andreas pushed a commit to branch core-updates
in repository guix.

commit 60594711e944413ac0d18687cce6828707ca8ecf
Author: Kaelyn Takata <[email protected]>
AuthorDate: Fri Apr 14 19:59:42 2023 +0000

    gnu: python-numpy: Skip failing tests on i686.
    
    * gnu/packages/python-xyz.scm (python-numpy): Skip failing tests on i686.
    
    Signed-off-by: Andreas Enge <[email protected]>
---
 gnu/packages/python-xyz.scm | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/gnu/packages/python-xyz.scm b/gnu/packages/python-xyz.scm
index 0f2b801046..5dae483e56 100644
--- a/gnu/packages/python-xyz.scm
+++ b/gnu/packages/python-xyz.scm
@@ -135,6 +135,7 @@
 ;;; Copyright © 2023 Gabriel Wicki <[email protected]>
 ;;; Copyright © 2023 Amade Nemes <[email protected]>
 ;;; Copyright © 2023 Bruno Victal <[email protected]>
+;;; Copyright © 2023 Kaelyn Takata <[email protected]>
 ;;;
 ;;; This file is part of GNU Guix.
 ;;;
@@ -6421,6 +6422,15 @@ include_dirs = ~:*~a/include~%"
                               ;; x86_64 CPUs such as the Core 2 Duo (see:
                               ;; https://github.com/numpy/numpy/issues/22170).
                               "and not test_rint_big_int "
+                              ;; The huge_array test is too large for 32-bit 
(see:
+                              ;; https://bugs.gentoo.org/843599 and
+                              ;; https://bugs.gentoo.org/846548).
+                              ;; TestKind.test_all is a Fortran type failure
+                              ;; that may be toolchain or environment related.
+                              #$@(if (or (target-x86?) (target-arm32?))
+                                     `(" and not 
test_identityless_reduction_huge_array"
+                                       " and not (TestKind and test_all)")
+                                   '())
                               ;; These tests seem to fail on machines without
                               ;; an FPU is still under investigation upstream.
                               ;; https://github.com/numpy/numpy/issues/20635

Reply via email to