New name for assert_lits_equal ?

[PragTob]

equal implies full equality aka == which seems confusing (and was pointed out as a downside to introducing assertions into our code base).

Not sure what the best name for this is, I introduced a small wrapper called assert_lists_contain_same for now - it's probably not the best for now happy for other variants but I wonder ig this would be a change considered for the library itself?

Wrapper for reference, yes it's just a module for now not a macro:

      import Assertions, except: [assert_lists_equal: 2, assert_lists_equal: 3]
      import AssertionAdjustments

defmodule ssertionAdjustments do
  require Assertions

  def assert_lists_contain_same(list_a, list_b) do
    Assertions.assert_lists_equal(list_a, list_b)
  end

  def assert_lists_contain_same(list_a, list_b, comparison) do
    Assertions.assert_lists_equal(list_a, list_b, comparison)
  end
end

[devonestes]

Funny enough, when I think of two lists as "equal," I always think of the contents and not the order, which is why I think this bug comes up so often in so many tests for folks. I (and I think many others) think of == as saying "Make sure these two lists have the same things in them," but what == really says is "Make sure these two lists have the same things in them, and make sure the lists are in the same order," so in my mind == is something more than "equal."

That said, I can totally see how this name might be confusing to many others, though, even though it makes sense to me, so I'm happy to explore a name change here. How about assert_lists_equivalent? Looking at the definition for equivalent I like this third one of corresponding or virtually identical especially in effect or function, so that seems like a potentially good fit. My biggest issue there is that I can never remember how to spell it, but tab complete should make that not so much of a problem 😉

[beligante]

Funny enough, when I think of two lists as "equal," I always think of the contents and not the order, which is why I think this bug comes up so often in so many tests for folks. I (and I think many others) think of == as saying "Make sure these two lists have the same things in them," but what == really says is "Make sure these two lists have the same things in them, and make sure the lists are in the same order," so in my mind == is something more than "equal."

That said, I can totally see how this name might be confusing to many others, though, even though it makes sense to me, so I'm happy to explore a name change here. How about assert_lists_equivalent? Looking at the definition for equivalent I like this third one of corresponding or virtually identical especially in effect or function, so that seems like a potentially good fit. My biggest issue there is that I can never remember how to spell it, but tab complete should make that not so much of a problem 😉

I share the same thoughts as you @devonestes .
Giving my 2cents on the discussion you've raised, there is another possible solution for your naming where assert_lists_equal could also accept an option to ensure the order. Like this

Assertions.assert_lists_equal(list_a, list_b, ensure_order: true)

This way you could keep the name as it is, and for devs it's cleaner what the default behaviour is

[timfjord]

Having the ensure_order option would be great

[PragTob]

I mean lists are inherently an ordered structure - otherwise sort etc. wouldn't make sense. It's totally legitimate to record a ranking of something in a list - and then lists with the same elements but in a different order aren't equal.

I don't think ensure_order is a good solution. If you want that you can just do assert list == other_list - I don't think assertions needs to replace default assertions.

I also don't like equivalent - as for many functions they are not equivalent (esp. pattern matching as order matters there.). List has a ton of function where the order matters (delete_at, insert_at, ends_with, first, last ....). As a word it's also very close to "equal".

I'll also raise that the wording of "equal" is especially confusing since ==/2 is explicitly called "equal":

Equal to operator. Returns true if the two terms are equal.

(emphasis mine)

I think when considering the definition of "equal" in a language, it'd be best to anchor it on the definition of equal in that language.

As such, I still think a rename would be good for intuitively making it more understandable what the helper tries to achieve.

[timfjord]

Unfortunately, assert list == other_list doesn't allow you to use the convenient assert_structs_equal (and other similar functions)