Open zabirauf opened 9 years ago
It agree. However I tried conda and it fails on ARM processors as it wants to install a binary image that is compiled for x86_64. The best processors in terms of low watt for the same CPU speed are ARM so for me this is important as I would run my jupyter icsharp 247.
For sure if Jupyter team finds out why pip fails on ARM then this would be the cleaner way.
I found a way that works which I posted on my Google doc.
@minrk @denfromufa @zabirauf With the install script already in github and with this command line syntax the behavior of windows and mono version is identical. YEAHHHH
jupyter notebook --KernelManager.kernel_cmd="['mono', '/usr/local/bin/myjupytertest/icsharp/Kernel/bin/Release/iCSharp.Kernel.exe', '{connection_file}']" --Session.key='' --Session.keyfile='' --ip=10.1.9.5 --MappingKernelManager.default_kernel_name='icsharp'
@minrk @denfromufa @zabirauf A BIG THANK YOU TO ALL OF YOU!!!!
conda works on Raspberri Pi 2 (arm 7). It was enabled only couple of weeks ago.
@awb99 great to know everything is working with both Mono and .NET, please document in wiki extra steps for Mono
@denfromufa https://docs.google.com/document/d/1UMN7WFRQ43KrF1yWcjsMDNqD6NcRe2x5fL99v2aRdPk/edit?usp=sharing
On page 4 I have documented what i did to install Jupyter on Linux. Conda itself works, but to install Jupyter it downloads binary packets that only work for x86.
I will edit the wiki (If I am permissioned for this)
This is great, can we work with Jupyter team to install c# kernel on try.jupyter.org? I think those are Linux machines.
On Sun, Sep 13, 2015, 10:48 AM awb99 notifications@github.com wrote:
@denfromufa https://github.com/denfromufa https://docs.google.com/document/d/1UMN7WFRQ43KrF1yWcjsMDNqD6NcRe2x5fL99v2aRdPk/edit?usp=sharing
On page 4 I have documented what i did to install Jupyter on Linux. Conda itself works, but to install Jupyter it downloads binary packets that only work for x86.
I will edit the wiki (If I am permissioned for this)
— Reply to this email directly or view it on GitHub https://github.com/zabirauf/icsharp/issues/13#issuecomment-139889362.
@zabirauf @minrk @denfromufa WIKI has now Install / Run instructions for Debian Unix.
@zabirauf related to the ScriptCS.Contracts related to StyleCop reference subproject issue - this should be opened with the ScriptCS team. Basically the subProject ScriptCS does not compile because of a NuGet error to get Stylecop. If the SAME project ScriptCS is compiled from within the subfolder, then it works all ok. So I would guess this is not an issue of icSharp... unless the main project caller has to set some paths correclty before compiling the subproject. But this I doubt as it is onlyt Styleop reference that does not work.
since python is required for jupyter/ipython anyway, so pip/conda seem to be as a good cross-platform package manager option. at least some other kernels already use this mechanism:
https://pypi.python.org/pypi?:action=browse&show=all&c=586