Number Formats: The Representation of Large Numbers in C - Cryptography in C and C++ - page 15

For some functions it is necessary to process numbers that have more digits
than can be accommodated by a CLINT object. For these cases the variants of the
CLINT type are defined by
typedef unsigned short CLINTD[1+(CLINTMAXDIGIT<<1)];
and
typedef unsigned short CLINTQ[1+(CLINTMAXDIGIT<<2)];
which can hold double, respectively four times, the number of digits.
As support personnel to aid in programming, the module flint.c defines
the constants nul_l , one_l ,and two_l , which represent the numbers 0, 1, and 2
in CLINT format; and in flint.h there are corresponding macros SETZERO_L() ,
SETONE_L() ,and SETTWO_L() , which set CLINT objects to the corresponding values.