I've run across a strange bug that seems to be hardware dependent: it is
repeatable on an Ubuntu server, but I can't replicate it on my Macbook Pro
or on Cocalc. It occurs when factoring rational polynomials, and gives an
error message from pari: "bug in gerepile, significant pointers lost,
please report" (full traceback and input below). To trigger the bug, one
factors two different polynomials consecutively (factoring the second on
its own does not trigger the problem).
The bug seems to be quite sensitive to the polynomials involved. Deleting
any subset of factors seems to not trigger the bug, and multiplying through
by the denominator causes it to disappear as well (though other examples
still trigger the bug even when all coefficients are integral).
When I try factoring in gp, I don't get an error (thus why I'm writing to
sage-devel rather than Pari).
If anyone has ideas on how to diagnose this, or reproduce it in a format
more easily reportable to the Pari developers, I'd be happy to hear it. In
the meantime, calling squarefree_decomposition and then factoring each of
the resulting terms seems to be a functional workaround for my purposes.
David
*System info:*
Freshly built Sage 8.7
Ubuntu 18.04.1 LTS (GNU/Linux 4.15.0-30-generic x86_64)
CPUs: 48 Intel Xeon CPU E5-2697 v2 @ 2.70GHz
RAM: 256GB in 16 Hynix DDR3 DIMMs (the bug occurs when only a small amount
of memory is in use)
Hard drive: Samsung SSD 840 EVO 1TB (381GB free)
*Input and traceback:*
sage: R. = QQ[]
sage: f = (x - 1)^8 * (x^2 - 2/3*x + 1)^2 * (x^2 + 2/3*x + 1)^2 * (x^4 +
3*x^3 + 37/9*x^2 + 3*x + 1) * (x^4 + 1/3*x^3 - x^2 + 1/3*x + 1)^2 * (x^16 +
2*x^14 + 211/81*x^12 + 28/9*x^10 + 269/81*x^8 + 28/9*x^6 + 211/81*x^4 +
2*x^2 + 1)^2
sage: g = (x - 1)^32 * (x^2 + 5/3*x + 1)^16
sage: factor(f)
(x - 1)^8 * (x^2 - 2/3*x + 1)^2 * (x^2 + 2/3*x + 1)^2 * (x^4 + 3*x^3 +
37/9*x^2 + 3*x + 1) * (x^4 + 1/3*x^3 - x^2 + 1/3*x + 1)^2 * (x^16 + 2*x^14
+ 211/81*x^12 + 28/9*x^10 + 269/81*x^8 + 28/9*x^6 + 211/81*x^4 + 2*x^2 +
1)^2
sage: factor(g)
---
PariError Traceback (most recent call last)
in ()
> 1 factor(g)
/home/sage/sage-8.7/local/lib/python2.7/site-packages/sage/arith/misc.pyc
in factor(n, proof, int_, algorithm, verbose, **kwds)
2563 except TypeError:
2564 # Maybe the factor() method doesn't have a proof option
-> 2565 return m(**kwds)
2566
2567
/home/sage/sage-8.7/local/lib/python2.7/site-packages/sage/rings/polynomial/polynomial_element.pyx
in sage.rings.polynomial.polynomial_element.Polynomial.factor
(build/cythonized/sage/rings/polynomial/polynomial_element.c:35617)()
4252 R = self._parent.base_ring()
4253 if hasattr(R, '_factor_univariate_polynomial'):
-> 4254 return R._factor_univariate_polynomial(self, **kwargs)
4255
4256 G = None
/home/sage/sage-8.7/local/lib/python2.7/site-packages/sage/rings/rational_field.pyc
in _factor_univariate_polynomial(self, f)
1348
1349 """
-> 1350 G = list(f._pari_with_name().factor())
1351
1352 # normalize the leading coefficients
cypari2/gen.pyx in cypari2.gen.Gen.factor()
cypari2/handle_error.pyx in cypari2.handle_error._pari_err_handle()
PariError: bug in gerepile, significant pointers lost, please report
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
