Issues · numpy/numpy · GitHub

Labels Milestones New issue

BUG: meson backend breaks two-stage callback-based F2PY builds that work with distutils

component: numpy.f2py

#31249

· xianrui5891 opened

on Apr 15, 2026

DOC: Clarify and document a two-stage F2PY callback workflow for larger Fortran programs

04 - Documentation

component: numpy.f2py

#31248

· xianrui5891 opened

on Apr 15, 2026

DISCUSS,DRAFT: Flexible `dtype.descr` and name registration

15 - Discussion

#31243

· seberg opened

on Apr 14, 2026

BUG: `unique_all/counts/inverse` source omits `sorted=False` despite docstring claim of equivalence

#31241

· komaksym opened

on Apr 14, 2026

ENH: port loops_minmax.dispatch.c.src to C++ Highway

component: SIMD

#31239

· csp256 opened

on Apr 13, 2026

ENH: use existing SIMD reduction for reversed unit-stride integer min/max

01 - Enhancement

component: SIMD

#31232

· csp256 opened

on Apr 13, 2026

BUG: `Generator.choice(replace=False, p=...)` silently ignores `shuffle` parameter

component: numpy.random

#31210

· sweverett opened

on Apr 10, 2026

ENH: Support importing complex `array`s from standard library

01 - Enhancement

#31204

· hpkfft opened

on Apr 10, 2026

·

Potential oversight from gh-29575: self-triggered deprecation warnings

#31192

· neutrinoceros opened

on Apr 10, 2026

·

BUG: `np.clip` has surprising object-dtype behavior for huge Python ints (`0-D` return type and `where` without `out`)

#31145

· francof2a opened

on Apr 4, 2026

BUG: `np.timedelta64()` and `np.datetime64()` do not raise any warning

#31103

· riku-sakamoto opened

on Mar 30, 2026

ENH: Improve performance of reductions like `np.sum`

01 - Enhancement

#31098

· eendebakpt opened

on Mar 29, 2026