Adding the ARG feature to build with any user for juju-1

[mbruzek]

• Removed the run.sh script based on the work that was done in master.
• Updated the README.md with the work done in master.

Just trying to keep this juju-1 branch fresh

[kwmonroe]

any reason you went with ADD over COPY here? docker docs say COPY is preferred if you're not relying on ADD auto extraction capabilities:

https://docs.docker.com/engine/userguide/eng-image/dockerfile_best-practices/#/add-or-copy

[kwmonroe]

Since #21 went with COPY, please s/ADD/COPY here too.

[kwmonroe]

why remove these exports from the user's .bashrc? they seem handy to have set for anyone entering the jujubox.

[kwmonroe]

oh, i see in master that these aren't set in the .bashrc either.. do ENV doohickeys persist when the image is run? if so, i see why these .bashrc additions aren't needed.

[chuckbutler]

Yeah, ENV exports persist unless you run sudo, in which point anything in the env was going to be stripped anyway. 🧀

[mbruzek]

Requested change committed, please re-review.

[chuckbutler]

I had one minor comment. the rest of this looks good to me.

I validated the build routine as follows:

docker build --build-arg JUJU_USER=foobar -t jujubox-1 .

everything appears to be in order here. Just a minor nit about the juju-1 bin naming and you have no control over that. Maybe we could symlink it for the user so its transparent? is that a bad idea?

[chuckbutler]

It might still be prudent to include information about how to run juju1

the binary has changed to juju-1, i was surprised when i fired up the assembled container, typed in juju and it responded with command not found :|

lameee

[mbruzek]

Thanks for review. I thought about using a symlink, but I don't want to step on the packaging if they ever were to fix that.