I think the true value of this limit is zero:
(%i1) limit(x*gamma_incomplete(1/2,x^2+%i*x),x, inf);
limit: variable must be a symbol or subscripted symbol; found: sin(x)
This similar limit is OK, I think:
(%i14) limit(x*gamma_incomplete(1,x^2+%i*x),x, inf);
(%o14) 0
We can make this bug "go away" by changing the call to
$limit
ingamma-upper-trans
to a call tolimit-catch
Calling
limit-catch
is some what different than calling$limit.
And maybe the true bug is somewhere else--that is,gamma-upper-trans
never should have been called with the arguments that give rise to this bug.But I think this fix is likely harmless?
Better patch is to use the errcatch macro:
Thanks to Kris Katterjohn for reminding me about this macro.
Thanks for working on it, I agree ERRCATCH is a good solution.