Rbs weird behaviour on modules

SR20VET

Today I ran into this weird behaviour when scrambling my scripts into an rbs.

Lets say I have the folowing script:

module MyModule1
  def self.load_second_script
    Sketchup;;require 'second_script'
  end#def
end#module

in seconds_script.rb(s) i have:

module MyModule1
  module MyModule2

  end#module
end#module

next, I do:

MyModule1.load_second_script

While testing with rb, this eventually gives

MyModule1;;MyModule2

as a valid module.
With rbs this results in

MyModule1;;MyModule1;;MyModule2

Has anyone noticed this before?

Regards

Dan Rathbun

@sr20vet said:

Has anyone noticed this before?

YES. ThomThom first noticed this, and has filed a bug report.

The error is within the Sketchup::require() method.

To workaround, you will need to force the evaluation to occur within the TOPLEVEL_BINDING, thus:

module MyModule1
  def self.load_second_script
    eval("Sketchup;;require('second_script')",TOPLEVEL_BINDING,'second_script.rbs',1)
  end#def
end#module

thomthom

It also happens with SketchupExtension classes - where you cannot bypass the Sketchup::require call and I had to create a proxy .rb file to load the actual .rbs file.

SR20VET

Ok, thanks for the workaround. You should expect that no mather you use rb's or rbs's, they behave the same way...