Clarification on multiline defs

[vincent-lg]

Greetings,

First of all, thank you for the guide.

I have some difficulty understanding two points that seem (to me) to conflict. I'm not saying they're not exact, but I have trouble guessing what they mean, maybe others will too:

• 
  Run single-line defs that match for the same function together, but separate
  multiline defs with a blank line.
  ^[link]

  def some_function(nil), do: {:error, "No Value"}
  def some_function([]), do: :ok
  
  def some_function([first | rest]) do
    some_function(rest)
  end

• 
  If you have more than one multiline def, do not use single-line defs.
  ^[link]

  def some_function(nil) do
    {:error, "No Value"}
  end
  
  def some_function([]) do
    :ok
  end
  
  def some_function([first | rest]) do
    some_function(rest)
  end
  
  def some_function([first | rest], opts) do
    some_function(rest, opts)
  end

The way I see it, the second point seems to contradict the first one. No one-line clause whenever we have a clause of the same function spanning multiple lines? This arises a lot when dealing with recursion. Some clauses (the ones dealing with empty collections in particular) can be extremely short while some can be more verbose. As a teacher (and writer), I'd prefer to follow some consistency.

From what I've seen, the first point seems to be used by most Elixir developers: one-line clauses are grouped together but multiline clauses are isolated by a blank line. That's the convention I personally follow.

I probably misunderstood the second point. A clarification (on this issue or in the README itself) would be great.

Thank you again,

[christopheradams]

Given that the formatter enforces blank lines between multiline defs, the style conventions could be reworded as:

1. Don't separate single-line defs of the same function with blank lines.
2. If you have more than one multiline def of the same function, don't use single-line style at all.

The first rule seems like a sensible style suggestion. The second one I'm unsure about. I can imagine (like you suggest) a recursive function with a series of single-line defs, followed by two multi-line defs, followed by another series of single-line defs, etc. Re-writing them as all as multiline just to follow this style rule wouldn't seem sensible.

Does anyone have strong opinions about the multiple-function-defs rule?

See #39 and #116

[dotdotdotpaul]

A happy medium might be: If your function has multiple implementations, single-line implementations at the "top" should be single-spaced (ie. don't separate the lines), but after the first multi-line implementation, double-space (blank line between) all subsequent functions, and don't use single-line functions after. There are a few cases I can think of where this would be a pain in the butt, but few and far between. This probably comes up mostly with a catch-all function that "just returns an error" and would love to make that a single line function. But personally, my "style" would look like: # Batch the one-liners at the top def foo(nil), do: nil def foo(<simple case>), do: nil # Now space out for the multi-line functions def foo(<case A>) do ... end def foo(<case B>) do ... end # Would love to make this a one liner... # def foo(arg), do: {:error, "Invalid arg #{inspect(arg)}"} # But should be a multi-line to be consistent... def foo(arg) do {:error, "Invalid arg #{inspect(arg)}"} end I think mixing single- and multi-line implementations is okay, but limit the single-line implementations "to the top". "Once you go multi-, you never go back." :) ...Paul

…

On Mon, Jul 11, 2022 at 3:55 AM Christopher Adams ***@***.***> wrote: Given that the formatter enforces blank lines between multiline defs, the style conventions could be reworded as: 1. Don't separate single-line defs of the same function with blank lines. 2. If you have more than one multiline def of the same function, don't use single-line style at all. The first rule seems like a sensible style suggestion. The second one I'm unsure about. I can imagine (like you suggest) a recursive function with a series of single-line defs, followed by two multi-line defs, followed by another series of single-line defs, etc. Re-writing them as all as multiline just to follow this style rule wouldn't seem sensible. Does anyone have strong opinions about the multiple-function-defs <https://github.com/christopheradams/elixir_style_guide#multiple-function-defs> rule? See #39 <#39> and #116 <#116> — Reply to this email directly, view it on GitHub <#204 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAANTBKROWKKTRLVBPAFGJTVTOEKNANCNFSM53FONQIQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[christopheradams]

Looking again at the multiple-function-defs rule, it was introduced in #39 in order to close #23, but the stated rationale of "not mixing single-line and multi-line for more than one multi-line def" was accepted without discussion.

I think we can agree on the rule to remove blank lines between single-line defs and add them between multi-line defs:

def some_function(nil), do: {:error, "No Value"}
-
def some_function([]), do: :ok
+
def some_function([first | rest]) do
  some_function(rest)
end

But if we were to add another multi-line def, do we really need to re-write everything to be multi-line? Any small benefit to readability I think is outweighed by the downsides of having to re-write perfectly fine single-line defs, which would complicate the commit diff (and any time you re-write something you could introduce bugs or regressions).

I would propose dropping the multiple-function-defs rule since it seems unnecessarily difficult to understand, use, and maintain the style.