FLINT 3.3

FLINT 3.3.1 (2025-06-16)#

• Bug fixes

  • Set permutation vector in gr_mat_lu_classical even if there are zero columns (Fredrik Johansson).

FLINT 3.3.0 (2025-06-11)#

Main contributors: Albin Ahlbäck (AA), Ricardo Buring (RB), Lars Göttgens (LG), Max Horn (MH), Fredrik Johansson (FJ), Jean Kieffer (JK), Marc Mezzarobba (MM), Vincent Neiger (VN).

• Features

  • Add generic fraction fields (gr_ctx_init_gr_fraction) (FJ).
  • Allow rings to overload _gr_mat_charpoly and provide such overloads for fmpz, fmpq, mpn_mod and fmpz_mod. Furthermore, remove fq_*_mat_charpoly_danilevsky and fmpz_mod_mat_charpoly_berkowitz, and wrap fq_*_mat_minpoly and fq_*_mat_charpoly around GR functions (FJ).
  • Remove fq_zech_mat_reduce_rows (FJ).
  • Add various mulmod, div_preinv and compose_mod methods for gr_poly (FJ).
  • Add fmpz_mod_mat_pow_ui (LG).
  • Add gr_poly_scalar_mul (FJ).
  • Implement GR_METHOD_OTHER_MUL for gr_poly (FJ).
  • Add gr_mat_pow_(si|fmpz|fmpq|scalar) (LG, FJ).
  • Add gr_mat_func_jordan (FJ).
  • Add nmod_poly_mat_transpose and fq_default_mat_transpose (LG).
  • Add different powmod-methods for gr_poly, based on methods from fq_poly (FJ).
  • Add gr_poly_(add|sub)_scalar (RB).
  • Support GR objects in flint_printf (FJ).
  • Add support for parsing builtin functions like sqrt(x) in gr_set_str (FJ).
  • Recognize when generic polynomial rings are UFDs (MM).
  • Add flint_set_throw (AA).
  • Add gr_generic_poly_factor_roots for factoring generic polynomials over algebraically closed fields providing gr_poly_roots (MM).
  • Add gr_factor for gr_poly (MM).
  • Add arf_nint (FJ).
  • Add gr_poly_gcd_subresultant (FJ).
  • Add generic gr_gcd for fields (FJ).
  • Add gr_mpoly_inv (FJ).
  • Implement gr_poly_gcd over any UFD, not only fields (FJ).
  • Add gr_canonical_associate and corresponding implementations for several types (FJ).
  • Add gr_mat_companion and gr_mat_companion_fraction to construct companion matrices (RB).
  • Add gr_randtest_invertible (RB).
  • Add gr_mat_randsimilar (RB).
  • Add flint_merge_sort and flint_sort (MM).
  • Add gr_vec_sort for sorting generic vectors (MM).
  • Add _gr_vec_shuffle, gr_vec_permute, gr_vec_contains (MM).
  • Add gr_in_domain and gr_check (MM).
  • Add gr_poly_leading_taylor_shift (MM).
  • Add shift equivalence, dispersion, and shiftless decomposition of generic polynomials (gr_poly_shift_equivalent, gr_poly_dispersion, gr_poly_shiftless_decomposition) (MM).
  • Add gr_mat_mul_rosowski (FJ).
  • Add acb_theta_one to compute a single theta function (JK).
  • Add parameter ab in acb_theta_jet (JK).

• Examples

  • Add example script to check n_is_prime against list of Fermat pseudoprimes (FJ).
  • Replace generated code with strings in huge_expr.c (FJ).

• Performance

  • Use stride instead of row pointers in all matrix types (FJ).
  • Refactor characteristic polynomial and retune it (FJ).
  • Use delayed canonicalization in mpn_mod_mat_reduce_row and fmpz_mod_mat_reduce_row (FJ).
  • Improve method choice in gr_mat_rank (FJ).
  • Inline operations in _gr_fmpz_set (FJ).
  • Use udiv_qrnnd in n_preinvert_limb whenever CPU has fast division (AA, FJ, VN).
  • Have gr_divexact over a gr_poly ring dispatch to gr_poly_divexact (FJ).
  • Rewrite the acb_theta module for substantially improved performance (JK).
  • Overload _gr_vec_mul for nmod types (FJ).
  • Optimize _fmpz_vec_add and _fmpz_vec_sub for small coefficients by inlining (FJ).

