After commit: "Consistently integrate trigonometric functions
via complex exponentials" we no longer use special integration
routines working with tangent. Advantage is that new
routines for integration in terms of special functions
now apply also in trigonometric case. In particular
new code should handle more cases of functions integrable
in terms of Si, Ci, fresnelC, fresnelS than the old
one. Also, integrator is simpler: the special routines
for integrationg rational functions of tangent are now
removed (in fact pattern matching routines working in terms
of tangent can no longer apply, so they are removed).
So integrator simultaneusly got stronger and smaller.
As side effect of the change several integral now take
different form. AFAICS the change of form has mostly
neutral effect. More precisely, location of poles
of arguments of logarithms changed. For some functions
it means that now we have pole in the path of integration,
so some definite integrals that used to work now fail.
OTOH some definite integrals which failed due to singularity
in the indefinite integral now work. The change is semi
random and looks neutral on average. However, if you
observe change that you think is especially bad, please
report. In general such problems are quite hard (in a
sense unsolvable), but sometimes in specific cases
we can get improvement.
BTW: This change required preparations. Bronstein
introduced tangent case to get nicer integrals of
trigonometric functions. Currently FriCAS obtains
nice integrals in different way (as I wrote results
differ, but without earlier changes would be worse
than obtained via tangents).
--
Waldek Hebisch
[email protected]
--
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 http://groups.google.com/group/fricas-devel.
For more options, visit https://groups.google.com/d/optout.
