These are chat archives for symengine/symengine

25th
Feb 2016
Iris Lui
@irislq
Feb 25 2016 01:09
I made a mistake, I meant "use the name UnivariatePolynomial for the new polynomial with Expression coefficients" not UnivariateSeries
Isuru Fernando
@isuruf
Feb 25 2016 01:47
Sure and the previous one to UnivariateIntPolynomial
Charles Chen
@chenchfort
Feb 25 2016 02:26
When do we need SYMENGINE_NOEXCEPT?
Isuru Fernando
@isuruf
Feb 25 2016 02:42
When you are adding noexcept to functions, you need to use this as some compilers may not have that feature built in
Abhinav Agarwal
@abhinavagarwal07
Feb 25 2016 02:49
How can we implement infinity ?
Isuru Fernando
@isuruf
Feb 25 2016 03:03
@chenchfort, thanks for reminding me of noexcept. I opened #825
@abhinavagarwal07, what do you mean?
Abhinav Agarwal
@abhinavagarwal07
Feb 25 2016 07:26
I was thinking that we should implement infinity and complex infinity.
Ralf Stephan
@rwst
Feb 25 2016 14:07
@certik Yes. I just used #651 as is, but the numbers don't differ much anyway. It looks like Pynac can be improved there by using hardware ints < 2^64instead of GMP for everything, but not by much probably.
Abhinav Agarwal
@abhinavagarwal07
Feb 25 2016 14:20
I'm implementing logGamma, I was wondering if I should use Gamma::is_canonical() and directly return log( gamma(...)) if it's false or evaluate it at only arg = 1,2,3 as in sympy.
Isuru Fernando
@isuruf
Feb 25 2016 14:41
I think the point of loggamma is that gamma can get quite large and for computations it's better to have a log version of it. So, evaluating as log(gamma()) and returning the log is not desired.
Abhinav Agarwal
@abhinavagarwal07
Feb 25 2016 14:52
+1
another importance of loggamma is : digamma is derivative of loggamma
Iris Lui
@irislq
Feb 25 2016 20:46
Does the Mul class support multiplication with non-integer multiplication? Like a*x?
Iris Lui
@irislq
Feb 25 2016 22:39
Er, disregard my last question I figured it out