• Bug fixes

  • Fix bug in fmpz_mat_charpoly_modular where wrong results were given if output polynomial was not initially set to zero by caller (FJ).
  • Fix memory leak in gr_mat_minpoly_field (FJ).
  • Fix aliasing for gr_poly_div_newton_n_preinv (FJ).
  • Fix memory leak in gr_mat_fflu (FJ).
  • Fix invalid write in _gr_nmod_vec_reciprocals when length is zero (FJ).
  • Fix various memory leaks found by the address sanitizer (RB).
  • Fix cases where gr_poly were not correctly normalized over a degenerate ring (FJ).
  • Fix gr_set_str with duplicate generator names (MM).
  • Fix fmpz_mat_minpoly and fmpq_mat_minpoly to return the correct result for the zero matrix (MH).
  • Fix bug in arb_nint (FJ).
  • Fix fmpz_mat_minpoly_modular with large coefficients (FJ).
  • Fix context bug in _gr_gr_series_mod_gen (RB).
  • Propagate failure in acb_poly_find_roots if numerical stage returns NaN (FJ).
  • Fix bug in arb_div giving NaN when the numerator contained zero at high precision (FJ).
  • Fix NaN output for [+/-inf +/- inf] in arb_get_interval_arf and arb_get_interval_mpfr (FJ).
  • Correct signature of gr_ctx_clear to have void return type (FJ).
  • Fix function signatures in internal callbacks for flint_parallel_binary_splitting (Oscar Benjamin, FJ).

• Build system

  • Push pragmas into macros, allowing them to be set during configuration (AA).
  • Fix an overwriting of CFLAGS in configure.ac (AA).
  • Re-enable -Wmaybe-uninitialized (FJ).
  • Enable -Werror=newline-eof (AA).
  • Accommodate for Meteor Lake, Raptor Lake and Tiger Lake (VN).
  • Make sure Comet Lake enables fft_small (VN).
  • Error when trying to build FLINT on a non-Windows system (AA).
  • Do not use find_library() in flint_ctypes (MM).

• Tests

  • Add more test code for preinv division for gr_poly (FJ).
  • Add tests for gr_mat_exp and gr_mat_log (FJ).
  • Add test for fmpz_mod_mat_transpose (LG).
  • Test gr_poly over random rings (MM).
  • Have gr_ctx_init_random generate polynomial contexts (FJ).
• Profiling

• Maintenance

  • Replace some fq_poly and fmpz_mod_poly methods with GR wrappers (FJ).
  • Include standard headers instead of using builtins (AA).
  • Refactor gr_mat_exp_jordan and gr_mat_log_jordan (FJ).
  • Delegate (fmpz|fmpz_poly|fmpq)_mat_transpose to gr_mat (LG).
  • Add issue templates to Github (AA, Lars Kastner).
  • Optimize for code size in some less performance-critical functions (FJ).
  • Fix various undefined behaviors reported by GCC's address sanitizer (FJ).
  • Add FLINT_UABS macro (FJ).
  • Refactor nullspace and rref in GR, and make fq_*_mat use these instead (FJ).
  • Use gr_poly-methods for powmod in fmpz_mod_poly and fq_*_poly (FJ).
  • Use gr_mat_rank_lu in fmpz_mod_mat_rank and fq_*_mat_rank (FJ).
  • Remove underscore in front of nfloat_ctx_(set|get)_real_prec to match documentation (Joel Dahne).
  • Add missing newlines at end of file (FJ).
  • Fix ENABLE_AVX2=OFF for CMake (AA).
  • Add gr_poly_coeff_ptr and deprecate gr_poly_entry_ptr (RB).

• Continuous integration

  • Use temporary Nemo branch for CI due to the change of using stride instead of row pointers in matrices (AA).

• Documentation

  • Fix fmpz docstrings to clarify that GCD of integers and rationals is non-negative (MH).
  • Unify some docstrings (LG).
  • Fix docstrings for various LU functions (LG).
  • Fix typos (AA, MM, user202729).
  • Touch up formula in hypgeom.rst (FJ).
  • Fix docstring for fmpz_mat_snf_kannan_bachem (Dima Pasechnik).
  • Fix docstring for fmpz_get_nmod (David Lowry-Duda).
  • Minor fix in documentation of ulong_extras (VN).
  • Document n_factor_t structure (Matthew House).
  • Document gr_vec_entry_srcptr (MM).
  • Document truth_t functions (MM).

@misc{flint-doc-history-flint3-v3-3,
  author       = {{The FLINT team}},
  title        = {{FLINT 3.3}},
  howpublished = {FLINT: Fast Library for Number Theory, documentation, version 3.7.0-dev},
  year         = {2026},
  url          = {https://flintlib.org/doc/history/flint3/v3-3.html},
  note         = {Generated 2026-10-20}
}