Python using sympy 1.3 core dumps when calling rubi_integrate

[nasser1]

I am getting an Aborted (core dumped) from Python itself when using rubi_integrate in sympy 1.3

>which python
/opt/anaconda/bin/python
>python --version
Python 3.7.0

Here are the steps to reproduce it

>python
Python 3.7.0 (default, Jun 28 2018, 13:15:42) 
[GCC 7.2.0] :: Anaconda, Inc. on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> from sympy import *
>>> from sympy.integrals.rubi.rubi import rubi_integrate 
>>> d,a,x,r,a,b,c,n,e=symbols('d a x r a b c n e')
>>> rubi_integrate((d+e*x**r)**3*(a+b*ln(c*x**n))/x,x)
Fatal Python error: Cannot recover from stack overflow.

Current thread 0x00007f35940eb740 (most recent call first):
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/basic.py", line 321 in __eq__
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/expr.py", line 177 in __div__
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/decorators.py", line 132 in binary_op_wrapper
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/decorators.py", line 91 in __sympifyit_wrapper
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/numbers.py", line 697 in __div__
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/decorators.py", line 91 in __sympifyit_wrapper
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/numbers.py", line 1649 in __div__
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/decorators.py", line 91 in __sympifyit_wrapper
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/mul.py", line 1074 in _combine_inverse
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/operations.py", line 200 in _matches_commutative
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/mul.py", line 965 in matches
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/operations.py", line 205 in _matches_commutative
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/add.py", line 397 in matches
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/operations.py", line 217 in _matches_commutative
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/add.py", line 397 in matches
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/power.py", line 1308 in matches
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/operations.py", line 205 in _matches_commutative
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/mul.py", line 965 in matches
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/operations.py", line 217 in _matches_commutative
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/mul.py", line 965 in matches
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/operations.py", line 215 in _matches_commutative
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/add.py", line 397 in matches
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/operations.py", line 205 in _matches_commutative
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/add.py", line 397 in matches
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/operations.py", line 217 in _matches_commutative
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/add.py", line 397 in matches
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/core/basic.py", line 1541 in match
  File "/opt/anaconda/lib/python3.7/site-packages/sympy/integrals/rubi/utility_function.py", line 1836 in CollectReciprocals
  ...
Aborted (core dumped)
>

Running on Linux Manjaro 18

Thank you
--Nasser

[asmeurer]

In SymPy master I get

>>> rubi_integrate((d+e*x**r)**3*(a+b*ln(c*x**n))/x,x)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "./sympy/integrals/rubi/rubi.py", line 308, in rubi_integrate
    expr = expr.replace(sym_exp, exp)
NameError: name 'sym_exp' is not defined

Anyway, a core dump indicates a bug in Python, since we are not using any extension modules. Since it's a stack overflow, it means that even if Python had not crashed, it would have resulted in a RecursionError.

[asmeurer]

Oh I forgot Rubi requires Python 3.6. So apparently there is a bug that Rubi doesn't properly indicate that 3.5 isn't supported. This is what I get on 3.6:

