thr3ads.net - Rails - This modules thing somehow really does not work for me, please help. [Dec 2011]

If this information is useful, please help other people find it:
Share via:

2011-Dec-10 22:51 UTC

This modules thing somehow really does not work for me, please help.

Hey,

thank you in advance for reading this post. I struggle with
modularization / packaging / name spacing (however you want to call
it) via ruby modules. I have the feeling that somehow rails is
responsible for the problems I am facing. I hope, you can help me.

Here''s an example to demonstrate more exactly what I mean:

module Storage
  module Administration

    def self.do_something()
      # .. omitted
    end

  end
end

This does sometimes work under the rails console (development mode),
most of the time I get:

ruby-1.9.2-p290 :004 > Storage::Administration.do_something()
NoMethodError: undefined method `do_something'' for
Storage::Administration:Module

This one here works more often (but not always):

module Storage::Administration

    def self.do_something()
      # .. omitted
    end

end

What''s about the constant resolution operator (::)? As far as I know
it is only allowed to reference constants (like module or class
names), but not directly in definitions (module X::Y::Z).

As I know it from pure ruby the first example should do it. You just
nest the modules as you need and simply reopen a module if you want to
extend it.

So, the contents of my ruby files used for this example:

"storage.rb" file (in a directory called "storage"):

module Storage
end

And my "administration.rb" (in a subdirectory of "storage"
called
"administration"):

module Storage
  module Administration

  # .. omitted

  end
end

OR (as I also tried)

module Storage::Administration

    # .. omitted

end

What am I doing wrong? Why can''t rails (console, runner, etc.) access
my static method _all the time_?

Thank you very much for any suggestions!
ms

-- 
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Talk" group.
To post to this group, send email to
rubyonrails-talk-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-talk+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-talk?hl=en.

Peter Vandenabeele

2011-Dec-10 23:23 UTC

head link

Re: This modules thing somehow really does not work for me, please help.

On Sat, Dec 10, 2011 at 11:51 PM, ms <ms-cGBD8117FJM@public.gmane.org>
wrote:
> Hey,
>
> thank you in advance for reading this post. I struggle with
> modularization / packaging / name spacing (however you want to call
> it) via ruby modules. I have the feeling that somehow rails is
> responsible for the problems I am facing. I hope, you can help me.
>
> Here''s an example to demonstrate more exactly what I mean:
>
> module Storage
>  module Administration
>
>    def self.do_something()
>      # .. omitted
>    end
>
>  end
> end
>
> This does sometimes work under the rails console (development mode),
> most of the time I get:
>
> ruby-1.9.2-p290 :004 > Storage::Administration.do_something()
> NoMethodError: undefined method `do_something'' for
> Storage::Administration:Module
>
> This one here works more often (but not always):
>
> module Storage::Administration
>
>    def self.do_something()
>      # .. omitted
>    end
>
> end
>
> What''s about the constant resolution operator (::)? As far as I
know
> it is only allowed to reference constants (like module or class
> names), but not directly in definitions (module X::Y::Z).
>
> As I know it from pure ruby the first example should do it. You just
> nest the modules as you need and simply reopen a module if you want to
> extend it.
>
> So, the contents of my ruby files used for this example:
>
> "storage.rb" file (in a directory called "storage"):
>
> module Storage
> end
>
> And my "administration.rb" (in a subdirectory of
"storage" called
> "administration"):
>
> module Storage
>  module Administration
>
>  # .. omitted
>
>  end
> end
>
> OR (as I also tried)
>
> module Storage::Administration
>
>    # .. omitted
>
> end
>
> What am I doing wrong? Why can''t rails (console, runner, etc.)
access
> my static method _all the time_?
>
> Thank you very much for any suggestions!
>
Under which directory exactly are you saving these files ?
Which version of rails?

If you use files under /lib, this may help (since recent versions of Rails,
files under lib are not by default auto-loaded).

    # Custom directories with classes and modules you want to be
autoloadable.
    config.autoload_paths += %W(#{config.root}/lib)
    config.autoload_paths += Dir["#{config.root}/lib/**/"]

to understand the issue, I suggest you try step by step> Storage
> Storage::Administration
> Storage::Administration.functionto see how far it works and where exactly it breaks.

