On Mon, Feb 6, 2017 at 10:44 AM, Waldek Hebisch <[email protected]> wrote: > oldk1331 wrote: >> >> > I mean text output: diffing output before and after change I >> > see several changes to results of 'mapleok' (those are nastly >> > examples that cross branch cuts). >> >> This patch fixes 'in249a:=integrate((sin(z)/(cos(z)-1))^(1/3), z= >> 0..%pi,"noPole")' >> As for other differences, I shall ask first: >> The "I" in the mapleok.input, means "%i", right? >> I think that should get fixed first. > > Original tests come from Vladimir Bondarenko and were in > Maple syntax. In Maple 'I' is imaginary unit. They were > incorrectly translated to Axiom syntax so that 'I' became > just an identifies. I think it makes sense to keep them > that way. Namely, original tests had no parameters and > most of them crosses conventional branch cuts or passed > trough singularites. Definite integrator is supposed > to reject such integrals (giving "potentialPole") and > it seem to do reasonable job. Having 'I' as parameters > made test less singular and more interesting, in particular > for some values of parameter we can get sensible results. > I was thinking about adding variants of tests with '%i' > but at least for now I decided that extra testing for > such variants is not worth extra time spent running tests. > > In a sense all those tests could be discarded as wrong > input. But _changes_ of results help discover real > bugs. > > -- > Waldek Hebisch
OK, I'll go over the differences caused by my patch. But, if "I" is changed into "%i", then the integrator goes into different path (integrate vs. complexIntegrate), and I think "I vs. %i" will have a impact on "sign". -- 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 post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/fricas-devel. For more options, visit https://groups.google.com/d/optout.
