>
> I was pleased to see FixUnderscores in the image.  It seems to work,
> grumbling about only a few methods in Kernel and Morphic, the two
> packages against which I ran it.  Is the plan to leave it to
> individuals, or to convert the entire image prior to release of 3.9?

We will see since recompiling all the packages may introduce problems  
(since we do not have atomic load).
The plan is to try to do it.


> Will the prohibition against underscores in selectors be removed at  
> the
> same time?

I do not think so. Someone should propose a change first :)


> The ToolBuilder discussion flew by me as it was happening - too much
> life in the way.  Seeing it in the image lead me to do some catch up
> reading, and the yes/no decision on packaging was hard to miss.  There
> was mention of some one-time fixes that should be applied and removed;
> no argument.  I would package it, named as some type of one-shot fix,
> apply and then remove the package.  I suspect the rest wants to  
> live as
> a package that is more or less critical to the tools in the image.   
> Dumb
> question: is the builder a code generator, or is it runtime support  
> that
> renders view from specs as they open?  So far, it looks like the  
> latter
> to me.
>
> Bill
>
>
>
> Wilhelm K. Schwab, Ph.D.
> University of Florida
> Department of Anesthesiology
> PO Box 100254
> Gainesville, FL 32610-0254
>
> Email: bills at anest4.anest.ufl.edu
> Tel: (352) 846-1285
> FAX: (352) 392-7029
>
>
>