This code below works reliably in Rails 3.1.x

$ cat lib/storage/administration.rb
module Storage
  module Administration
    def self.foo
     "foo"
    end
  end
end

$ rails c
Loading development environment (Rails 3.1.3)
001:0> Storage
=> Storage
002:0> Storage::Administration
=> Storage::Administration
003:0> Storage::Administration.foo
=> "foo"
004:0> quit

$ rgrep -i autoload config/application.rb
    # Custom directories with classes and modules you want to be
autoloadable.
    config.autoload_paths += %W(#{config.root}/lib)
    config.autoload_paths += Dir["#{config.root}/lib/**/"]

I would expect this code to be reliable on your side as well.

HTH,

Peter

-- 
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Talk" group.
To post to this group, send email to
rubyonrails-talk-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-talk+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-talk?hl=en.

Frederick Cheung

2011-Dec-11 14:04 UTC

head link

Re: This modules thing somehow really does not work for me, please help.

On 10 Dec 2011, at 22:51, ms <ms-cGBD8117FJM@public.gmane.org> wrote:
> Hey,
> 
> thank you in advance for reading this post. I struggle with
> modularization / packaging / name spacing (however you want to call
> it) via ruby modules. I have the feeling that somehow rails is
> responsible for the problems I am facing. I hope, you can help me.
> 
It sounds like the issue is with rails'' autoloading system
> 
> What''s about the constant resolution operator (::)? As far as I
know
> it is only allowed to reference constants (like module or class
> names), but not directly in definitions (module X::Y::Z).
> module X::Y::Z is fine, but requires that X::Y already exists, whereas

module X
  module Y
    module Z
    end
  end
end

Doesn''t. This also means that in the first case rails will look (via
const_missing) for X::Y (in x/y.rb) whereas in the second case it
won''t. If in x/y.rb you added some methods to Y and x/y.rb had never
been loaded but z.rb had been then X::Y would still exist because of z.rb, but
those methods from y.rb wouldn''t.
> As I know it from pure ruby the first example should do it. You just
> nest the modules as you need and simply reopen a module if you want to
> extend it.
> 
> So, the contents of my ruby files used for this example:
> 
> "storage.rb" file (in a directory called "storage"):
> 
> module Storage
> end
Actually, rails would expect to find Storage in storage.rb at the top level, ie
not in the storage folder> 
> And my "administration.rb" (in a subdirectory of
"storage" called
> "administration"):
> 
> module Storage
>  module Administration
> 
>  # .. omitted
> 
>  end
> end
> 
> OR (as I also tried)
> 
> module Storage::Administration
> 
>    # .. omitted
> 
> end
Ditto, rails expects to find Storage::Administration in
storage/administration.rb
> What am I doing wrong? Why can''t rails (console, runner, etc.)
access
> my static method _all the time_?
> Probably a combination of not putting things quite in the right place and using
the
module X
  module Y
  end
end
style, which doesn''t force rails to load x.rb, leading to different
behaviour depending on whether you reference X before X::Y or the other way
around.

Fred> Thank you very much for any suggestions!
> ms
> 
> -- 
> You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Talk" group.
> To post to this group, send email to
rubyonrails-talk-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
> To unsubscribe from this group, send email to
rubyonrails-talk+unsubscribe@googlegroups.com.
> For more options, visit this group at
http://groups.google.com/group/rubyonrails-talk?hl=en.
> 
-- 
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Talk" group.
To post to this group, send email to
rubyonrails-talk-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-talk+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/rubyonrails-talk?hl=en.

Peter Vandenabeele

2011-Dec-11 18:10 UTC

head link

Re: This modules thing somehow really does not work for me, please help.

