[Numpy-discussion] glibc memory corruption when running numpy.test()

[Hoyt Koepke]

Attached.

I had a bunch of issues getting things to install with lapack and
ATLAS.  In the end I had to specify the following environment
variables (+ appropriate command line options to ATLAS & lapack) to
get it to work.  If there's an easier way, let me know.


export FLAGS='-march=core2 -mtune=core2 -m64'
export CFLAGS="$FLAGS"
export CPPFLAGS="$FLAGS"
export CXXFLAGS="$FLAGS"
export FFLAGS="$FLAGS"
export F77FLAGS="$FLAGS"
export LDFLAGS="$FLAGS"


When compiling numpy, I had to manually add '-shared' to LDFLAGS to
get it to work.


Thanks!
--Hoyt



On Wed, Nov 19, 2008 at 11:33 PM, David Cournapeau
<cournapeau at cslab.kecl.ntt.co.jp> wrote:
> On Wed, 2008-11-19 at 23:23 -0800, Hoyt Koepke wrote:
>> Hello,
>>
>> Sorry about this bug report.... I know how much us programmers like
>> these kind of bugs :-/.
>>
>> In testing the latest svn version of numpy (6083), I get a memory
>> corruption error:
>>
>> test_csingle (test_linalg.TestSolve) ... ok
>> test_double (test_linalg.TestSolve) ... ok
>> test_empty (test_linalg.TestSolve) ... ok
>> Check that matrix type is preserved. ... ok
>> Check that matrix type is preserved. ... ok
>> test_nonarray (test_linalg.TestSolve) ... ok
>> test_single (test_linalg.TestSolve) ... ok
>> Ticket #652 ... *** glibc detected *** python: free(): invalid next
>> size (fast): 0x0000000002c80ef0 ***
>>
>> ... and it hangs.
>>
>> I don't get any errors in my installation.  The system is a 64bit
>> intel xeon server running debian unstable, compiler is gcc-4.3.2
>
> Could you give us the complete build log (e.g. when build from scratch:
> remove the build directory and give us the log of python setup.py build
> &> build.log) ?
>
> You have tens of failures which are more likely due to a build problem,
>
> David
>
> _______________________________________________
> Numpy-discussion mailing list
> Numpy-discussion at scipy.org
> http://projects.scipy.org/mailman/listinfo/numpy-discussion
>



-- 

++++++++++++++++++++++++++++++++++++++++++++++++
+ Hoyt Koepke
+ University of Washington Department of Statistics
+ http://www.stat.washington.edu/~hoytak/
+ hoytak at gmail.com
++++++++++++++++++++++++++++++++++++++++++
-------------- next part --------------
A non-text attachment was scrubbed...
Name: build.log.gz
Type: application/x-gzip
Size: 6346 bytes
Desc: not available
URL: <http://mail.python.org/pipermail/numpy-discussion/attachments/20081119/6e114c84/attachment.bin>