On Sat, 2003-03-01 at 00:43, Akim Demaille wrote: > > I don't think this is the right approach to the problem. We should > first design the interface that config.status needs (mostly > s/@foo@/val/), and implement a tool that does this with existing > tools. > > You would give the list of substitution to perform, and then the tool > would make them.
That is exactly what I am proposing, in fact. I don't understand your objection. > This would allow using Perl, etc. I think there is a reason that Perl is not used thus far, and this certainly not a reason to start using it now. Bob
