Resolve "Document and publish the programming and testing procedure"

Closes #3

See merge request be-cem-edl/fec/hardware-modules/rf-att-4cha!2

no need to remove FTDI driver after EEPROM is programmed

Resolve "Provide a mechanism to program an empty EEPROM"

Closes #2

See merge request be-cem-edl/fec/hardware-modules/rf-att-4cha!1

The new scheme also work for empty EEPROMs, as they come from the
factory after production.

We basically ignore everything and erase the eeprom and update all
fields.

We also correctly setup the drivers so that when the attenuator is
powered up, all four channels are correctly initialised at -40dB.

Also ask separately for serial number to facilitate the use of barcode scanners

in case the device still has the FTDI VID and PID, the ftdi_sio module
might be loaded, in which case some of the udev matches will not have
attributes such as "busnum", "devnum" (and certainly not "serial").

We can be more careful by checking if  udev_device_get_sysattr_value()
returns NULL before doing the strcpy() to avoid segfaults.

to make sure that commit bac67a8c is conserved for reference

This merges proposed-master (with all david's modifications) with master
after correcting in master the versions of the EDA files

This merges into master the 'mise en forme' of the library
so that it is usable by FESA developers, documented and sporting a
python test program.

This implements the rewriting of a board's serial number.

Conflicts:
	software/lib/Makefile
	software/lib/liboattnusb.c
	software/lib/liboattnusb.h
	software/lib/oautest