABLED !!
#
def self.unchecked?()
  #@@validation < 8 # PC
  (@@validation & MF_CHECKED).nonzero? ? false : true
end


---

EDIT: made Boolean methods use constants and bitwise comparison for cross-platform use, (which greatly simplifys the check for MF_GRAYED.)

[thomthom]

This is what I always use for validation procs:

Code: Select all

  def self.can_paste_proc
    if self.can_paste?
      MF_ENABLED
    else
      MF_DISABLED | MF_GRAYED
    end
  end


I though MF_DISABLED prevented the action from triggering. And that MF_GRAYED where used only for the visual effect.

[thomthom]

hmm... testing in SU8M1, you're right. MF_DISABLED seem to do nothing. Only MF_GRAYED seems to affect anything.
Though I imagine there is a reason I use MF_DISABLED | MF_GRAYED - just can't remember why. Can't find any notes on it.

[thomthom]

Same behaviour on SU6, 7.1 Windows.

[Dan Rathbun]

Though I imagine there is a reason I use MF_DISABLED | MF_GRAYED - just can't remember why. Can't find any notes on it.

I'll bet it's a cross-platform thangie... it does not seem to matter to Windows if the DISABLED bit is set, so no harm is done.

Need a Mac guy to test it and let us know.

[Dan Rathbun]

This is what I always use for validation procs:

Just being clear that the methods above are not the actual validation_proc. I try not to put conditional evaluations into the menu or toolbar procs, if I can do the decision in another place in the code. In most cases, the validation state does not change until the user actually clicks on the item (or button,) so it's better to change the menu item's (or button's) valid state within the command proc, or some other method that gets called. So when the change is made I just change the value of a @@validation class var (or @validation instance var,) and let Sketchup just pick up the value of that var when it must do a UI draw.

So creating a menu item, would look like:
item = menu.add_item('Some Menu Label') { ..execute_code.. }
menu.set_validation_proc(item) { @@validation }

[thomthom]

Yea, that'd probably be better.

IN my case I'd need an observer, but it'd still be more efficient.

[thomthom]

Tested on OSX SU8.0M1 - Same behaviour, MF_DISABLED does nothing.

[Dan Rathbun]

hmm... testing in SU8M1, you're right. MF_DISABLED seem to do nothing. Only MF_GRAYED seems to affect anything.

Same behaviour on SU6, 7.1 Windows.

Tested on OSX SU8.0M1 - Same behaviour, MF_DISABLED does nothing.

Thanks go to ThomThom for filing a Google BugReport on this, 30 JAN 2011.

[thomthom]

I still wonder why I've been using MF_DISABLED...

[Dan Rathbun]

Covering your butt.. just in case the API changes in the future?

[thomthom]

Possibly. Maybe it was in another language where you had to gray and disable. ...Visual Basic...? Can't think of anything else where I've dealt with menus...