Filter::Simple: eliminate /\C/

Filter::Simple: eliminate /\C/

Since that construct is now deprecated, don't use it. It turns out that
it's usage was almost certainly wrong anyway, and a simple /./s is better.

It was temporarily converting literals within the src text to "$;NNNN$;"
where the Ns are 4 octets that encode the index of the literal.  It was
then using /(\C{4})/ to later extract those octets. But since the octets
were originally inserted as chars anyway, they should be extracted as
chars too.

I suspect the original code could have failed on utf8 source once the
number of literals exceeded 0x80.