Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 14 Apr 2013 22:13:42 -0400
From:      "John W. O'Brien" <john@saltant.com>
To:        freebsd-python@freebsd.org
Subject:   Re: [CFT] math/py-pandas: Update to 0.11.0rc1
Message-ID:  <516B6256.6050202@saltant.com>
In-Reply-To: <51699935.3040300@saltant.com>
References:  <51699935.3040300@saltant.com>

next in thread | previous in thread | raw e-mail | index | archive | help
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 04/13/2013 01:43 PM, John W. O'Brien wrote:
> I am seeing a few test failures on my system which should have time
> to investigate tomorrow.
> 
> pandas.tests.test_format.TestDataFrameFormatting 
> pandas.tests.test_index.TestMultiIndex

I am having trouble reproducing these reliably.

In one case (test_index:TestMultiIndex.test_legacy_pickle), the
failing test passes when I build from the github tag, but I don't know
enough to know where to look next.

https://github.com/pydata/pandas/issues/3360

In the other, stranger case
(test_format:TestDataFrameFormatting.test_to_string_repr_unicode):

    A) the test fails when I run it as part of the
TestDataFrameFormatting suite, but passes when I run it alone
    B) the test fails on one machine, but fails on another, highly
similar machine
    C) inspection of the code indicated by the traceback reveals that
a thrown UnicodeEncodeError in core.format._encode_diff() is not
caught in core.format.SeriesFormatter.to_string()

C by itself is strange enough to give me pause, because I don't think
my python-fu is yet sufficiently strong to plumb these depths.

Under what circumstances might

    try:
        func_that_always_raises_spam()
    except spam:
        eggs()

fail to invoke eggs()?

Does this smell more like a bug in pandas, a bug in related software
(e.g. nose, numpy), or some snafu in my dev't environment?

I would like to do whatever legwork is needed to ensure a high-quality
port of pandas 0.11.0 for FreeBSD, and so I will appreciate any advice
python@ would care to provide. Can anybody else reproduce these test
failures with the patch I posted? Is there a (preferred) way to run
the unit tests in redports?

Thank you.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)

iQEcBAEBAgAGBQJRa2JVAAoJEBRzAKlhyP/F7BkH/1KebkEHery/O5w6kN069a/+
ST7/ZSzuMF7k8sbV1NzBymCqCE22t6iemvLD6y+oI+JyEBKrwWjEAY8RBenyUN73
EMGpa0VJh4ZowUrITAWmeXJ73y+4wRTxXpq+Ns2lkdm01kkMcpaCC8TRzb1KDBKg
Nd7DTdjl3UvBRN2pjqXiUG8QXlZorukQjlkDAfgYI5671Gp6nY//ABLuaIfwFghK
DMbWuSHxmsRmLOEwSuWo7WUF1p+D2WKUGc+smaC9VLa4Q+ppdD1CvRiwNORZac5l
jglCMG75NuRbiVlHEafzVLKTXX8t1bRZgNeD+pXMpLpDEhaEaZyimCt2lmcnI4E=
=pHWA
-----END PGP SIGNATURE-----



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?516B6256.6050202>