chruby does not automatically see new Ruby installs

[FooBarWidget]

I'm considering using chruby in a prebuilt machine image that is going to be publicly available for Ruby developers. However I ran into an unpleasant usability surprise.

I installed chruby, then I ran:

ruby-install ruby 1.8.7

Then I ran:

chruby

The output is empty. This was confusing. Why did it not see my new Ruby? I logged out and logged in again, and this time it saw my new Ruby. I shouldn't have to do this.

Reading the source code, it seems that chruby only queries the list of Rubies during evaluation of source.sh. You should query it more often. Or, at the very least, you should provide a command for reloading the list of Rubies, and updating ruby-install to tell the user about the need to relogin/running the reload command.

[havenwood]

@FooBarWidget The 0.4.0 branch includes chruby-install, which integrates with ruby-install and reloads $RUBIES after install so they're immediately available: https://github.com/postmodern/chruby/blob/0.4.0/share/chruby/install.sh

[postmodern]

Also you can simply restart the shell or open a new terminal tab.

[FooBarWidget]

I know how to solve the problem, but I am concerned about the usability for
people who have few unix skills. It was not obvious that the list is cached
in RUBIES (I thought it was only used to override autodetection of Rubies).

I happen to have the intuition that the problem is probably a caching
issue, so I tried relogging in, but there are lots of folks who don't know
that and will just give up in frustration.

At the very least, you should document the need to relogin after installing
a Ruby.

Sent from my Android phone.
On Oct 13, 2013 1:26 AM, "Postmodern" notifications@github.com wrote:

Also you can simply restart the shell or open a new terminal tab.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/212#issuecomment-26208204
.

[judofyr]

I use this helper:

function chruby_reload() {
  RUBIES=(~/.rubies/*)
}

chruby_reload

[postmodern]

The other option we were discussing was to hardcode /opt/rubies/* and $HOME/.rubies/* into the chruby function that lists/selects rubies. However, this would change the meaning of RUBIES, from being the authoritative list of rubies to the list of user specified rubies.

[judofyr]

RUBIES could also be a glob:

RUBIES=('~/.rubies/*')

[postmodern]

Ah yes, someone else recommended using double for loops. I think I will have to play around with bash dir-glob expansions.

[postmodern]

This appears to work, even with ruby dirs that contain spaces:

RUBIES=("/opt/rubies/*" "$HOME/.rubies/*")
for ruby in ${RUBIES[@]}; do
    echo "Testing: $ruby"
    if [[ -d "$ruby" ]] && [[ "$ruby" == *$1* ]]; then
        echo "Found: $ruby"
        exit
    fi
done