On Sat, Nov 13, 2004 at 11:33:49AM -0800, Michael Wolfson wrote:
> 
> On Nov 10, 2004, at 11:51 PM, Eduardo Horvath wrote:
> 
> >.  It appears that the trap happens before the FCode is done being
> >interpreted.  I wonder if it could be a tokenizer issue.  You can try
> >the attached version of fgen and see if the bootblock it generates
> >works any better than the in-tree version.
> 
> Well, at least it lets me get more information.
> 
> Does this help at all?

What it tells me is that the problem is during the initial FCode
evaluation, not during the bootblock execution.  This could be
a hardware issue, buggy OF, or something caused as a sideffect
of buggy FCode generation.

Eduardo