From: David Alan Black <dblack@...>
Date: 2001-10-29T08:50:29+09:00
Subject: [ruby-talk:23700] Re: newbie syntax question?

Hello --

On Mon, 29 Oct 2001, Chochain Lee wrote:

> Hi,
>
> I ran into something like the following code which gave me a small
> surprise. Can some explain why?
>
> puts (123 && 456).type    # => 456
> puts ((123 && 456).type)  # => Fixnum

In the first one, you're asking for Ruby to evaluate the type of the
expression "puts (123 && 456)".  As a side effect, that expression
prints the value of (123 && 456) (which is 456).  And then it
evaluates the type (which actually is NilClass, because that's the
return value of puts).  But it doesn't do anything with it :-)

In other words, your first example is parsed as:

  (puts(123 && 456)) .type

In the second one, by grouping all that stuff together you're asking
Ruby to, *first*, to evaluate (123 && 456), *then* to determine the
type of the result (which is Fixnum), and *then* to print out the
result of that determination.

If you rewrite the first one like this, you might see more clearly
what's happening:

   puts(123 && 456) .type

The (123 && 456) binds more tightly to puts than it does to .type.
In fact, if you use the -w flag, Ruby will warn you about this:

  candle:~$ ruby -we 'puts (123 && 456).type'
  -e:1: warning: puts (...) interpreted as method call
  456

That warning means: even though the gap before the "(" makes it look
like the (...) part binds with ".type", it really binds as
(puts(...)) .type.  (The way not to get this warning is to not put
a space between the method name and the parens.)


David

-- 
David Alan Black
home: dblack@candle.superlink.net
work: blackdav@shu.edu
Web:  http://pirate.shu.edu/~blackdav