Implement not, shl and shr in mpz library.  Add function to create mpzs
on the stack, used for memory efficiency when rhs is a small int.
Factor out code to parse base-prefix of number into a dedicated function.

This change allows to build unix version without libreadline installed.

Remove param from python API of ext_register

Minor cleanup.

Don't need to wrap bitfields in their own struct.  Compiler does the
correct thing without it.

Update LED configuration

Print an error if an invalid BOARD is specified

Add EXTI support

Fix floating point printing of negative numbers in the range (0, -1.0)

Also change int -> machine_int_t where appropriate.

With this option selected, only HID on its own works, not VCP+HID.

Revamp usrsw to use new exti code.
Add user switch callback function.

Stack layout is different when -g used, so need to handle the pendsv
jump differently.  Addresses Issue #315.

Add Configurable LFN support to FatFS

Some functionality is still missing (eg and, or, bit shift), and some
things are buggy (eg subtract).

strtonum clashes with BSD function of same name, and our version is
different so warrants a unique name.  Addresses Issue #305.

parse: Refactor parse node encoding to support full range of small ints.

Based on suggestion by @dpgeorge at
https://github.com/micropython/micropython/pull/313

Specifically, VM's small ints are 31 bit, while parser's only 28. There's already
MP_OBJ_FITS_SMALL_INT(), so, for clarity, rename MP_FIT_SMALL_INT() to
MP_PARSE_FITS_SMALL_INT().

TODO: De-duplicate DECODE_UINT, etc. definitions.

Assuming we have truncating (floor) division, way to do ceiling division
by N is to use formula (x + (N-1)) / N. Specifically, 63 bits, if stored
7 bits per byte, require exactly 9 bytes. 64 bits overflow that and require
10 bytes.