Re: algorithmic question

Charlie,

MARK_REFLOW_b() is a macro. The '//#define' is picked up by the ibex 
preprocessor, thus one thing expanded going to 'build'.

-Jeff

Charles Goodwin wrote:

>On Thu, 2004-04-01 at 02:38 -0800, Adam Megacz wrote:
>  
>
>>Charles Goodwin <[email protected]> writes:
>>    
>>
>>>There definitely needs to be two different directions for resizing, one
>>>for the parent to resize and propogate changes downwards and another for
>>>a child resizing with changes propogated upwards.
>>>      
>>>
>>Right; see NEEDS_REFLOW (upwards) and reflow() (downwards).
>>    
>>
>
>Only NEEDS_REFLOW and reflow() don't exist - do you really mean
>MARK_REFLOW and repack()?
>
>What's the difference between MARK_REFLOW and MARK_REFLOW_b?
>
>Also MARK_REFLOW_b is never declared anywhere in Box.java or anywhere
>else, although it is used several times:
>
>[email protected] ibex $ egrep -r MARK_REFLOW_b src/org/ibex/*
>src/org/ibex/Box.java:    //#define MARK_REFLOW_b for(Box b2 = b; ...
>src/org/ibex/Box.java:        MARK_REFLOW_b;
>src/org/ibex/Box.java:            MARK_REFLOW_b;
>
>(Why doesn't this make the compiler barf?)
>
>Perhaps that could be causing some of the unecessary repack() calls?
>  
>