> > Trying to catalogue and assign a permanent place and number to every > single optimization message a compiler can generate is a much much much > worse idea, IMHO.
In the same way numbering warning messages is a bad idea (yes ICC and a couple other compilers do but we should not). -- Pinski
