Question

I’m looking at the code for Rails 3. In rails/cli.rb, which is loaded during the startup process, it makes a call to Rails::ScriptRailsLoader.exec_script_rails! which in turn replaces the current ruby process by reloading the application a second time with a call to exec:

exec RUBY, SCRIPT_RAILS, *ARGV if in_rails_application?

I’m not that familiar with the Rails codebase, so my question is what is the rationale for this? Why doesn’t Rails just continue starting up in the same original process? I’m sure there is a good reason, I just don’t know it.

Answer 1

I think the idea is that ‘rails’ as a command line executable is used in two ways:

• to initialize apps (outside of a Rails application)
• to do things for an app (ex launch a server — inside a Rails application dir)

For example, outside a rails app:

$ rails -h
Usage:
  rails new APP_PATH [options]
...

And inside a rails app:

$ rails -h
Usage: rails COMMAND [ARGS]
...

It looks like exec_script_rails! is used to pick which interface the rails command presents to the user. This comment and a look at ScriptRailsLoader suggest that the method only calls the exec if the user is inside a rails app – specifically a directory where ‘script/rails’ exists. In that case, the exec replaces the original process with ‘script/rails’ to make these equivalent:

$ rails ...
$ ruby script/rails ...

Otherwise the script continues on in the same process. So the rationale is that it lets the ‘rails’ command behave in a context-specific way that preserves whatever ‘script/rails’ does.