On Sun, Dec 11, 2011 at 3:04 PM, Frederick Cheung <
frederick.cheung-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
>
> On 10 Dec 2011, at 22:51, ms <ms-cGBD8117FJM@public.gmane.org> wrote:
>
> > Hey,
> >
> > thank you in advance for reading this post. I struggle with
> > modularization / packaging / name spacing (however you want to call
> > it) via ruby modules. I have the feeling that somehow rails is
> > responsible for the problems I am facing. I hope, you can help me.
> >
>
> It sounds like the issue is with rails'' autoloading system
>
> >
> > What''s about the constant resolution operator (::)? As far as
I know
> > it is only allowed to reference constants (like module or class
> > names), but not directly in definitions (module X::Y::Z).
> >
> module X::Y::Z is fine, but requires that X::Y already exists, whereas
>
> module X
>  module Y
>    module Z
>    end
>  end
> end
>
> Doesn''t. This also means that in the first case rails will look
(via
> const_missing) for X::Y (in x/y.rb) whereas in the second case it
won''t. If
> in x/y.rb you added some methods to Y and x/y.rb had never been loaded but
> z.rb had been then X::Y would still exist because of z.rb, but those
> methods from y.rb wouldn''t.
>
Thanks for the very interesting explanation.

If I understand your reasoning correctly, the logic would be that
in the cases with separate module definitions:

1) module X::Y::Z is loaded from the file ../lib/x/y/z.rb
2) this implicitely also defines the constant X::Y
    (based solely on information in file ../lib/x/y/z.rb
     and not trying to read ../lib/x/y.rb)
3) when X::Y is used later on, it is never really loaded from the
    file ../lib/x/y.rb , so what is defined there can be
    mysteriously missing (depending on the order modules where loaded).

However, I could not produce this behavior as I understood it.

As a test, I made 2 implementations of the module in ./lib/x/y/z.rb

In both cases I have in ../lib/x/y.rb

.../lib/x$ cat y.rb
puts "loading module X / module Y in file ../lib/x/y.rb"

module X
  module Y
  end
end


Test 1: separate module definitions
===========================
.../lib/x/y$ cat z.rb
module X
  module Y
    module Z
      def self.foo
        "Z.foo"
      end
    end
  end
end

Results:

$ rails c
Loading development environment (Rails 3.1.3)
001:0> ActiveSupport::Dependencies.logger = Rails.logger  #=> <...>
002:0> ActiveSupport::Dependencies.log_activity=true  #=> true
003:0> X::Y::Z
loading module X / module Y in file ../lib/x/y.rb
=> X::Y::Z

in logfile:
Dependencies: called load_missing_constant(Object, :X)
Dependencies: called load_missing_constant(X, :Y)
Dependencies: called require_or_load(".../lib/x/y.rb", nil)
Dependencies: loading .../lib/x/y
Dependencies: called load_file(".../lib/x/y.rb", ["X::Y",
"Y"])
Dependencies: called new_constants_in("X", :Object)
Dependencies: New constants: X::Y
Dependencies: loading .../lib/x/y.rb defined X::Y
Dependencies: called load_missing_constant(X::Y, :Z)
Dependencies: called require_or_load(".../lib/x/y/z.rb", nil)
Dependencies: loading .../lib/x/y/z
Dependencies: called load_file(".../lib/x/y/z.rb",
["X::Y::Z", "Y::Z", "Z"])
Dependencies: called new_constants_in("X::Y", "Y", :Object)
Dependencies: New constants: X::Y::Z
Dependencies: loading .../lib/x/y/z.rb defined X::Y::Z


Test 2: (in 1 direct line module X::Y::Z)
============================
.../lib/x/y$ cat z.rb
module X::Y::Z
  def self.foo
    "Z.foo"
  end
end

Results:

$ rails c
Loading development environment (Rails 3.1.3)
001:0> ActiveSupport::Dependencies.logger = Rails.logger  #=> <...>
002:0> ActiveSupport::Dependencies.log_activity=true  #=> true
003:0> X::Y::Z
loading module X / module Y in file ../lib/x/y.rb
=> X::Y::Z

In logifle:

Dependencies: called load_missing_constant(Object, :X)
Dependencies: called load_missing_constant(X, :Y)
Dependencies: called require_or_load(".../lib/x/y.rb", nil)
Dependencies: loading .../lib/x/y
Dependencies: called load_file(".../lib/x/y.rb", ["X::Y",
"Y"])
Dependencies: called new_constants_in("X", :Object)
Dependencies: New constants: X::Y
Dependencies: loading .../lib/x/y.rb defined X::Y
Dependencies: called load_missing_constant(X::Y, :Z)
Dependencies: called require_or_load(".../lib/x/y/z.rb", nil)
Dependencies: loading .../lib/x/y/z
Dependencies: called load_file(".../lib/x/y/z.rb",
["X::Y::Z", "Y::Z", "Z"])
Dependencies: called new_constants_in("X::Y", "Y", :Object)
Dependencies: New constants: X::Y::Z
Dependencies: loading .../lib/x/y/z.rb defined X::Y::Z



