diff file for comm.lisp
Maxima's formulae for the derivatives of acsc, asec,
acsch,
and asech are incorrect in the left half plane. To
illustrate, we should have asin(1/x) - acsc(x) = 0
for all x # 0; however,
(C2) ratsimp(diff(asin(1/x) - acsc(x),x));
(D2) -SQRT(x^2-1)*(ABS(x)-x)/(x^4-x^2)
This only vanishes for x > 0. I've attached
a diff file for comm.lisp that fixes these
problems. And I attached an rtest file
that tests the derivatives of the inverse
trig functions.
Barton
diff file for comm.lisp
rtest-like code
Logged In: YES
user_id=501686
Problem shown here is OK now (5.9.3cvs):
ratsimp(diff(asin(1/x) - acsc(x),x)); => 0
Patch was applied to src/comm.lisp and committed as r1.12.
Test cases found their way into the cvs as
tests/rtest_diff_invtrig.mac and 5.9.3cvs passes all tests.
Closing this report as fixed.
Log in to post a comment.