The `r` channel has been considered part of `defaults` since `conda` version `4.3.0`. So it should already be taken into consideration by `conda` installs without having to explicitly add the channel. Further the `conda-forge` channel has incorporated an ever growing, healthy stack of R packages. As such it appears all of the R packages used in this stack now come from `conda-forge` and not `defaults`. Given all of this, it seems safe to drop the `r` channel from explicit addition to the channel list.
- any files the user should be able to write should have group `user-permissions` with `g+rwX`
- remove `chown` from start.sh because it is no longer needed
- add `fix-permissions` script for setting the user-writable permissions on a path
- user-permissions group as GID 10000 (is there a reason for it to have a different value?)
- containers can set group with `--group-add user-writable` if they want to run with a different uid/gid
(without -u root -e NB_UID -e NB_GID, which make this unnecessary)
- add a few missing `--system` flags to conda config
- install default notebook config to /etc/jupyter instead of ~/.jupyter
- add a few missing `conda clean`s
base-notebook defines environment variables for the Conda install path and the
notebook user. However, in some instances, these locations were still hardcoded.
Let’s use the variables instead.
* create `.juliarc.jl` start up script to point Julia to conda libraries, specifically the hdf5 libraries installed through the h5py package in the scipy stack
* add the HDF5 Julia package
Update README.md to document HDF5 addition
Add -F option to force config load
Add the -F flag to julia expression execution to load .juliarc.jl before execution.
Bite the bullet and preinstall it so that plotting libs that default
to using desktop rendering just work (matplotlib, ggplot, ...)
out of the box without having to get configuration right beforehand
(e.g., %matplotlib inline ahead of matplotlib import)
Only adds ~100k to the image size
(c) Copyright IBM Corp. 2015