So, my experiments seem to suggest that in both cases
that I tried, there is an explicit order where all intermediate
modules are loaded in turn (with load_missing_constants),
even if only the deepest module would be strictly required to
evaluate X::Y::Z

Could you show an example of where the problem you described
occurs. Or maybe this behavior has changed recently?

Thanks,

Peter

-- 
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Talk" group.
To post to this group, send email to
rubyonrails-talk-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-talk+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-talk?hl=en.

Frederick Cheung

2011-Dec-11 19:30 UTC

head link

Re: This modules thing somehow really does not work for me, please help.

On Dec 11, 6:10 pm, Peter Vandenabeele
<pe...-jNuWw7i2w7syMbTcgqFhxg@public.gmane.org> wrote:
> So, my experiments seem to suggest that in both cases
> that I tried, there is an explicit order where all intermediate
> modules are loaded in turn (with load_missing_constants),
> even if only the deepest module would be strictly required to
> evaluate X::Y::Z
>
> Could you show an example of where the problem you described
> occurs. Or maybe this behavior has changed recently?
Yeah I mixed things up a bit there as well as slightly mis-
remembering. if you do things as rails expects, and rely entirely on
the const_missing hooks then either way works. but if you deviate from
that then things are different.

For example if z.rb is

module X
  module Y
    module Z
    end
  end
end

and you do require_dependency ''x/y/z'' then x/y.rb
won''t be loaded

but if you do

module X::Y::Z
end

then it will.

I guess the point I was trying to make was that if ruby ends up
reading z.rb then using module X::Y::Z guarantees that x.rb and x/y.rb
will be read, but the other form doesn''t, although the usual scenarios
would mean that x.rb and x/y.rb would have been read earlier on
anyway.

Fred


-- 
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Talk" group.
To post to this group, send email to
rubyonrails-talk-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-talk+unsubscribe@googlegroups.com.
For more options, visit this group at
http://groups.google.com/group/rubyonrails-talk?hl=en.

Peter Vandenabeele

2011-Dec-11 21:36 UTC

head link

Re: Re: This modules thing somehow really does not work for me, please help.

On Sun, Dec 11, 2011 at 8:30 PM, Frederick Cheung <
frederick.cheung-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> wrote:
> On Dec 11, 6:10 pm, Peter Vandenabeele
<pe...-jNuWw7i2w7syMbTcgqFhxg@public.gmane.org> wrote:
>
> > So, my experiments seem to suggest that in both cases
> > that I tried, there is an explicit order where all intermediate
> > modules are loaded in turn (with load_missing_constants),
> > even if only the deepest module would be strictly required to
> > evaluate X::Y::Z
> >
> > Could you show an example of where the problem you described
> > occurs. Or maybe this behavior has changed recently?
>
> Yeah I mixed things up a bit there as well as slightly mis-
> remembering. if you do things as rails expects, and rely entirely on
> the const_missing hooks then either way works. but if you deviate from
> that then things are different.
>
> For example if z.rb is
>
> module X
>  module Y
>    module Z
>    end
>  end
> end
>
> and you do require_dependency ''x/y/z'' then x/y.rb
won''t be loaded
>
> but if you do
>
> module X::Y::Z
> end
>
> then it will.
>
> I guess the point I was trying to make was that if ruby ends up
> reading z.rb then using module X::Y::Z guarantees that x.rb and x/y.rb
> will be read, but the other form doesn''t, although the usual
scenarios
> would mean that x.rb and x/y.rb would have been read earlier on
> anyway.
>

Thanks for the additional info. I was now able to reproduce your point.

