Problem with field borderwidth 2

[Richard Gaskin]

Dar Scott wrote:
> On Dec 30, 2004, at 1:50 PM, Richard Gaskin wrote:
> 
>> Rev now supports native appearances for fields on XP when the field's 
>> default values are left in place (borderWidth set to 2).  So what your 
>> screen show at <http://www.sanke.org/RevTestPage/>  is showing is XP's 
>> rendering of the field border.
>>
>> You'll see a similar thing on OS X: while the appearance differs from 
>> XP, when a field border is set to the default value of 2 OS routines 
>> are used to render the control instead of the engine's.
>>
>> I think this should be overridable.  Standard and Rectangle style 
>> buttons also take on native appearances, but you can override this by 
>> specifying a backgroundColor.
>>
>> Perhaps we need a Bugzilla request so that if the borderColor is set 
>> the OS no longer draws the control and it uses the engine's internal 
>> routines.
>>
>> Please post the bug number after you log it so we can vote for that 
>> enhancement.
>>
>> In the meantime a workaround is to set the style of the field to 
>> "shadow" instead of "rectangle" or "scrolling", and then to prevent 
>> the shadow from drawing set its shadowwidth to 0 (zero).  Shadow-style 
>> fields are always drawn by the engine, so you should get what you're 
>> after with those.
>
> Where did you get all this info?

It's funny, but I honestly don't know.  I even called Ken to see how he 
learned this, and he doesn't know either.  Alien visitation?

I think I learned some of it in an email with Scott Raney when he was 
doing support for MetaCard (the OS X appearances were put into the 
engine before its acquisition by RunRev);  when I first saw the 
relationship between borderWidth and OS rendering I filed a bug report 
(he explained it's a feature).  The rest I learned through experimentation.

This should definitely be documented, and I feel there should be an 
explicit property for the control to determine whether it uses native 
appearances or not.  Relying on magic combinations of what should be 
irrelevant properties (who would think that changing the borderWidth 
would determine whether the OS renders the control?) is, to be as polite 
as possible, suboptimal. ;)

What shall we call this new property?  Has this been proposed in 
Bugzilla yet?

--
  Richard Gaskin
  Fourth World Media Corporation
  __________________________________________________
  Rev tools and more: http://www.fourthworld.com/rev