"Richard A. O'Keefe" wrote:
> 
> 
> Well, we have the "self assert: [...]" mechanism.  Perhaps some way of
> turning that off would do.  But it isn't relevant here, because there
> are correct results to produce, and the present code produces incorrect
> results.
I thought the  assert: package by vassily Bykov was ported to squeak and posted to the list. I wonder if (and if not why not) this has been added to the base image. The package provides for disabling assertions with abolutely no runtime costs (if you dont count the avoerhaed of loading some extra bytes with the literal header of a method).

HTH,
Torge