Feature #10318: [PATCH 0/n] Let underscore be positionally matched arg to omit binding obvious variable.

• Status: Open
• Assignee: -
• Target version: -

[gogotanaka (Kazuki Tanaka)]

I’m so sorry for just proposing this future without PATCH right now ...

(I suppose it contains too much issues.)

After some discussions, let me implement this future.

Of course I need to obtain your consent.

Anyway, I want to let underscore be positionally matched arg and omit to binding obvious variable when it comes to proc, lambda and block .

# let items be a Array of String. (e.g. items = ['foo', 'bar', 'foobar']

# Before
items.select { |item| item.include?('foo') }

# After
items.select { _.include?('foo') }

Even if block has more than 2 args, I want to omit one of them.

(I know it has much pain ... just idea)

# let nums be a Array of Integer. (e.g. nums = [1, 2, 3, 4]

# Before
nums.inject { |sum, num| sum + num }

# After
nums.inject { |sum| sum + _ }
#↑ It seems to difficult to hundle .. but looks aesthetic or understandable.

Let me know how do you think about it or what problem is.

thank you. gogo.

[gogotanaka (Kazuki Tanaka)]

• File convert_tabs_to_spaces.patch convert_tabs_to_spaces.patch added

[sawa (Tsuyoshi Sawada)]

When there is more than one parameter passed to a block, your proposal will create a problem.

When the number of parameters passed to a block is more than the number of block variables in ||, then destructive assignment should apply, but your proposal will make this not happen. It will be incompatible with many existing code.

[nobu (Nobuyoshi Nakada)]

Why expanding tabs?

[gogotanaka (Kazuki Tanaka)]

@Tsuyoshi Sawada san

Thank you for your comment. I got your point.

As you said (and I mentioned also) it has much pain, so I can't implement right now.

How do you think about the case giving a argument?

I mean like that.

['foo', 'bar', 'foobar'].select { _.include?('foo') }
['foo', 'bar', 'foobar'].each { eval(_) }

[gogotanaka (Kazuki Tanaka)]

@Nobuyoshi Nakada san

Thank you for your comment.

Starting to implement this issue, I find spaces and tabs being mixed.

And I suppose this might make us confusion and there are no reason spaces and tabs be mixed, but I am afraid a conflict.

This is why I upload this patch at first.

[nobu (Nobuyoshi Nakada)]

Welcome to the underground holy war. ;)
You may know that ruby-style.el has (indent-tabs-mode . t).

[gogotanaka (Kazuki Tanaka)]

@Nobuyoshi Nakada san

It was an honor to encounter such a underground holy war.

Sure. I'll follow what you said.

Thank you.

[gogotanaka (Kazuki Tanaka)]

oops... I found totally incompatible code in Rails project.
I gave up this approach. Nonetheless, I still wanna introduce concept omitting obvious variable. (Like Scala's lambda calculus)
#10394 does make sense so far?

[naruse (Yui NARUSE)]

• Target version deleted (3.0)