The Problem with Implicit Scoping in CoffeeScript

Post by cheald » Thu, Dec 22, 2011, 7:21 PM UTC

the_mitsuhiko wrote: post #96

mhansen wrote: post #90

I'll copy below jashkenas' longer answer from the old github issue about this, https://github.com/jashkenas/coffee-script/issues/712#issuec... """ Sorry, folks, but I'm afraid I disagree completely with this line of reasoning -- let me explain why: Making assignment and declaration two different "things" is a huge mistake. It leads to the unexpected global problem in JavaScript, makes your code more verbose, is a hug…

> As an existence proof, Ruby gets along just fine without it. Missing that Ruby stops scoping variables at a method and uses separate lexical scoping rules for constants thereby avoiding this issue mostly.

    class Foo
      def bar
        "bar"
      end

      def baz
        bar = "foo"
        puts bar
      end

      def bang
        puts bar
      end
    end

    foo = Foo.new
    foo.baz  # => "foo"
    foo.bang # => "bar"

Post by cheald » Thu, Dec 22, 2011, 7:57 PM UTC

showell30 wrote: post #122

cheald wrote: post #121

Earlier quoted context omitted.

Also, Ruby can shadow method names with local variable names just fine: class Foo def bar "bar" end def baz bar = "foo" puts bar end def bang puts bar end end foo = Foo.new foo.baz # => "foo" foo.bang # => "bar"

Well, this is a case where the equivalent CS behaves exactly like Ruby. class Foo bar: -> 'bar' baz: -> bar = "foo" console.log bar bang: -> console.log @bar() foo = new Foo() foo.baz() # => "foo" foo.bang() # => "bar"

    def bar
      "I called bar!"
    end

    def foo
      puts bar
      bar = "I manually assigned bar"
      return bar
    end

    puts foo()
    puts bar()

    # =>

    I called bar!
    I manually assigned bar
    I called bar!

    bar = ->
      "I called bar!"

    foo = ->
      console.log bar()
      bar = "I manually assigned bar!"
      return bar

    console.log foo()
    console.log bar()

    # =>

    I called bar!
    I manually assigned bar!
    TypeError: string is not a function

Post by showell30 » Thu, Dec 22, 2011, 8:25 PM UTC

cheald wrote: post #123

showell30 wrote: post #122

Earlier quoted context omitted.

Well, this is a case where the equivalent CS behaves exactly like Ruby. class Foo bar: -> 'bar' baz: -> bar = "foo" console.log bar bang: -> console.log @bar() foo = new Foo() foo.baz() # => "foo" foo.bang() # => "bar"

Not quite; @bar is "this.bar", and is an unambiguous reference, not just "bar". In Ruby, because self can be implicit, it's literally just "bar" in both cases, and Ruby gives the local variable precedence over the method name. You can still invoke the method via self.bar, but accessing the method as just "bar" is shadowed by the variable. Additionally, even though the method may be invoked by referencing bar, you can…

  bar = ->
    "I called bar!"

  foo = ->
    console.log bar()
    bar = "I manually assigned bar!"
    return bar

  console.log foo()
  console.log bar()
  

  log = (data) ->
    console.log data
    
  enhance_logging = ->
    i = 0
    log = (data) ->
      i += 1
      console.log i, data

  log "no line numbers"
  enhance_logging()
  log "line one"
  log "line two"
  

  > coffee foo.coffee 
  no line numbers
  1 'line one'
  2 'line two'

Post by barrkel » Fri, Dec 23, 2011, 4:27 AM UTC

jashkenas wrote: post #120

barrkel wrote: post #117

Earlier quoted context omitted.

You haven't said anything new to me here; I took all this for granted and it isn't what I'm criticizing. You haven't actually addressed my summary of the apparent reasoning for why accidental assignments to top-level symbols isn't a pathology of CoffeeScript. The bit I'm criticizing is the failure mode of "if you happen to try to use the same name for two different things within the same lexical scope, it won't work"…

I see. The reason I can't address your "failure mode" with an explicit suggestion is because it isn't a failure mode in CoffeeScript: If you assign a new value to a variable in an inner scope, the variable now has a new value. If you're thinking "I want to use the same name for two different variables" ... the answer is: choose a different (better) name for one of them. But perhaps I'm still not getting at the answer…

Post by barrkel » Fri, Dec 23, 2011, 1:14 PM UTC

showell30 wrote: post #128

barrkel wrote: post #127

Earlier quoted context omitted.

You keep stating and restating how and why CoffeeScript doesn't support shadowing of symbols via lexical scoping. That's not what I'm criticizing. Lexical shadowing is not what I'm advocating. I think your choice is fine in so far as it goes. It has its logic. What I am criticizing is how it can fail. The top scope is different, quantitatively and qualitatively, from almost all nested scopes. It's much larger, and sp…

barrkel, I think you should create a new programming language with more sensible scoping than CoffeeScript, so that developers can create bug-free code while working in parallel on thousand-line codebases without close reviews. I would start by using "$" as a sigil to prevent top-level variables from being reassigned in nested scopes. It would be a visual shorthand that indicates that you are definitely not creating…