[lnkForumImage]
TotalShareware - Download Free Software
Confronta i prezzi di migliaia di prodotti.
Asp Forum
 Home | Login | Register | Search 


 

Forums >

comp.lang.ruby

strict pragma & scope

Tom Allison

1/20/2006 10:41:00 AM

I'm still reading "the book" and am trying to see if I got this right.
Sometimes the book isn't direct enough on "this is how you will be programming
in ruby".

Is there anything like perl strict pragma in ruby?

Scope. It seems that there are some different approaches to scope. For
example, objects created in a loop remain in scope after you leave the scope of
the iterator. So my favorite temporary variables of i, x, foo, and bar are all
now permanent variable elements in my code blocks from their first use.
True/False?
11 Answers

Robert Klemme

1/20/2006 11:02:00 AM

0

Tom Allison wrote:
> I'm still reading "the book" and am trying to see if I got this right.
> Sometimes the book isn't direct enough on "this is how you will be
> programming in ruby".
>
> Is there anything like perl strict pragma in ruby?

Not exactly. Ruby has command line option -w that will help with
warnings.

> Scope. It seems that there are some different approaches to scope.
> For example, objects created in a loop remain in scope after you
> leave the scope of the iterator. So my favorite temporary variables
> of i, x, foo, and bar are all now permanent variable elements in my
> code blocks from their first use. True/False?

Both: it depends on how you do it.

>> def foo() 5.times {|i| p i}; puts "after", i end
=> nil
>> foo
0
1
2
3
4
NameError: undefined local variable or method `i' for main:Object
from (irb):4:in `foo'
from (irb):5
from :0

In this case you have to declare i before the loop block to see it
afterwards.

>> def foo() for i in 0..5 do p i end; puts "after", i end
=> nil
>> foo
0
1
2
3
4
5
after
5
=> nil

Kind regards

robert

Austin Ziegler

1/20/2006 1:06:00 PM

0

On 20/01/06, Tom Allison <tallison@tacocat.net> wrote:
> I'm still reading "the book" and am trying to see if I got this right.
> Sometimes the book isn't direct enough on "this is how you will be programming
> in ruby".

This is, in part, because not everyone programs in Ruby quite the same
way. I know that my style is very distinctive from others' style (and
my style is pretty "mainstream," nonetheless).

> Is there anything like perl strict pragma in ruby?

Not quite, but Ruby is more strict than Perl in any case. Running ruby
-w (even in your bangpath line) will give you more information.

> Scope. It seems that there are some different approaches to scope. For
> example, objects created in a loop remain in scope after you leave the scope of
> the iterator. So my favorite temporary variables of i, x, foo, and bar are all
> now permanent variable elements in my code blocks from their first use.
> True/False?

I believe that if you do a "for" loop, your statement is true.

irb(main):001:0> (1..10).each { |i| }
=> 1..10
irb(main):002:0> i
NameError: undefined local variable or method `i' for main:Object
from (irb):2
irb(main):003:0> for i in 1..10
irb(main):004:1> end
=> 1..10
irb(main):005:0> i
10

The idiomatic way to loop in Ruby is with #each or other iterator items.

Now, if you've defined i *outside* of your iterator, current Ruby will
overwrite it.

irb(main):001:0> i = 0
=> 0
irb(main):002:0> (1..10).each { |i| }
=> 1..10
irb(main):003:0> i
=> 10

I believe that there will be changes to this moving forward, but I
can't remember the exact scope changes.

-austin
--
Austin Ziegler * halostatue@gmail.com
* Alternate: austin@halostatue.ca

Gene Tani

1/20/2006 1:56:00 PM

0


Tom Allison wrote:
> I'm still reading "the book" and am trying to see if I got this right.
> Sometimes the book isn't direct enough on "this is how you will be programming
> in ruby".
>
> Is there anything like perl strict pragma in ruby?
>
> Scope. It seems that there are some different approaches to scope. For
> example, objects created in a loop remain in scope after you leave the scope of
> the iterator. So my favorite temporary variables of i, x, foo, and bar are all
> now permanent variable elements in my code blocks from their first use.
> True/False?

this was a good writeup of for and #each
http://www.sitepoint.com/forums/showthread.ph...

