<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On Tue, Jan 3, 2017 at 9:00 AM, Yaroslav Halchenko <span dir="ltr"><<a href="mailto:lists@onerussian.com" target="_blank">lists@onerussian.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Sorry for coming too late to the discussion and after PR "addressing"<br>
the issue by issuing an error was merged [1]. I got burnt by new<br>
behavior while trying to build fresh pandas release on Debian (we are<br>
freezing for release way too soon ;) ) -- some pandas tests failed since<br>
they rely on previous non-erroring behavior and we got numpy 1.12.0~b1<br>
which included [1] in unstable/testing (candidate release) now.<br>
<br>
I quickly glanced over the discussion but I guess I have missed<br>
actual description of the problem being fixed here... what was it??<br>
<br>
previous behavior, int**int->int made sense to me as it seemed to be<br>
consistent with casting Python's pow result to int, somewhat fulfilling<br>
desired promise for in-place operations and being inline with built-in<br>
pow results as far as I see it (up to casting).</blockquote><div><br></div><div>I believe this is exactly the behavior we preserved. Rather, we turned some cases that previously often gave wrong results (involving negative integer powers) into errors.</div><div><br></div><div>The pandas test suite triggered this behavior, but not intentionally, and should be fixed in the next release:</div><div><a href="https://github.com/pandas-dev/pandas/pull/14498">https://github.com/pandas-dev/pandas/pull/14498</a><br></div></div></div></div>