Hi, Could someone with the proper Trac permissions close the following tickets: http://scipy.org/scipy/numpy/ticket/884 http://scipy.org/scipy/numpy/ticket/902 Fixed in trunk, when fix_float_format was merged back. http://scipy.org/scipy/numpy/ticket/955 Fixed in trunk + tests. http://scipy.org/scipy/numpy/changeset/6283 Addressed in trunk? Also, the following could be checked/reviewed: http://scipy.org/scipy/numpy/ticket/975 Possibly also fixed? http://scipy.org/scipy/numpy/ticket/629 Does the fix_float_format branch address printing longdouble nans (which I guess was the last issue)? Could the currently commented-out test be uncommented, cf. at the end of numpy/core/tests/test_scalarmath.py? Is float32 still a problem? (The test passes for me.) -- Pauli Virtanen
On Sat, Feb 28, 2009 at 9:11 PM, Pauli Virtanen <pav@iki.fi> wrote:
Hi,
Could someone with the proper Trac permissions close the following tickets:
http://scipy.org/scipy/numpy/ticket/884 http://scipy.org/scipy/numpy/ticket/902 Fixed in trunk, when fix_float_format was merged back.
http://scipy.org/scipy/numpy/ticket/955 Fixed in trunk + tests.
Closed.
http://scipy.org/scipy/numpy/changeset/6283 Addressed in trunk?
I think there are issues with this on windows (removing files with open handles fail), I will check and comment/close depending on the result.
Also, the following could be checked/reviewed:
http://scipy.org/scipy/numpy/ticket/975 Possibly also fixed?
http://scipy.org/scipy/numpy/ticket/629 Does the fix_float_format branch address printing longdouble nans (which I guess was the last issue)?
Yes it should fix this long-standing issue, but only in some cases I believe (when printing is not done through our work in numpyos.c). I will still check that the test works, at least. David
Sat, 28 Feb 2009 21:30:27 +0900, David Cournapeau wrote: [clip: #951]
I think there are issues with this on windows (removing files with open handles fail), I will check and comment/close depending on the result. [clip: #629] Yes it should fix this long-standing issue, but only in some cases I believe (when printing is not done through our work in numpyos.c). I will still check that the test works, at least.
Thanks! -- Pauli Virtanen
participants (2)
-
David Cournapeau -
Pauli Virtanen