Thanks guys.
The update() and build() calls didn't fix it. But a checkout() of PyCall
followed by a new call to build() worked it out.
I don't actually use IJulia. I installed it so that I could demonstrate it
to colleagues. Guess I'll leave it uninstalled for now :)
Dave.
On Monday, 14 September 2015 18:10:21 UTC+2, John Pearson wrote:
>
> Steven (PyCall author) posted on this a few days ago.
>
> The bug has been fixed, you just need to
>
> Pkg.checkout("PyCall")
>
> which will get the latest version of the code from master.
>
> Note, though, that even with this fixed, I'm having some problems with
> plotting in the IJulia notebook (dead kernels), though things work just
> fine from the REPL.
>
> On Monday, September 14, 2015 at 11:53:51 AM UTC-4, David Higgins wrote:
>>
>> Hi,
>>
>> I'm not sure where exactly my problem is located (in code/interfaces) so
>> any help would be appreciated.
>>
>> I did an update of all of my packages a few nights ago and ran into the
>> same problem detailed here:
>> https://groups.google.com/d/msg/julia-users/XqzNceNwa2Y/ZhahE-kQAwAJ
>> Basically the 'cp()' command was being called during the IJulia update
>> with a parameter which was not valid in v0.3 of Julia. So that particular
>> update bombed out.
>>
>> I didn't initially realise, however I have not done any other changes on
>> my system in the interim, but my ability to access Python libraries via
>> PyCall has been broken since that update. My main use is via PyPlot. This
>> is really killing me for work, I had moved entirely to Julia, so I really
>> need to sort this out.
>>
>> The best reduction of the problem which I can find is the following:
>>
>> julia> using PyCall
>>
>> julia> @pyimport math
>>
>> Could not find platform independent libraries <prefix>
>>
>> Could not find platform dependent libraries <exec_prefix>
>>
>> Consider setting $PYTHONHOME to <prefix>[:<exec_prefix>]
>>
>> ImportError: No module named site
>>
>> I don't typically use Python, so I'm not much good at it, but I can see
>> that this is the error you would normally get if the python libraries were
>> not where they're supposed to be (and is generated in Python not Julia). I
>> just have the regular Anaconda installation in order to provide matplotlib
>> access to Julia. Python is still working fine from the command-line and has
>> full access to its libraries.
>>
>> I've tried deleting the Anaconda installation and doing a re-install.
>> I've also tried deleting and reinstalling Julia. It seems to me that either
>> I'm missing an environment setting somewhere, or there's a new bug in one
>> of my libraries (PyCall??), or a file somehow got deleted during the failed
>> call to cp().
>>
>> I'm working on a Mac. Oh, and Julia completely crashes after this error!
>>
>> Anybody able to give advice or help? Please.
>>
>> Thanks,
>> David.
>>
>
