Dancing links random failures

[vbraun]

The output of one_solution is random:

**********************************************************************
File "src/sage/combinat/matrices/dancing_links.pyx", line 27, in sage.combinat.matrices.dancing_links
Failed example:
    sorted(x.one_solution(ncpus=8))
Expected:
    [0, 1]
Got:
    [2, 3]
**********************************************************************
1 item had failures:
   1 of  20 in sage.combinat.matrices.dancing_links
    [175 tests, 1 failure, 8.38 s]
----------------------------------------------------------------------
sage -t --long src/sage/combinat/matrices/dancing_links.pyx  # 1 doctest failed
----------------------------------------------------------------------

Also, does ncpus=8 really use 8 cores? We already parallelize doctests! Tests should definitely not x8 that.

CC: @seblabbe @videlec

Component: combinatorics

Keywords: random_fail

Author: Sébastien Labbé

Branch/Commit: 2df8c97

Reviewer: Volker Braun

Issue created by migration from https://trac.sagemath.org/ticket/24424

[seblabbe]

Commit: 2df8c97

[seblabbe]

Branch: u/slabbe/24424

[seblabbe]

New commits:

2df8c97

24424: fixing random doctests failures

[seblabbe]

Replying to @vbraun:

Also, does ncpus=8 really use 8 cores? We already parallelize doctests! Tests should definitely not x8 that.

I changed 8 to 2. The dancing links problem in question uses only 6 rows. So, it is not something that will use a lot of cpu time (something like 1ms).

[vbraun]

Reviewer: Volker Braun

[vbraun]

Author: Sébastien Labbé

[vbraun]

Changed branch from u/slabbe/24424 to 2df8c97