1 | | Another source of test failures I've found--possibly different from the PRNG differences--is that iteration order over a number of objects seems to be different; particularly iterating over some matrix groups. I'm not sure why, or if it's expected. But it seems mostly arbitrary--you get the same group elements just in a different order. |

| 1 | Another source of test failures I've found--possibly different from the PRNG differences--is that iteration order over a number of objects seems to be different; particularly iterating over some matrix groups. I'm not sure why, or if it's expected. But it seems mostly arbitrary--you get the same group elements just in a different order. For example, in the case of: |

| 2 | |

| 3 | {{{ |

| 4 | sage: SGA = SymmetricGroupAlgebra(QQ, WeylGroup(["A",3], prefix='s')) |

| 5 | sage: SGA.an_element() |

| 6 | }}} |

| 7 | |

| 8 | I now get |

| 9 | |

| 10 | {{{ |

| 11 | s1*s2*s3 + 3*s3*s2 + 2*s3 + 1 |

| 12 | }}} |

| 13 | |

| 14 | where previously it returned |

| 15 | |

| 16 | {{{ |

| 17 | 2*s1*s2*s3*s2*s1 + 3*s1*s2*s3*s1 + s1*s2*s3 + 1 |

| 18 | }}} |

| 19 | |

| 20 | This is of course meant to be somewhat arbitrary, though it should be deterministic (and it ''is'', the result just changed due to some change in GAP when iterating over the underlying matrix group). |