>>> rubi_integrate((d+e*x**r)**3*(a+b*ln(c*x**n))/x,x)
Traceback (most recent call last):
  File "./sympy/polys/polyutils.py", line 209, in _parallel_dict_from_expr_if_gens
    monom[indices[base]] = exp
KeyError: (1/x)**(1/3)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "./sympy/integrals/rubi/rubi.py", line 322, in rubi_integrate
    results = rubi.replace(Integral(expr, var), max_count = 10)
  File "/Users/aaronmeurer/anaconda3/envs/python36/lib/python3.6/site-packages/matchpy/matching/many_to_one.py", line 796, in replace
    replacement, subst = next(iter(self.matcher.match(subexpr)))
  File "/Users/aaronmeurer/anaconda3/envs/python36/lib/python3.6/site-packages/matchpy/matching/many_to_one.py", line 99, in __iter__
    for _ in self._match(self.matcher.root):
  File "/Users/aaronmeurer/anaconda3/envs/python36/lib/python3.6/site-packages/matchpy/matching/many_to_one.py", line 148, in _match
    yield from self._match_transition(transition)
  File "/Users/aaronmeurer/anaconda3/envs/python36/lib/python3.6/site-packages/matchpy/matching/many_to_one.py", line 162, in _match_transition
    yield from self._match_regular_operation(transition)
  File "/Users/aaronmeurer/anaconda3/envs/python36/lib/python3.6/site-packages/matchpy/matching/many_to_one.py", line 310, in _match_regular_operation
    for new_state in self._check_transition(transition, subject, False):
  File "/Users/aaronmeurer/anaconda3/envs/python36/lib/python3.6/site-packages/matchpy/matching/many_to_one.py", line 209, in _check_transition
    yield from self._match(transition.target)
  File "/Users/aaronmeurer/anaconda3/envs/python36/lib/python3.6/site-packages/matchpy/matching/many_to_one.py", line 148, in _match
    yield from self._match_transition(transition)
  File "/Users/aaronmeurer/anaconda3/envs/python36/lib/python3.6/site-packages/matchpy/matching/many_to_one.py", line 160, in _match_transition
    yield from self._match_commutative_operation(transition.target)
  File "/Users/aaronmeurer/anaconda3/envs/python36/lib/python3.6/site-packages/matchpy/matching/many_to_one.py", line 292, in _match_commutative_operation
    self._check_constraints(variable, restore_constraints, restore_patterns)
  File "/Users/aaronmeurer/anaconda3/envs/python36/lib/python3.6/site-packages/matchpy/matching/many_to_one.py", line 235, in _check_constraints
    if not constraint(self.substitution):
  File "/Users/aaronmeurer/anaconda3/envs/python36/lib/python3.6/site-packages/matchpy/expressions/constraints.py", line 190, in __call__
    return self.constraint(**args)
  File "./sympy/integrals/rubi/constraints.py", line 3211, in cons_f590
    return PseudoBinomialPairQ(u*x**(m/p), v, x)
  File "./sympy/integrals/rubi/utility_function.py", line 2766, in PseudoBinomialPairQ
    lst1 = PseudoBinomialParts(u, x)
  File "./sympy/integrals/rubi/utility_function.py", line 2748, in PseudoBinomialParts
    if PolynomialQ(u, x) and Greater(Expon(u, x), S(2)):
  File "./sympy/integrals/rubi/utility_function.py", line 1059, in PolynomialQ
    deg = degree(u, x.base)
  File "./sympy/polys/polytools.py", line 4458, in degree
    p, _ = poly_from_expr(f, gen)
  File "./sympy/polys/polytools.py", line 4266, in poly_from_expr
    return _poly_from_expr(expr, opt)
  File "./sympy/polys/polytools.py", line 4288, in _poly_from_expr
    rep, opt = _dict_from_expr(expr, opt)
  File "./sympy/polys/polyutils.py", line 366, in _dict_from_expr
    rep, gens = _dict_from_expr_if_gens(expr, opt)
  File "./sympy/polys/polyutils.py", line 305, in _dict_from_expr_if_gens
    (poly,), gens = _parallel_dict_from_expr_if_gens((expr,), opt)
  File "./sympy/polys/polyutils.py", line 215, in _parallel_dict_from_expr_if_gens
    "the set of generators." % factor)
sympy.polys.polyerrors.PolynomialError: (1/x)**(1/3) contains an element of the set of generators.

[RituRajSingh878]

In SymPy master I get

>>> rubi_integrate((d+e*x**r)**3*(a+b*ln(c*x**n))/x,x)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "./sympy/integrals/rubi/rubi.py", line 308, in rubi_integrate
    expr = expr.replace(sym_exp, exp)
NameError: name 'sym_exp' is not defined

Anyway, a core dump indicates a bug in Python, since we are not using any extension modules. Since it's a stack overflow, it means that even if Python had not crashed, it would have resulted in a RecursionError.

@asmeurer can you give me some suggestions to solve this issue.

[asmeurer]

I think getting rubi to give a proper error in Python < 3.6 would be a good start.