Conda.jl is a community contribution as a package, not part of the core
language. Likewise it is not strictly required. It does make using PyCall
significantly more convenient. Maybe you might consider making a Pip.jl
package that does something similar?

On Mon, Nov 2, 2015 at 4:18 PM, <[email protected]> wrote:

> Yes.
>
> The practical problem is out of date packages and subtle incompatibilities
> between "system" installs of Python and the private one.
>
> I now use the "Julia" Python as my 2.7 version and a system Python from
> PSF for 3.5.
>
> Conda is limited to Python 3.4 and matplotlib 1.4.3.
>
> Conda et al will always lag a bit with the risk of more incompatibilities
> creeping in.
>

Reply via email to