.../lib/x/y$ cat z.rb
module X
  module Y
    module Z
      def self.foo
        "Z.foo"
      end
    end
  end
end

and using ''require_dependency'' the result the functionality in
the
file x/y.rb is not loaded when loading module Z:

$ rails c
Loading development environment (Rails 3.1.3)
001:0> ActiveSupport::Dependencies.logger = Rails.logger
002:0> ActiveSupport::Dependencies.log_activity=true
003:0> require_dependency ''x/y/z''
=> true

# In log this gives:
Dependencies: called require_or_load(".../lib/x/y/z.rb", nil)
Dependencies: loading .../lib/x/y/z
Dependencies: called load_file(".../lib/x/y/z.rb",
["X::Y::Z", "Y::Z", "Z"])
Dependencies: called new_constants_in("X::Y", "Y", :Object)
Dependencies: New constants: X::Y::Z, X
Dependencies: loading .../lib/x/y/z.rb defined X::Y::Z, X

#  X::Y::Z is defined as expected
#  X is also defined (that is an ancestor)
#  X::Y and Y are not defined (while these are ancestors too)

004:0> X::Y.foo
NoMethodError: undefined method `foo'' for X::Y:Module
from (irb):4
from .../gems/railties-3.1.3/lib/rails/commands/console.rb:45:in
`start''
from .../gems/railties-3.1.3/lib/rails/commands/console.rb:8:in `start''
from .../gems/railties-3.1.3/lib/rails/commands.rb:40:in `<top
(required)>''
from script/rails:6:in `require''
from script/rails:6:in `<main>''
005:0> require_dependency ''x/y''
loading module X / module Y in file ../lib/x/y.rb
=> true

# This shows in the log:
Dependencies: called require_or_load(".../lib/x/y.rb", nil)
Dependencies: loading .../lib/x/y
Dependencies: called load_file(".../lib/x/y.rb", ["X::Y",
"Y"])
Dependencies: called new_constants_in("X", :Object)
Dependencies: New constants:

006:0> X::Y.foo
=> "Y.foo"


With the version:

/lib/x/y$ cat z.rb
module X::Y::Z
  def self.foo
    "Z.foo"
  end
end

The result is:

$ rails c
Loading development environment (Rails 3.1.3)
001:0> ActiveSupport::Dependencies.logger = Rails.logger
002:0> ActiveSupport::Dependencies.log_activity=true
003:0> require_dependency ''x/y/z''
loading module X / module Y in file ../lib/x/y.rb
=> true

# log now shows a recursive resolution

Dependencies: called require_or_load(".../lib/x/y/z.rb", nil)
Dependencies: loading .../lib/x/y/z
Dependencies: called load_file(".../lib/x/y/z.rb",
["X::Y::Z", "Y::Z", "Z"])
Dependencies: called new_constants_in("X::Y", "Y", :Object)
Dependencies: called load_missing_constant(Object, :X)
Dependencies: called load_missing_constant(X, :Y)
Dependencies: called require_or_load(".../lib/x/y.rb", nil)
Dependencies: loading .../lib/x/y
Dependencies: called load_file(".../lib/x/y.rb", ["X::Y",
"Y"])
Dependencies: called new_constants_in("X", :Object)
Dependencies: New constants: X::Y
Dependencies: loading .../lib/x/y.rb defined X::Y
Dependencies: New constants: X::Y::Z, X
Dependencies: loading .../lib/x/y/z.rb defined X::Y::Z, X

004:0> X::Y.foo
=> "Y.foo"

Thanks again,

Peter

-- 
You received this message because you are subscribed to the Google Groups
"Ruby on Rails: Talk" group.
To post to this group, send email to
rubyonrails-talk-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
To unsubscribe from this group, send email to
rubyonrails-talk+unsubscribe-/JYPxA39Uh5TLH3MbocFF+G/Ez6ZCGd0@public.gmane.org
For more options, visit this group at
http://groups.google.com/group/rubyonrails-talk?hl=en.

Rails - Dec 2011 - This modules thing somehow really does not work for me, please help.

This modules thing somehow really does not work for me, please help.

Re: This modules thing somehow really does not work for me, please help.

Re: This modules thing somehow really does not work for me, please help.

Re: This modules thing somehow really does not work for me, please help.

Re: This modules thing somehow really does not work for me, please help.

Re: Re: This modules thing somehow really does not work for me, please help.