I've been tracking this in a debugger.  There is a bug in the Wang
factorization algorithm.  The difference comes at line 920 of
factortools.py, where a random number is used.  I can try and see if I
can figure this out, but it probably won't be easy without
understanding of the algorithm.

Mateusz, can you look at this?  For reference, the input to dmp_zz_wang is

f = [[[mpz(2)]], [[]], [[mpz(1), mpz(-1)], [mpz(-1), mpz(1),
mpz(0)]]], u = 2, K = ZZ, mod=None.

Aaron Meurer

On Wed, May 2, 2012 at 3:47 PM, Aaron Meurer <[email protected]> wrote:
> I can reproduce this with ./bin/test --seed=13604685
> sympy/polys/tests/test_polyroots.py
>
> I put in some print statements, and it seems that roots is returning
> {1: 1, y/2 - sqrt(8*x**2 + y**2 - 2*y + 1)/2 + 1/2: 0, y/2 +
> sqrt(8*x**2 + y**2 - 2*y + 1)/2 + 1/2: 0} when it should be returning
> {1: 1, y/2 - sqrt(8*x**2 + y**2 - 2*y + 1)/2 + 1/2: 1, y/2 +
> sqrt(8*x**2 + y**2 - 2*y + 1)/2 + 1/2: 1} (the difference is the
> values).
>
> Aaron Meurer
>
> On Wed, May 2, 2012 at 3:05 PM, Joachim Durchholz <[email protected]> wrote:
>> I just got a nonreproducible assertion failure.
>> I.e. bin/test (with no additional options) gave me an AssertionError on line
>> 293 of sympy/polys/tests/test_polyroots.py.
>> Sure enough, that line compares two hashes for equality.
>>
>> Rerunning the test in isolation did not reproduce the error, nor did
>> rerunning the test suite in full reproduce.
>> It does reproduce with the right random seed, 13604685, on my current
>> workdir and on current master.
>>
>> Here's the first few lines of bin/test output:
>>
>> ============================= test process starts
>> ==============================
>> executable:   /usr/bin/python  (2.7.3-final-0)
>> architecture: 64-bit
>> cache:        yes
>> ground types: python
>> random seed:  13604685
>>
>> What now?
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "sympy" group.
>> To post to this group, send email to [email protected].
>> To unsubscribe from this group, send email to
>> [email protected].
>> For more options, visit this group at
>> http://groups.google.com/group/sympy?hl=en.
>>

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/sympy?hl=en.

Reply via email to