I want to use the pandoc markdown plugin for jekyll on Openshift. This requires changing the PATH variable, such that the pandoc executable can be placed in the $OPENSHIFT_DATA_DIR. I've forked the jekyll cartridge and attempted to mimic the way env is setup in bin/setup there.
The ruby path element was setup like this:
dirname $(scl enable ruby193 "which ruby") > $env_dir/OPENSHIFT_RUBY_PATH_ELEMENT
So, I tried this:
dirname $(scl enable ruby193 "printenv OPENSHIFT_DATA_DIR") > $env_dir/OPENSHIFT_JEKYLL_PATH_ELEMENT
Thinking that perhaps this counted as a ruby cartridge, even though the short name is "jekyll," I also tried:
dirname $(scl enable ruby193 "printenv OPENSHIFT_DATA_DIR") > $env_dir/OPENSHIFT_RUBY_PATH_ELEMENT
I also tried:
dirname echo $OPENSHIFT_DATA_DIR > $env_dir/OPENSHIFT_JEKYLL_PATH_ELEMENT
dirname echo $OPENSHIFT_DATA_DIR > $env_dir/OPENSHIFT_RUBY_PATH_ELEMENT
Finally, in env/
I placed a file OPENSHIFT_JEKYLL_PATH_ELEMENT.erb
with the following line:
<%= ENV['OPENSHIFT_DATA_DIR'] %>
My reason for attempting this last was due to the cartridge developer's guide:
The PATH variable is set by OpenShift with the base being /etc/openshift/env/PATH. If you provide an OPENSHIFT_{Cartridge-Short-Name}_PATH_ELEMENT, OpenShift will include the value when building the PATH when your scripts are run or an application developer does an interactive log on.
SSHing to the server and checking echo $PATH
revealed no change to the environment variable. I've read the other questions regarding openshift and setting the path and I either try those solutions above or they find some way to avoid setting the path.
I'm lost…