[Laszlo-dev] For Review: Change 20090414-ptw-G Summary: Correct path computation to work for canvas

[P T Withington]

Well the debugger is meant to serve both the average Joe and the power  
user.  The average Joe probably doesn't get what the #numbers are  
anyways, so I didn't think putting two in a row would bother him.   
This feature (distinguishing classes whose name is possibly confused  
for the global definition) has been in the debugger since forever, and  
it has certainly helped me on more than one occasion to figure out why  
an object was not behaving as I expected.  So, I'd like to leave some  
way to note this -- but I am open to suggestions on how to make it  
more user-friendly.

On 2009-04-14, at 13:48EDT, André Bargull wrote:

> Hm, I wonder if it is at all necessary to distinguish 'normal'  
> classes from anonymous classes. Actually average Joe shouldn't even  
> know that in certain cases (methods, handlers, constraints, etc.)  
> anonymous subclasses have been generated, should he?
>
> On 4/14/2009 7:35 PM, P T Withington wrote:
>> This is actually correct.  It is telling you that the class of the  
>> canvas is not just canvas, but an anonymous subclass of canvas.   
>> That's why the class is <canvas>#1, instead of just <canvas>, it's  
>> because the class is not lz.canvas, even though it identifies  
>> itself as a class named 'canvas'.
>> If you don't have any methods on the canvas, so that it does not  
>> have to make an anonymous subclass, you won't get the extra #.  The  
>> same will happen with any class that has to create an anonymous  
>> subclass because there are instance methods.
>> I'm open to better ideas on how to display these anonymous classes  
>> if you have one.
>> On 2009-04-14, at 13:04EDT, André Bargull wrote:
>>> Approved for the classname change, but in the debugger the canvas  
>>> is still displayed as:
>>> «<canvas>#1#0| This is the canvas»
>>> Instead of:
>>> «<canvas>#0| This is the canvas»
>>>