Feature #7346: object(...) as syntax sugar for object.call(...)

• Status: Rejected
• Assignee: matz (Yukihiro Matsumoto)
• Target version: 3.0

[rosenfeld (Rodrigo Rosenfeld Rosas)]

I propose for the parser to interpret "object(...)" as "object.call(...)". It should raise NoMethodError at runtime if object doesn't respond to "call".

This would read better than using "call":

double = -> n { n * 2 }

double(3) == 6

Sorry if this has already been proposed before (and rejected) but I couldn't find any references to something like this using Redmine's search interface.

[matz (Yukihiro Matsumoto)]

• Status changed from Open to Rejected

I have once tried, but it caused serious incompatibility problem for example:

p = Object.new
p(15)

So compromise with object.() syntax introduced in 1.9.

Matz.

[rosenfeld (Rodrigo Rosenfeld Rosas)]

Ah, ok, I didn't know about this syntax until now. What does the code above do?

[matz (Yukihiro Matsumoto)]

We easily forget conflict between method names and variable names, in a language like Ruby, where methods and variables have separated name space.

We expect p(15) to print 15 even when we have a variable named p in the scope.

Matz.

[rosenfeld (Rodrigo Rosenfeld Rosas)]

Ah, of course! :D I totally forgot about Kernel#p! :P

Yes, that makes total sense.

[nathan.f77 (Nathan Broadbent)]

@rosenfeld, I'll just mention that you can use Proc#[] in your example:

double = -> n { n * 2 }

double[3] == 6 #=> true

On Wednesday, 14 November 2012, rosenfeld (Rodrigo Rosenfeld Rosas) wrote:

Issue #7346 has been updated by rosenfeld (Rodrigo Rosenfeld Rosas).

Ah, of course! :D I totally forgot about Kernel#p! :P

Yes, that makes total sense.

Feature #7346: object(...) as syntax sugar for object.call(...)
https://bugs.ruby-lang.org/issues/7346#change-32860

Author: rosenfeld (Rodrigo Rosenfeld Rosas)
Status: Rejected
Priority: Normal
Assignee: matz (Yukihiro Matsumoto)
Category: core
Target version: Next Major

I propose for the parser to interpret "object(...)" as "object.call(...)".
It should raise NoMethodError at runtime if object doesn't respond to
"call".

This would read better than using "call":

double = -> n { n * 2 }

double(3) == 6

Sorry if this has already been proposed before (and rejected) but I
couldn't find any references to something like this using Redmine's search
interface.

--
http://bugs.ruby-lang.org/

[rosenfeld (Rodrigo Rosenfeld Rosas)]

Yes, I know, it is just that I prefer to read object.call(arguments) than object[arguments]. This is just a personal opinion, I know and I can change my mind some day about this :)