mangling of function names

[Martin von Loewis]

> It occurred to me, since we agreed today that since names of functions
> are not going to be substituted anyway, why don't we just eliminate
> them as substitution candidates?  That way there's no need to rewrite
> the name grammar, which pretty much works now.  What I'm proposing is
> another exception in the `Compression' section that states that the
> <name> of a function shall not be a substitution candidate.

That could be done. But then, this only safes a single slot in the
table. So for simplicity of the spec, I would not introduce that
special case.

Regards,
Martin