FILE test

ruby-talk
1

I noticed that a lot of scripts apply the following coding pattern:

if __FILE__ == $0
....
end

I know what is does but what kind of problems does it solve?

thx,

used-to-be-a-Smalltalker

2

It allows the file to be both a library (when required the if statement will not run) and executable.

Hope that helps.

James Edward Gray II

···

On Feb 21, 2006, at 8:58 AM, raving_ruby_rider wrote:

I noticed that a lot of scripts apply the following coding pattern:

if __FILE__ == $0
....
end

I know what is does but what kind of problems does it solve?

3

You will see this sort of thing in other scripting languages. __FILE__ contains the name of the file source file and $0 is the name of the currently executing script. So a file called hello.rb

class Hello
    def initialize(name)
        @name = name
    end

    def greet
        puts "Hello #{@name}"
    end
end

if __FILE__ == $0 then
    h = Hello.new('World')
    h.greet
end

you can then run this file with ruby hello.rb which will run the code at the bottom. However with tom.rb

require 'hello'

h = Hello.new('tom')
h.greet

when you run tom.rb the __FILE__ == $0 part of hello does not run as the file hello.rb but the currently executing script is tom.rb. This allows you to have a ruby file hold the class for inclusion by other scripts and also be a utility script in it's own right.

4

Peter Hickman wrote:

You will see this sort of thing in other scripting languages. __FILE__ contains the name of the file source file and $0 is the name of the currently executing script. So a file called hello.rb

...

when you run tom.rb the __FILE__ == $0 part of hello does not run as the file hello.rb but the currently executing script is tom.rb. This allows you to have a ruby file hold the class for inclusion by other scripts and also be a utility script in it's own right.

Not to mention the fact that it allows you to easily write unit tests for most of the code in a script.
V.-

···

--
http://www.braveworld.net/riva

____________________________________________________________________
http://www.freemail.gr - äùñåÜí õðçñåóßá çëåêôñïíéêïý ôá÷õäñïìåßïõ.
http://www.freemail.gr - free email service for the Greek-speaking.

5

Dňa Utorok 21 Február 2006 20:25 Damphyr napísal:

Peter Hickman wrote:
> You will see this sort of thing in other scripting languages. __FILE__
> contains the name of the file source file and $0 is the name of the
> currently executing script. So a file called hello.rb

...

> when you run tom.rb the __FILE__ == $0 part of hello does not run as the
> file hello.rb but the currently executing script is tom.rb. This allows
> you to have a ruby file hold the class for inclusion by other scripts
> and also be a utility script in it's own right.

Not to mention the fact that it allows you to easily write unit tests
for most of the code in a script.

Which I personally can't bear the sight of. Library is script is test? Nuh-uh.
Just my two cents.

David Vallner

6

So what is the 'ruby way' to store your unit tests? A separate require'd file?

···

On 2/21/06, David Vallner <david@vallner.net> wrote:

Which I personally can't bear the sight of. Library is script is test? Nuh-uh.

7

David Vallner wrote:

Dňa Utorok 21 Február 2006 20:25 Damphyr napísal:

Peter Hickman wrote:

You will see this sort of thing in other scripting languages. __FILE__
contains the name of the file source file and $0 is the name of the
currently executing script. So a file called hello.rb

...

when you run tom.rb the __FILE__ == $0 part of hello does not run as the
file hello.rb but the currently executing script is tom.rb. This allows
you to have a ruby file hold the class for inclusion by other scripts
and also be a utility script in it's own right.

Not to mention the fact that it allows you to easily write unit tests
for most of the code in a script.

Which I personally can't bear the sight of. Library is script is test? Nuh-uh. Just my two cents.

Well I actually put the unit tests in a different file.
The if $0==__FILE__ check allows me to require the script in the unit test file.
Following mostly the DRY principle and having a knack for organizing code allows you to group most of the functionality in objects (at which point you put them in a 'library' file and forget about it) or methods to be used by the 'top-level' script. Requiring the 'script' file allows me to unit test the methods without contriving manual tests.
I found it most valuable when I do parameter parsing and parameter/configuration validation in my command line scripts.
Cheers,
V.-

···

--
http://www.braveworld.net/riva

____________________________________________________________________
http://www.freemail.gr - äùñåÜí õðçñåóßá çëåêôñïíéêïý ôá÷õäñïìåßïõ.
http://www.freemail.gr - free email service for the Greek-speaking.

8

Adam Shelly wrote:

So what is the 'ruby way' to store your unit tests? A separate
require'd file?

Well, going by the Pickaxe book you end up with something like this:

./classfilename
->/doc
->/lib
->/test

in ./classfilename/lib you create your ruby source file classfilename.rb
in ./classfilename/test you create your ruby test/unit file
classfilename_tc.rb

In classfilename_tc.rb you put the following lines at the start:

···

#----------------------------------------------------------------
# The following prefixes ../lib to the active ruby load path
$:.unshift File.join(File.dirname(__FILE__), "..", "lib")

require 'test/unit'
require 'classfilename'

class Test_ClassFileName < Test::Unit::TestCase
...
#----------------------------------------------------------------

and then write your test cases as methods. When you run your test suite
you can invoke it from any palce on the system as the #unshift prefixes
the load path with the relative location of the classfilename.rb with
respect to the test case file.

Thus, assuming that for the example given above that ./ = ~/ruby then:

#ruby -w ~/ruby/classfilename/test/classfilename_tc.rb

will work whatever pwd you are in.

I love test/unit...

Regards,
Jim

--
Posted via http://www.ruby-forum.com/\.