Try setting the environment variable PYTHONHASHSEED as shown at the
top of the tests. If that reproduces the errors exactly, then some
part of your code is dependent on the values of hashes. This happens
if you iterate through a set or dictionary, or if you iterate through
the args of an Add or Mul, which are sorted by hash, and the result
depends on the order or iteration.

By the way, why are you using doctest -n?  The doctests should all
pass without the -n option.

Aaron Meurer

On Sat, Sep 7, 2013 at 2:52 PM, Sachin Joglekar <[email protected]> wrote:
> I am currently working on a PR to modify the mechanics core - #2412. The
> Travis build seems to be failing for a couple of doctests in the
> doc/src/modules/physics/mechanics/bicycle_example.rst file. Have a look here
> - https://travis-ci.org/sympy/sympy/jobs/11069792
> The problem is, neither Gilbert nor I are able to figure out why. The
> weirder stuff is, doing
> ./bin/doctest -n doc/src/modules/physics/mechanics/bicycle_example.rst
> on my machine gives no error once, and 5 errors the next time I run it
> (Without altering any code in the meantime). Same goes for the command
> without the '-n' parameter.
> Can anybody help us out with this? The PR is stuck at this issue, and we
> need to resolve it soon. (I have attached a file with the errors shown)
>
> --
> You received this message because you are subscribed to the Google Groups
> "sympy" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To post to this group, send email to [email protected].
> Visit this group at http://groups.google.com/group/sympy.
> For more options, visit https://groups.google.com/groups/opt_out.

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sympy.
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to