04:16 <odigity> If I'm creating a gem with an executable that will be installed, should I require 'bundler/setup' in my executable to ensure the versions specified in the gemspec are used?

12:17 <jhass> odigity: no

13:40 <odigity> jhass, after looking at examples, that's what I concluded. What I did instead: 1) in my executable, I just require one file from my gem (a Thor app) and run start 2) in my gem repo, I run "bundle --binstubs" to generate a binstub for my exe in my gem's bin/ dir after installing the gem to the system, so now it runs with the not-checked-in Gemfile in my gem repo

13:40 <odigity> Does that all seem correct?

13:41 <jhass> idk, I don't follow the silly exe/ convention

13:42 <jhass> just require your gem files in bin/ and use bundle exec bin/foo for dev

13:42 <odigity> jhass, I generated a local binstub in my gem repo to avoid having to type "bundle exec" every time, because that's annoying.

13:42 <odigity> and I don't like to depend on shell aliases, because then things stop working when you move to another system

13:42 <jhass> idk, my up arrow works well still

13:43 <odigity> When a user installs my gem and rubygems installs a wrapper in /usr/local/bin/myapp, what ensures that the command gets run with the gem versions specified in the gemspec?

13:46 <jhass> nothing

13:47 <odigity> Isn't that a problem?

13:47 <jhass> however iirc rubygems will error out when your gem gets activated and other incompatible versions are already activated

13:55 <odigity> jhass, well, I suppose that's something. still, it'd be nice if there was just one good tutorial that covered all these issues from start to finish