"
Hello, Tom !
I think, I have found a bug in your library. This simple code demonstrates an erroneous substraction and a right substraction:

fp_int t1, t2;

//Erroneous:
fp_2expt(&t1, 511); //t1 <= 2^511
fp_set(&t2, 1); //t2 <= 1 (and t2.used <= 1)
fp_sub (&t1, &t2, &t1); //t1 <= (t1 - t2); ERRONEOUS RESULT !!!
//Right:
fp_2expt(&t1, 511); //Once more: t1 <= 2^511
fp_set(&t2, 1); //t2 <= 1
t2.used = 16; //But now (t2.used <= 16 (i.e. 512 bits))
fp_sub (&t1, &t2, &t1); //t1 <= (t1 - t2); ALL OK !!!

Can you give me a comment ?
Best regards, Igor Koval (Ukraine)
"