I just think there are better/faster/easier/more efficient ways to do it.
Building what amounts to a full-on translater into the kernel is the hard
way to go. This was _one_ suggestion of _one_ method; most of the
suggestions that have crossed the list have a zero or one-instruction
performance overhead, and range from less-than-zero to massive memory
overheads.
> Oops data is not performance critical.
No, but its contents _are_ critical (not for performance, but for
debugging), and I don't like the idea of putting any more steps than
necessary between their source and the user/logfiles. (Particularly
anything dealing with large string tables, which could itself be a prime
candidate for OOPSing, esp in the alpha/beta phases...) Call it an
aesthetic preference... ;-)
Adam
-- He feeds on ashes; a deluded mind has led him Adam Bradley, UNCA Senior astray, and he cannot deliver himself or say, Computer Science "Is there not a lie in my right hand?" Isaiah 44:20 bradley@cs.unca.edu http://www.cs.unca.edu/~bradley <><