Forum: Ruby Meaning of "<<"?

Announcement (2017-05-07): www.ruby-forum.com is now read-only since I unfortunately do not have the time to support and maintain the forum any more. Please see rubyonrails.org/community and ruby-lang.org/en/community for other Rails- und Ruby-related community platforms.
Chris D. (Guest)
on 2009-02-13 16:06
(Received via mailing list)
Hi,

I'm currently trying to learn ruby, coming from a perl/unix
background. I've been working through one of the many online tutorials
and it's suddenly thrown in the "<<" operator.

Now, I'm familiar with a the use of "<<" as a bitwise left shift
and as a here doc, but this one throws me. The relevant URL is
http://rubylearning.com/satishtalim/mutable_and_im...
and here's a variation on the theme for illustrative purposes:

    str = 'one'
    str << 'two'
    puts str    # onetwo

It seems to me that << in this context is equivalent to +=, but I can't
find any documentation explaining /why/ this is the case, and why I
should use one in preference to the other. (Google really doesn't like
"<<".)

Any suggestions, please?
Cheers,
Chris
saurabh purnaye (Guest)
on 2009-02-13 16:15
(Received via mailing list)
Hi Chris,
If you see the string class documentation of ruby,
http://www.ruby-doc.org/core/classes/String.html#M000822
Its for concatenation..
You may find more here http://www.ruby-doc.org/

On Fri, Feb 13, 2009 at 7:33 PM, Chris D.
<removed_email_address@domain.invalid>wrote:

>
> Cheers,
> Chris
>
>


--
--
Thanks and Regards
Saurabh P.
+91-9922071155
skype: sorab_pune
yahoo & gtalk: saurabh.purnaye
msn:  removed_email_address@domain.invalid
Dido S. (Guest)
on 2009-02-13 16:18
(Received via mailing list)
Ruby is capable of doing operator overloading, and the << operator has
been consistently overloaded to do stuff like appending data (I
believe this was an idea that came from C++). Note also that use of <<
on strings is much faster than +=, since += will create a new string,
append the string you're adding to it, and replace the old string by
the new one, while, if I'm not mistaken << will append the data on its
right side in place, if I recall correctly.

Several other classes also overload the << operator. Using << on an
array will append the right hand side to the end of the array
(equivalent to Array#push). << on an IO object will write the object
on the right hand side, converting it to a string, just as someone
coming from C++ might expect.
Stefano C. (Guest)
on 2009-02-13 16:29
(Received via mailing list)
Alle Friday 13 February 2009, Chris D. ha scritto:
>
> Cheers,
> Chris

This is the documentation for the String << operator (that is, for the
String#<< method):

     str << fixnum        => str
     str.concat(fixnum)   => str
     str << obj           => str
     str.concat(obj)      => str

     From Ruby 1.8
------------------------------------------------------------------------
     Append---Concatenates the given object to _str_. If the object is a
     +Fixnum+ between 0 and 255, it is converted to a character before
     concatenation.

        a = "hello "
        a << "world"   #=> "hello world"
        a.concat(33)   #=> "hello world!"

The difference between << and + is that << modifies the first string,
while +
creates a new string, made of the caracters of the first string followed
by
the characters of the second. Too see the difference, look at this
example:

a = "first string"
b = "second string"
puts "Using +"
a + b
puts a
puts b
puts "Using <<"
a << b
puts a
puts b

The output is:

Using +
first string
second string
Using <<
first stringsecond string
second string

As you can see, in the second case, a has been modified, while in the
first
case it wasn't.

In general, to obtain documentation about ruby classes and methods, you
can
use the ri command:

ri 'String#<<'

which will give the documentation I posted above. There's also an online
API
reference at http://www.ruby-doc.org/core/

I hope this helps

Stefano
Chris D. (Guest)
on 2009-02-13 19:05
(Received via mailing list)
Chris D. <removed_email_address@domain.invalid> wrote:
> It seems to me that << in this context is equivalent to +=, but I can't
> find any documentation explaining /why/ this is the case [...]

Thank you all. Very enlightening.
Chris
Herman M. (Guest)
on 2009-02-13 21:10
(Received via mailing list)
Chris D. wrote:
> Now, I'm familiar with a the use of "<<" as a bitwise left shift
> and as a here doc, but this one throws me. [...]


Think of it as a method, not as an operator.
Every class can define a method that is called '<<'
just like any class can define a method called 'print' or 'count'.

Don't let this syntax fool you:

str = "basic"
str << "appended"

It is just a nice to read way of saying the
following, which explains it and works, too:

str = "basic"
str.<<( "appended" )

Got it?

Hrmn
Julian L. (Guest)
on 2009-02-14 07:05
(Received via mailing list)
I thought + was for concatenation. << actually modifies Its receiver

Blog: http://random8.zenunit.com/
Learn rails: http://sensei.zenunit.com/

On 14/02/2009, at 1:12 AM, saurabh purnaye 
<removed_email_address@domain.invalid>
Julian L. (Guest)
on 2009-02-14 07:12
(Received via mailing list)
I thought operator overloading was a non-oop way of specifying
multiple functions that each take different argument numbers? That's
not what's going on here. This is messages having the same name
exactly being sent to different classes having slightly different
meanings depending on the class of the receiver. It's encapsulation,
really... Each class encapsulates its meanings and can obviously have
different meanings for the same method names depending on the classes
in question.

Blog: http://random8.zenunit.com/
Learn rails: http://sensei.zenunit.com/
Gary W. (Guest)
on 2009-02-14 08:20
(Received via mailing list)
On Feb 14, 2009, at 12:10 AM, Julian L. wrote:

> I thought operator overloading was a non-oop way of specifying
> multiple functions that each take different argument numbers? That's
> not what's going on here.

