Possible bug in PARI handling...

[Emmanuel Charpentier]

Inspiration is this ask.sagemath question. Roughly, the original question was a discrepancy between the numerical results obtained by Sage and by Matlab when solving a set of three equations with three unknowns.

I wanted to compare the numerical results with those obtained by rearranging the original equations into rational fractions and solving their numerators (while watching out for null denominators). My code so far is in the attached SolutionExacte.sage.

What succeeds :

• Rearranging the equations in rational fractions
• creating the relevant polynomial ring
• creating the ideal J representing the set of solutions
• checking that its dimension is 0

What fails is enumerating the solutions : S=J.variety(QQbar) hasn’t returned for more than eight hours (!). When throwing the towel (*i. e. interrupting), I got the following message in a message box of the Jupyter notebook :

/usr/local/sage-9/local/lib/python3.7/site-packages/sage/rings/qqbar.py:2078: RuntimeWarning: cypari2 leaked 2081168 bytes on the PARI stack df = factor(d)

Questions :

1. Is that a known bug ?
2. Is that possibly a new bug ?
3. If so, how to file the most informative ticket ?
4. Are you aware of possible workarounds ?

HTH,

​

[Dima Pasechnik]

I guess it would be Singular, not Pari.

--
You received this message because you are subscribed to the Google Groups "sage-support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-support...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sage-support/ab9e3015-f0cf-43a8-84b3-5c9c7c12d1bao%40googlegroups.com.