Repository navigation

All pull requests

Labels (97)Milestones (2)New pull request

Open

(311)

Closed

(18,350)

BUG: linalg: improve 2x2 determinant accuracy using Kahan's algorithm
#33064sarhank8 opened Oct 16, 2026
ENH: Limited API fallbacks for reading str data
#33063prathamhole14 opened Oct 16, 2026Member
PERF: speed up dtype.name
#33059lazerg opened Oct 16, 2026Contributor
TYP: concrete number subtypes
#33052jorenham opened Oct 15, 2026Member 2.6.0 Release
2
BUG: Supply missing self-casts for legacy-backed DTypes
#33050ngoldbaum opened Oct 15, 2026Member
3
BUG: include dtype class in finfo cache key to avoid aliasing (gh-32947)
#33049Rvish-glitch opened Oct 15, 2026Contributor
5
TYP: append: add @deprecated overloads for python scalar coversions
#33047jorenham opened Oct 15, 2026Member 2.6.0 Release
3
BUG: detect ASIMDFHM at runtime on macOS ARM64
#33040Gusmao021 opened Oct 14, 2026
DOC: clarify that "pin the current major version" means numpy<3
#33038shakrishnan-nvidia opened Oct 14, 2026
DOC: Remove <1.23 versionadded/changed directives in _function_base
#33033michaelzhan1 opened Oct 14, 2026Contributor
3
BUG: Show warning for ignored CPU feature names
#33027mikamikasuki opened Oct 14, 2026
2
BUG: raise TypeError for file-like objects without fileno in fromfile
#33017AbhiRKeesara opened Oct 13, 2026Contributor
4
DOC: document VS Code and LLDB debugging on macOS
#32998ShivBoyPro opened Oct 13, 2026
ENH: add an internal indexed copying API for take, repeat, put, and putmask
#32996ngoldbaum opened Oct 12, 2026Member
3
DOC: clarify NpzFile archive member handling
#32992travisfortney opened Oct 12, 2026
CI/BLD: don't install ninja on pyodide in the build dependency group
#32980andyfaff opened Oct 12, 2026Member
1
BUG: fft: allow out together with any s in N-D transforms
#32973jayzuccarelli opened Oct 11, 2026Contributor
2
BUG: timsort on structured dtypes
#32963tylerjereddy opened Oct 8, 2026Contributor
2
MAINT: lazy imports on Python 3.15+ (PEP 810)
#32958eendebakpt opened Oct 8, 2026Contributor
3
BUG: release temporary references during string promoter initialization
#32957sa-faizal opened Oct 8, 2026
6
ENH: Limited API fallbacks for PyTuple_GET_ITEM and other accessor macros
#32951prathamhole14 opened Oct 8, 2026Member
BUG: fix nanquantile dtype for all-NaN first slice
#32940dave123981 opened Oct 7, 2026
3
BUG: fix overflow in np.round of float16
#32916tuan2k33 opened Oct 6, 2026Contributor
BUG: close the file in np.load when an .npz is invalid
#32915aryangorde6 opened Oct 6, 2026
5
DOC: Clarify half-to-even rounding behavior for quantile 'nearest' method
#32911ryanakml opened Oct 6, 2026
2

Navigation Menu

Repository navigation

All pull requests

BUG: linalg: improve 2x2 determinant accuracy using Kahan's algorithm

ENH: Limited API fallbacks for reading `str` data

PERF: speed up dtype.name

TYP: concrete `number` subtypes

BUG: Supply missing self-casts for legacy-backed DTypes

BUG: include dtype class in finfo cache key to avoid aliasing (gh-32947)

TYP: `append`: add `@deprecated` overloads for python scalar coversions

BUG: detect ASIMDFHM at runtime on macOS ARM64

DOC: clarify that "pin the current major version" means numpy<3

DOC: Remove <1.23 `versionadded/changed` directives in _function_base

BUG: Show warning for ignored CPU feature names

BUG: raise TypeError for file-like objects without fileno in fromfile

DOC: document VS Code and LLDB debugging on macOS

ENH: add an internal indexed copying API for take, repeat, put, and putmask

DOC: clarify NpzFile archive member handling

CI/BLD: don't install ninja on pyodide in the `build` dependency group

BUG: fft: allow out together with any s in N-D transforms

BUG: timsort on structured dtypes

MAINT: lazy imports on Python 3.15+ (PEP 810)

BUG: release temporary references during string promoter initialization

ENH: Limited API fallbacks for `PyTuple_GET_ITEM` and other accessor macros

BUG: fix nanquantile dtype for all-NaN first slice

BUG: fix overflow in np.round of float16

BUG: close the file in np.load when an .npz is invalid

DOC: Clarify half-to-even rounding behavior for quantile 'nearest' method

Pull requests