These are chat archives for symengine/symengine

Apr 2016
Isuru Fernando
Apr 27 2016 03:59
I'm refactoring functions.h and functions.cpp to reduce all the code duplication found there. I was thinking about having 3 base classes for functions with 1, 2, variable number of arguments. Any ideas for the names of these classes? . OneArgFunction, TwoArgFunction, MultiArgFunction?
Rajith Vidanaarachchi
Apr 27 2016 04:35

I'm failing a test for lambertw for the following case:

1) SymEngine Function #lambertw calculation of lambertw(e**-1) returns -1

 Failure/Error: expect(f).to eql(-1)
       expected: -1
       got: lambertw(E**(-1))

But when I run locally, I am getting the -1 .. I'm trying to figure out what's the reason.

Also, for another test case:

2) SymEngine Function #zeta calculation of zeta zero returns minus half

Failure/Error: `expect(f).to eql(-1/2)`
       expected: `-1`
       got: `-1/2`

Here the expected is set to -1/2, but it's taken as -1 for the test

Any help is appreciated.

Isuru Fernando
Apr 27 2016 04:38
@rajithv, you can format code with github markdown here as well
Isuru Fernando
Apr 27 2016 04:46
Second case is because of integer division. -1/2 is simplified to -1
First case, when is the wrong answer output? On travis?
Rajith Vidanaarachchi
Apr 27 2016 04:47
@isuruf yes, on travis
Isuru Fernando
Apr 27 2016 04:54
this was a bug fixed in #854
You should update SymEngine locally and you should get the same failure
Rajith Vidanaarachchi
Apr 27 2016 11:46


The appveyor build is giving an error,

Build started
git clone -q C:\projects\symengine-rb
git fetch -q origin +refs/pull/23/merge:
git checkout -qf FETCH_HEAD
Running Install scripts
if [%PLATFORM%] == [x64] set DEVKIT=C:\Ruby21-x64\DevKit
if [%PLATFORM%] == [Win32] set DEVKIT=C:\Ruby21\DevKit
The system cannot find the path specified.
Isuru Fernando
Apr 27 2016 11:57
Recent appveyor updates might be the reason. Can you login to the VM and see if devkit is installed and if so where it is?
Apr 27 2016 21:53
@isuruf After consulting with @shivamvats we have realized PR #913 has become very large and have decided to split it into smaller pull requests (PR #924 and PR #925 ). If you could review these, it would be very much appreciated.