Re: [sympy] Travis failing perhaps due to build differences

2014-03-29 Thread SAHIL SHEKHAWAT
Ah! sorry for that silly mistake.. I am correcting it now. Thanks aaron On Sunday, March 30, 2014 1:17:40 AM UTC+5:30, Aaron Meurer wrote: > > OH, I didn't see that. > > You should not use "is" to compare things. "is" does an exact object > comparison (in memory). So for instance > > a = [] > b

Re: [sympy] Travis failing perhaps due to build differences

2014-03-29 Thread Aaron Meurer
OH, I didn't see that. You should not use "is" to compare things. "is" does an exact object comparison (in memory). So for instance a = [] b = [] a is b will give False, because a and b are different lists. For immutable objects like tuples, Python reserves the right to cache the same object for

Re: [sympy] Travis failing perhaps due to build differences

2014-03-29 Thread SAHIL SHEKHAWAT
My guess is that somehow True and False are escaping from "not a.args is ()" which are checking if they have any arg and only then -BOOLEANTRUE can come. The tests were passed when i explicitly added true and false in the condition. On Sunday, March 30, 2014 1:09:28 AM UTC+5:30, Aaron Meurer wr

Re: [sympy] Travis failing perhaps due to build differences

2014-03-29 Thread Aaron Meurer
I guess the difference is that this is PyPy, but even on my computer, that test passes in PyPy. Do you have any guess what would cause it to get the -BOOLEANTRUE term (the - part in particular)? Aaron Meurer On Sat, Mar 29, 2014 at 12:51 PM, SAHIL SHEKHAWAT < sahilshekhawa...@gmail.com> wrote:

[sympy] Travis failing perhaps due to build differences

2014-03-29 Thread SAHIL SHEKHAWAT
For my pull request #7303 Travis is showing an error see https://travis-ci.org/sympy/sympy/jobs/21810070#L3374 while running the same code locally it is passing all the test and doctests. After discussing it on gitter with @zamrath i got to know that it