* ron minnich rminnich@lanl.gov [031002 06:18]:
On 1 Oct 2003, Eric W. Biederman wrote:
The one pass algorithm of the current config tool, already does not work 100% of the time, the evidence is some options that do not work properly if I move them. So we might as well rework the code so it operates in two passes, a parsing passes and then an option evaluation pass.
OK, OK, two passes.
Is this really closer to optimum than the current default..option behaviour? I don't really think that even having default ROM_SIZE in a mainboard config file would cause more mistakes than in the Options.lb file. But I may be wrong...
Stefan