[Python-modules-team] Bug#679948: Acknowledgement (regression (fixed upstream) in deducing the dtype for record array)

[Yaroslav Halchenko]

ok -- bug bisected to  be fixed in

$> PYTHONPATH=$PWD python -c "import numpy; print numpy.__version__; print numpy.array([('a', 1), ('b', 9.99)])"
1.7.0.dev-91f87e1                                                                                         
[['a' '1']
 ['b' '9.99']]

but unfortunately 1.7.0.dev-91f87e1 cannot be applied as is to 1.6.x series :-/
thus I would appreciate if someone numpy-source-code aware person could look
into adopting this fix.

commit 91f87e1f613630ff0ad9864017f059afcd6e57f1
Author: Bryan Van de Ven <bryan at Laptop-3.local>
Date:   Mon Feb 20 11:32:57 2012 -0600

    BUG: Fix ticket #1990.
    
    When an array is created from a sequence of numeric (boolean, int, float,
    complex) and string (bytes, str, unicode) values, the resulting array type
    is string, but only the string values were being used to choose the string
    length, leading to truncation of data.



-- 
Yaroslav O. Halchenko
Postdoctoral Fellow,   Department of Psychological and Brain Sciences
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
Phone: +1 (603) 646-9834                       Fax: +1 (603) 646-1419
WWW:   http://www.linkedin.com/in/yarik