Hi Doug

It may be good to have another to switch between looks. this way this is not
simplicity over multiple but simplicity and multiple choices.

This is what I'm trying to do by having  a visitor on the parse tree. We can
have multiple pretty-printer wihout the mess of switch statements.

So if you would have a solution for the system window this would be fun to
try.

Stef


on 1/7/02 6:35 AM, Doug Way at dway at riskmetrics.com wrote:

> 
> Tim Rowledge wrote:
>> 
>> ducasse <ducasse at iam.unibe.ch> is widely believed to have written:
>> 
>>> I was frustrated not to be able to find that so after fighting with other
>>> methods, I found the trick. The model of the SystemWindow should be
>>> specified first as follow then the color can be declared.
>> I had similar frustrations when doing the VMMakerTool stuff; very
>> irritating.
> 
> I've had problems figuring out the color handling of SystemWindows too... you
> also have complicated methods like #paneColor (which is what #setWindowColor:
> really sets, not the morph color), and then #existingPaneColor and
> #paneColorToUse, etc.
> 
> Perhaps, a couple of versions from now, if everyone is happy with the
> alternateWindowLook, we could dump support for the original window look and
> simplify some of the code a bit. (although the original window look is
> probably better for 1-bit displays right now...)
> 
> In any case, maybe a comment should be added to #setWindowColor: stating that
> the model must be set first, or something to that effect.
> 
> - Doug Way
> dway at riskmetrics.com
> 
>