Patch to allow pg_filedump to support reading of pg_filenode.map

Hello Hackers,

I recently had to work on a case where some catalog files were
corrupt and/or missing. One of the things we sought to inspect was
pg_filenode.map, but there was no tooling available to do so.

With the help of Álvaro H.. I've put together a patch to allow pg_filedump
to do some rudimentary decoding of pg_filenode.map, so that it's at least
human-readable. I had the idea to try to map the OIDs to relnames, but
that might get hairy, with TOAST table OIDs possibly changing (for pg_proc,
etc.)

It seems that Christoph Berg is the primary committer for the pg_filedump
project these days so he is cc'd here, but if there's some other place I
should be sending this kind of contribution to, please let me know.
Hopefully this will be helpful to others in the future.

Much appreciated,
--Richard