0.3

CHANGES:

• optimize several functions in Arith:
  • mul
  • pow for small bases
  • log2 and log2sup (5x faster)
  • log and logsup for bases 2, 16, 64, 10, 60
  • valuation_of_2
  • number_of_bits_set (7x faster)
• fix overflow in Arith.mul_quo (renamed to mul_equo) and Arith.gcdext
• let Arith.gcdext return minimal coefficients
• add Arith.sdiv
• add Arith.mul_{ediv,erem}
• add Arith.{gcd,gcdext,lcm}_of_seq
• add functions related to integer powers and roots:
  • Arith.isqrt_if_square
  • Arith.is_pow
  • Arith.is_pow2
  • Arith.kth_root
  • Arith.is_kth_pow
  • Arith.smallest_root
• add some classical arithmetic functions whose computation uses factorization:
  • Primes.divisor_pairs
  • Primes.sum_of_divisors
  • Primes.jordan
  • Primes.carmichael
  • Primes.mobius
  • Primes.derivative
• add Primes.order, Primes.order_with_known_multiple, Primes.order_mod_prime_pow
• Primes.factors now performs some iterations of Fermat’s factor searching
• BREAKING: rename sequence-related functions:
  • Arith.sum_seq -> sum_of_seq
  • Arith.prod_seq -> prod_of_seq
  • Primes.prime_seq -> gen_primes
  • Primes.primes -> iter_primes
• BREAKING: rename logarithm functions:
  • Arith.log -> Arith.ilog
  • Arith.log2 -> Arith.ilog2
  • Arith.logsup -> Arith.ilogsup
  • Arith.log2sup -> Arith.ilog2sup
• BREAKING: rename Arith.mul_quo to mul_equo
• BREAKING: add notation Arith.( ** ) for integer exponentiation
• fix Arith.range_down not being exposed in the interface