January 2022 - GiNaC-list

factorization bug
by Ivan Vasilyev 05 Feb '22

05 Feb '22

Hi! I believe there's an intermittent bug in GiNaC::factor() function. Sometimes it leaves uncancelled terms with seemingly random coefficients like: w*(w^2*(K^4*B^2+B^2-2*K^2*B^2)+196*w^2*K^4+196*w^2+C^2-196*w^2*(1+K^4-2*K^2)-392*w^2*K^2)*A The minimal code to reproduce the bug in GiNaC 1.8.2 is: #include <ginac/ginac.h> #include <iostream> int main() { std::string input("w^3*B^2*A-2*w^3*K^2*B^2*A+w*C^2*A+w^3*K^4*B^2*A"); while (true) { std::ostringstream s; GiNaC::parser reader; s << factor(reader(input)); if (s.str().length() <= input.length()) continue; std::cout << s.str() << '\n'; return 0; } } Also, this problem can be reproduced on https://daninet.github.io/ginac-wasm/ with the same expression being calculated several times in a row: factor(w^3*B^2*A-2*w^3*K^2*B^2*A+w*C^2*A+w^3*K^4*B^2*A); With best regards, Ivan.

3 3

Re: [GiNaC-list] to_polynomial function does not return a polynomial
by Feng Feng 14 Jan '22

14 Jan '22

Dear Vladimir, Thanks so much for checking this, and thanks for the quick update on the git! Best regards! Feng 在 2022年1月13日 22:35，Vladimir V. Kisil<V.Kisil(a)leeds.ac.uk> 写道： Dear Feng, Thanks for spotting this typo. Indeed currently the following lines symbol x("x"); ex a = pow(x+1/x,2); cout << a << endl; exmap mp1, mp2; ex p = a.to_polynomial(mp1); cout << " = " << p << "\n with " << mp1 << endl; p = a.expand().to_polynomial(mp2); cout << " = " << p << "\n with " << mp2 << endl; produce the output with not a polynomial in the first instance: (x^(-1)+x)^2 = (x^(-1)+x)^2 with {} = 2+symbol3^2+x^2 with {symbol3==x^(-1)} If to_rational() is replaced by to_polynomial() as you suggested, then we do have (although slightly different) polynomial outputs in both cases: (x+x^(-1))^2 = (x+symbol3)^2 with {symbol3==x^(-1)} = 2+x^2+symbol4^2 with {symbol4==x^(-1)} I am attaching a one-line patch to fix this and copy the message to the developer list. Best wishes, Vladimir -- Vladimir V. Kisil http://www.maths.leeds.ac.uk/~kisilv/ Book: Geometry of Mobius Maps https://doi.org/10.1142/p835 Soft: Geometry of cycles http://moebinv.sourceforge.net/ Jupyter notebooks: https://github.com/vvkisil?tab=repositories >>>>> On Tue, 11 Jan 2022 07:10:40 +0000, Feng Feng <f.feng(a)outlook.com> said: FF> Dear Vladimir V. Kisil, FF> Thanks very much for suggestion to use expand() method on e3 FF> first. FF> And I have given a look at the source code at FF> power::to_polynomial FF> https://www.ginac.de/reference/normal_8cpp_source.html#l02688 FF> ... FF> and I wonfer if there is a typo on line 2691, the FF> basis.to_rational can be replaced with basis.to_polynomial? FF> Best regards! Feng

1 0

Re: [GiNaC-list] to_polynomial function does not return a polynomial
by Feng Feng 13 Jan '22

13 Jan '22

Dear Vladimir V. Kisil, Thanks very much for suggestion to use expand() method on e3 first. And I have given a look at the source code at power::to_polynomial https://www.ginac.de/reference/normal_8cpp_source.html#l02688 2688 ex power::to_polynomial(exmap & repl) const 2689 { 2690 if (exponent.info(info_flags::posint)) 2691 return pow(basis.to_rational(repl), exponent); 2692 else if (exponent.info(info_flags::negint)) 2693 { 2694 ex basis_pref = collect_common_factors(basis); 2695 if (is_exactly_a<mul>(basis_pref) || is_exactly_a<power>(basis_pref)) { 2696 // (A*B)^n will be automagically transformed to A^n*B^n 2697 ex t = pow(basis_pref, exponent); 2698 return t.to_polynomial(repl); 2699 } 2700 else 2701 return pow(replace_with_symbol(pow(basis, _ex_1), repl), -exponent); 2702 } 2703 else 2704 return replace_with_symbol(*this, repl); 2705 } and I wonfer if there is a typo on line 2691, the basis.to_rational can be replaced with basis.to_polynomial? Best regards! Feng

2 1

GiNaC 1.8.2 released
by Richard B. Kreckel 01 Jan '22

01 Jan '22

Happy New Year! GiNaC 1.8.2 has just been released. Here's what's new: * Fix elusive bug in comparing relational objects. * Ensure modular_form_kernel::series() includes an Order term. Happy hacking! -richy. -- Richard B. Kreckel <https://in.terlu.de/~kreckel/>

1 0