Use the compiler flag

    -DNTL_TRANSITION

to help with the transition to NTL 3.0 from earlier versions.

This has the effect of undeclaring certain functions
whose semantics in 3.0 is different than in versions < 3.0.
Thus, one can use the compiler to find the trouble spots.

THE LIBRARY CAN NOT BE COMPILED WITH THIS FLAG!
ONLY USE TO FIND TRANSITION PROBLEMS IN CLIENT CODE.

Undeclares all shift operators for NTL arithmetic type;
in versions < 3.0, << was a conversion operator; now it
is a shift operator.

Undeclares division functions in ZZX;  in versions < 3.0,
these were defined in terms of pseudo-division; now they are
defined as ordinary division with an error being raised if the
result is not integral.  Explicit pseudo-division functions
are now provided for the old semantics.

Undeclares the UpdateMap function in for ZZ_pX and zz_pX;
in versions < 3.0, the output always had length n;
now high-order zeroes are stripped.

Undeclares the conversion from GF2X to GF2EX functions;
in versions < 3.0, this was defined as creating a constant
polynomial by reduction modulo GF2E::modulus();
now, it is defined as a coefiicient-wise "lift".
GF2X and GF2EX happened to be called BB and BB_pX in 
versions < 3.0.

Declares assignment and copy for RR to be private.  The
semantics of these have changed from "copy and round to
current precision" to "exact copy".