<div class="gmail_quote">On 9 September 2011 06:47, Paul Berry <span dir="ltr">&lt;<a href="mailto:stereotype441@gmail.com">stereotype441@gmail.com</a>&gt;</span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="gmail_quote"><div><div></div><div class="h5"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Paul,<br>
<br>
This is not a windows specific bug. It looks like numpy creates many identical type objects on all OSes, so type object identity/equality means nothing.<br>
<br>
Take for example what &quot;pydoc numpy&quot; outputs on my Linux system:<br>
<br>
CLASSES<br>
    __builtin__.object<br>
        broadcast<br>
        dtype<br>
        flatiter<br>
        generic<br>
[...]<br>
            number<br>
[...]<br>
                integer<br>
                    signedinteger<br>
                        int16<br>
                        int16<br>
                        int32<br>
                        int32<br>
                        int64(signedinteger, __builtin__.int)<br>
                        int64(signedinteger, __builtin__.int)<br>
                        int64(signedinteger, __builtin__.int)<br>
                        int64(signedinteger, __builtin__.int)<br>
                        int64(signedinteger, __builtin__.int)<br>
                        int8<br>
                        int8<br>
                    unsignedinteger<br>
                        uint16<br>
                        uint16<br>
                        uint32<br>
                        uint32<br>
                        uint64<br>
                        uint64<br>
                        uint64<br>
                        uint64<br>
                        uint64<br>
                        uint8<br>
                        uint8<br>
<br>
You were just lucky on Linux so far, and Windows unlucky, but your luck may change with different numpy versions, and or other systems. Who knows.<br></blockquote><div><br></div></div></div><div>Wow.  That is enough to convince me that my change is too brittle.  Good call, Jose.</div>

<div><br></div><div>Do you have any idea if the developers of numpy know about this?  It seems like a bug to me, and I want to file a bug report, but unfortunately the server at <a href="http://projects.scipy.org" target="_blank">projects.scipy.org</a> is giving me errors.  I&#39;ll try again later today.</div>
</div></blockquote><div><br></div><div>Ok, bug report filed, if anyone is curious to follow along: <a href="http://projects.scipy.org/numpy/ticket/1950">http://projects.scipy.org/numpy/ticket/1950</a></div><div><br></div>
</div>