RCR 331: Remove 'true' parameter from #attr

I would like to see *all* Boolean parameters replaced with expressive methods names (instance_methods(false) being another one that comes to mind. I'm less sold on the attr :foo= thing... it kind of violates the tidiness of just saying :foo and having the automatic syntactic sugar assignment thing. Then again, it only violates it in a way that's invisible when it's actually used.

I would actually think it would make sense for attr to do what attr_accessor does currently. _reader and _writer could then be the specializations.

David Black

---

I agree with David. It doesn't make sense that attr_writer, attr_reader, and attr_accessor accept multiple attribute names, and attr only accepts one. I think the most common case is attr_accessor, so attr should match that:

 class Foo
   attr :bar, :baz, :bur
 end

---

I agree with David Black.

---

I think it would be better either way, thoug I lean toward do less rather then more and just make it the same a attr_reader. I think #attr itself is used rarely enough that 'attr :foo, :foo=' are a fine way to handle both in its case. I also have to confess that I fins myself shying more and more away from these writters and am leaning toward #attr_setter, i.e. where 'foo' would return the value 'foo 42' would set it. I think Ara is on to something here. Obviously thats another reason I think alias :attr :attr_reader --cause otherwise I say alias :attr :attr_setter ;)

T.

---

I'm with David Black on this one.

---

i'm simply for replacing attr_accessor with attr - since it's the most commonly used anyway. attr :foo #=> #foo; #foo= attr_reader :foo #=> #foo attr_writer :foo #=> #foo=

i'm less in favor with the use of foo(42) to assign stuff though - i think having = makes the whole thing nice to use (+= comes to mind)

---

I am not in favour of the originally submitted suggestion; I would rather make attr_accessor and attr aliases for one another, not attr_reader and attr. You would still have special handling for this case if you make it so that "attr :foo, :foo=" does what you're suggesting. --Austin

---

Hmm... Do you all really use attr_accessor that much? I find I usually only need attr_reader when I think about it. That's why I'd prefer the finer control of attr :foo, :foo=. OTOH, Perhaps a Railsism?

  attr :foo, :writer => true

In anycase, either way is better than no way.

--T.

---

In anycase, either way is better than no way.

I take that back "alias :attr :attr_accessor" can have bad backward compatability results.

  class X
    def foo=(x)
      ...
    end
    attr :foo
  end

And

  class X
    attr :foo
    def method_missing(sym,*args)
      if /=$/ =~ sym.to_s
        ...
  end

Etc.

T.

---

I am OK for this proposal, once we can settle down the exact behavioir. Should attr be alias to the attr_accessor, or attr_reader? I think I prefer making it alias to attr_reader, since it encourages encapsulation, which is good. Any opinion?

matz.

---

I prefer attr_accessor because I *do* use attr_accessor that often. -Austin

---

That said, as long as one is chosen and documented that way, I see no problem with either one. I will probably continue to use attr_accessor even if attr is made equivalent. What I (personally) would rather not see is "attr :foo=" being meaningful; if attr is going to mean attr_reader, fine. But attr_reader :foo= fails for a meaningful reason. Alternatively, we can look at adopting something like Ara Howard's "traits" library which takes the whole idea to a very interesting level -- I don't use it precisely because it isn't core. -Austin

---

I agree with Austin, if its documented, it doesnt really matter which one is picked.

---

"What I (personally) would rather not see is "attr :foo=" being meaningful"

Although, I point out that it is meaningful when aliasing an accessor:

  attr_accessor :a
  alias :b, :a
  alias :b=, :a=

T.

---