Me too. I mean, I found things worked much more smoothly after following this advice to use Conda. Also, Windows and OSX are supported immediately. Also, for my application, the Ubuntu package versions are very old, and out of date. This problem is solved as well.
On Friday, October 7, 2016 at 10:24:36 PM UTC+2, Steven G. Johnson wrote: > > > > On Friday, October 7, 2016 at 9:10:44 AM UTC-4, David van Leeuwen wrote: >> >> Hello, >> >> For a tiny package <https://github.com/davidavdav/CHull.jl> that depends >> on PyCall and python's scipy.spatial I am trying to engineer a >> `.travis.yml`, >> > > You can always do ENV["PYTHON"]="" to force PyCall to install its own > Python distro (via Conda), and do pyimport_conda("scipy.spatial", "scipy") > to make Conda install scipy for you as needed. > > Steven > > PS. As explained in the PyCall README, I don't recommend using @pyimport > in Modules. Instead, do > const spatial = PyNULL() > and then, in your __init__ function, do copy!(spatial, > pyimport_conda("scipy.spatial", "scipy")) .... this way, you can put > __precompile__(true) at the top of your module and safely precompile it. > > PPS. I see from your source code that you had some confusion about > Base.show vs. Base.display. Never override Base.display for this sort of > thing. See the new manual section for more info: > > http://docs.julialang.org/en/latest/manual/types/#custom-pretty-printing >