Agreed. This is not 'overloading' in the C++ sense of the term.

> This is messages having the same name exactly being sent to
> different classes having slightly different meanings depending on
> the class of the receiver. It's encapsulation, really... Each class
> encapsulates its meanings and can obviously have different meanings
> for the same method names depending on the classes in question.

Disagree. Encapsulation doesn't seem like the right concept to
describe this pattern.

I think that 'abstraction' might be a better description.  When the
same method provides analogous semantics across multiple classes, the
method name becomes a sigil for the abstract semantics.  So '<<' is a
sigil for 'alter via concatenation' (IO, String, Array) but is also a
sigil for 'shift to the left' (Fixnum, Date, IPAddr).

The term 'protocol' is often used to describe this sort of pattern
across classes.

Good library design comes from exploiting these sorts of patterns.

Gary W.
Chris D. (Guest)
on 2009-02-14 11:05
(Received via mailing list)
Herman M. <removed_email_address@domain.invalid> wrote:
> It is just a nice to read way of saying the
> following, which explains it and works, too:

> str = "basic"
> str.<<( "appended" )

> Got it?

More syntatic sugar, like the "def instance_var=" setter method?

Thanks,
Chris
7stud -. (Guest)
on 2009-02-14 11:23
Julian L. wrote:
> I thought operator overloading was a non-oop way of specifying
> multiple functions that each take different argument numbers?
>

That's called *function* overloading.  In C++, function overloading and
operator overloading are two different things.  Operator overloading
requires a specific function signature for a given operator.  Function
overloading on the other hand is used to create multiple function
signatures for a given function name.
Julian L. (Guest)
on 2009-02-14 17:24
(Received via mailing list)
Ah! What's operator overloading then?

Blog: http://random8.zenunit.com/
Learn rails: http://sensei.zenunit.com/
Rick D. (Guest)
on 2009-02-14 20:47
(Received via mailing list)
On Sat, Feb 14, 2009 at 4:21 AM, 7stud -- <removed_email_address@domain.invalid>
wrote:

> Julian L. wrote:
> > I thought operator overloading was a non-oop way of specifying
> > multiple functions that each take different argument numbers?
> >
>
> That's called *function* overloading.  In C++, function overloading and
> operator overloading are two different things.


That's not the way I see it.  In C++ operator overloading is
accomplished by
defining an overloaded function with special syntactic sugar to give it
the
same name as an existing operator:

e.g.

Class Complex {
   Complex operator+(int);
   Complex operator+(double);
   Complex plus(int);
   Complex plus(double);
   ...
};

This class declares that it has both overloaded versions of the binary +
operator as well as of the plus member function.


Operator overloading
> requires a specific function signature for a given operator.  Function
> overloading on the other hand is used to create multiple function
> signatures for a given function name.
>

As indicated above, they both do this.

Overloading means that the compiler distinguishes at compile time
between
calls to the same function or operator with the same name, based on the
type(s) of the arguments.

As this depends on having static types, overloading, either of functions
or
operators really isn't a concept which is applicable to dynamically
typed
languages like Ruby.  Use cases like mixed-mode arithmetic are handled
by
different techniques like the Numeric.coerce method in Ruby or double
dispatching in Smalltalk.

A side note on C++.  Overloading is one of the mechanisms C++ provides
to
define polymorphism.  Virtual functions are another, and probably better
known, Virtual functions can be overridden (which is different from
overloaded) by providing a new implementation in a subclass.  Although
virtual functions can be overloaded, it's not for the faint of heart:
http://answers.google.com/answers/threadview?id=379269

It's a good example of why, when you scratch a little below the surface,
statically typed languages like C++ and dynamically typed language like
Ruby
are very different beasts and it's dangerous to try to apply knowledge
of
one to the other without some deep thinking.

--
Rick DeNatale

Blog: http://talklikeaduck.denhaven2.com/
Twitter: http://twitter.com/RickDeNatale
7stud -. (Guest)
on 2009-02-14 20:49
Rick Denatale wrote:

> Operator overloading
>> requires a specific function signature for a given operator.  Function
>> overloading on the other hand is used to create multiple function
>> signatures for a given function name.
>>
>
> As indicated above, they both do this.
>

I stand corrected.  Thank you.
Ken B. (Guest)
on 2009-02-16 04:40
(Received via mailing list)
On Sat, 14 Feb 2009 04:21:42 -0500, 7stud -- wrote:

> Julian L. wrote:
>> I thought operator overloading was a non-oop way of specifying multiple
>> functions that each take different argument numbers?
>>
>>
> That's called *function* overloading.  In C++, function overloading and
> operator overloading are two different things.  Operator overloading
> requires a specific function signature for a given operator.  Function
> overloading on the other hand is used to create multiple function
> signatures for a given function name.

Ruby restricts you to defining operators in the class of the left-hand
side object. C++ has no such restriction. You could presume that in C++,
the + operator is defined in the global namespace as:

  int operator + (int, int)
  double operator + (double, double)
  double operator + (int, double)
  double operator + (double, int)

Thus, when you create a class Date and define the + operator as
  Date operator + (Date const &, int)
you really are overloading the + operator (in the function overloading
sense of the term).

C++ also allows you to define
class Date{
  public:
    Date operator+ (int);
}
which equivalent to the global namespace version, and the name operator
overloading comes along for the ride.

Ruby, which is more restrictive, doesn't have function overloading at
all, so operator overloading has taken on a new meaning (simply
redefining the operator on a new class) that's a bit of a misnomer if
you
try to relate it back to the C++ term whence it originated.

--Ken
This topic is locked and can not be replied to.