Retire mahonia for code.google.com/p/go.text/encoding

[GoogleCodeExporter]

Mahonia maintainers have decided to discontinue support, and suggest
go.text/encoding to achieve the same ends.

https://code.google.com/p/mahonia/code.google.com/go.text/encoding/

Below, a more current incarnation of the library: https://godoc.org/golang.org/x/text

An example of a cutover:

oov/the_platinum_searcher@c5c5960

A brief look at go.text/encoding indicates it's a very different approach. This
is a nice-to-have, but not yet critical unless a manhonia bug appears that
effects go-dbf.

[LindsayBradford]

Biggest difference to work out between the libraries is that Mahonia automatically bridges the charset mappings to their matching encoders/decoders. It's then a piece of cake to ask for a new encoder/decoder, based only on the charset name string.

In the example cutover link above, the authors had to build their own decoder/charset name bridging.

Either this is something offered by the golang.org/x/text library (but I'm missing), or I'd need to backfill this charset/encoding bridge within the go-dbf library. The backfilling option is unappealing given my lack of knowledge on a) which chartsets need supporting, and b) which encoder/decoders then match the charset names I need to support.