25 Mar
2010
25 Mar
'10
12:17 a.m.
I stumbled upon this using Sage, but developers there have traced it to a problem in GiNaC by running the following in ginsh:
atan2(-Pi,0);
The response is: power::eval(): division by zero atan2 appears to work for other values, including (positive) (Pi,0), but (-Pi,0) fails.