On Sun, Aug 21, 2022 at 09:43:55PM +0800, Qian Yun wrote:
> There is a regression caught by Nasser:
>
> https://12000.org/my_notes/CAS_integration_tests/reports/summer_2022/REGRESSION_REPORTS/fricas/reportsection10.htm
>
> integrate(((x^2-3*x-4)*log(1+x)+((-x^3+3*x^2+4*x)*exp(x+7)+3*x^2-10*x-8)*log(x)+(-x^2+3*x+4)*exp(x+7)+x^3-4*x^2-5*x)/(x^3-3*x^2-4*x),x)
>
> It returns integral unevaluated instead of giving correct answer.
Yes, thanks for info. This can be simplifed to
integrate(D(log(x)*log(x + 1), x), x)
> Git bisect (via fricas0-repo) gives commit
> c6039c20b834990ca4db708cf5e4e8e53e605b78, which is
> "More polylog cases".
Yes, error is in polylog_int in intpar.spad. Trouble is that
current integrator is doing recursion kernel-by-kernel while
such things are easier when working simultanously with all
relevant kernels. Note that
D(log(x)*log(x + 1), x) = (1/x)*log(x+1) + (1/(x+1))*log(x)
(1/x)*log(x+1) should integrate to dilog. More generally,
we would like to do
(a/x)*log(x+1) + (b/(x+1))*log(x)
which has elementary integral only when a = b and otherwise
needs dilog.
--
Waldek Hebisch
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/fricas-devel/20220821200756.GA6641%40fricas.math.uni.wroc.pl.
