another maxima integration bug

[Frédéric Chapoton]

var('s,x,y')
q = sqrt(2)
g = (x - y) * sin((x - y) / s)
g1 = integral(g({s:q}), y)
g2 = integral(g({s:q}), y, algorithm='sympy')
g3 = integral(g, y).subs({s:q})

Then g2 and g3 have the correct derivative, but not g1.

possibly related to #30378 and #31530

[Erik Bray]

Ticket retargeted after milestone closed

[Matthias Köppe]

Moving tickets to milestone sage-9.2 based on a review of last modification date, branch status, and severity.

[Matthias Köppe]

Moving to 9.4, as 9.3 has been released.

[Frédéric Chapoton]

EDIT:

in sage 9.4.b0, numerically :

the derivative of g1 is wrong ; those of g2 and g3 are ok

[Frédéric Chapoton]

directly in maxima:

maxima: a: sin((x-y)/sqrt(2))*(x-y);                                            
sin((x-y)/sqrt(2))*(x-y)

maxima: g: integrate(a,y);                                                      
((2^(3/2)+2)*sin(y/sqrt(2)-x/sqrt(2))+(((-2^(3/2))-2)*(y/sqrt(2)-x/sqrt(2))+((-sqrt(2))-2)*x)*cos(y/sqrt(2)-x/sqrt(2)))/2+sqrt(2)*x*cos(y/sqrt(2)-x/sqrt(2))

[Frédéric Chapoton]

This should be reported upstream to maxima. Karl-Dieter, would you please do that ?

For info, the correct answer as given by sympy is

sqrt(2)*(x - y)*cos(-1/2*sqrt(2)*(x - y)) + 2*sin(-1/2*sqrt(2)*(x - y))

[Karl-Dieter Crisman]

​https://sourceforge.net/p/maxima/bugs/3795/