Gary Wright

1/20/2006 2:37:00 PM

0


On Jan 20, 2006, at 5:41 AM, Tom Allison wrote:
> Scope. It seems that there are some different approaches to
> scope. For example, objects created in a loop remain in scope
> after you leave the scope of the iterator. So my favorite
> temporary variables of i, x, foo, and bar are all now permanent
> variable elements in my code blocks from their first use.
> True/False?

The non-block looping constructs: while, unless, for/in do *not*
create a new scope.

The block looping constructs: each, loop do create a new scope.
The scoping behavior exists because of the block, not because the
methods each or loop have some special properties.

So basically your question is about blocks, not about loops.

Matz has acknowledge that the behavior of block arguments and block
local variables
is confusing and is exploring some alternatives.

See Austin's post for an illustration of the current behavior of
block arguments and
local variables within a block.



Gary Wright

Tom Allison

1/20/2006 8:06:00 PM

0


>> Is there anything like perl strict pragma in ruby?
>
>Not quite, but Ruby is more strict than Perl in any case. Running ruby
>-w (even in your bangpath line) will give you more information.
>


Thanks for the examples. As soon as I get back to my ruby installed
machine (not here) I'll play.
Meanwhile...
I don't think ruby -w quite does what perls 'use strict' pragma will
allow.

Under strict every variable must be identified, if not vartyped.
my $foo;

This prevents me from doing something like:

my $super_ids;
...
print $super_id,"\n";
(note the typo)
This is nice when it comes to trying to sort out bugs. Otherwise I'm
trying to find out why my variable $super_ids keeps coming up empty.
And after a few hours I'll probably realize that somewhere in my code I
messed up.

James Gray

1/20/2006 8:11:00 PM

0

On Jan 20, 2006, at 2:05 PM, Tom Allison wrote:

> I don't think ruby -w quite does what perls 'use strict' pragma will
> allow.

That's correct, -w is similar to Perl's warnings.

> Under strict every variable must be identified, if not vartyped.
> my $foo;
>
> This prevents me from doing something like:
>
> my $super_ids;
> ...
> print $super_id,"\n";
> (note the typo)

Ruby requires we declare a local variable before we use it, so you
are all set. ;)

James Edward Gray II

Logan Capaldo

1/20/2006 8:38:00 PM

0


On Jan 20, 2006, at 3:05 PM, Tom Allison wrote:

> my $super_ids;
> ...
> print $super_id,"\n";

In ruby...

super_ids = nil # assign to declare
...
print super_id, "\n"

If you run this you'll get
NameError: undefined local variable or method `super_id' for main:Object

Austin Ziegler

1/20/2006 8:59:00 PM

0

On 20/01/06, James Edward Gray II <james@grayproductions.net> wrote:
> > Under strict every variable must be identified, if not vartyped.
> > my $foo;
> >
> > This prevents me from doing something like:
> >
> > my $super_ids;
> > ...
> > print $super_id,"\n";
> > (note the typo)
>
> Ruby requires we declare a local variable before we use it, so you
> are all set. ;)

s/declare/set/

There are no variable declarations in Ruby.

The more likely problem that Tom is to find is with instance
variables. It would be nice to have some way of detecting missing
assignments to instance variables, but only have it show up when
requested.

-austin
--
Austin Ziegler * halostatue@gmail.com
* Alternate: austin@halostatue.ca

Jenda Krynicky

2/28/2007 1:11:00 PM

0

Logan Capaldo wrote:
> On Jan 20, 2006, at 3:05 PM, Tom Allison wrote:
>
>> my $super_ids;
>> ...
>> print $super_id,"\n";
>
> In ruby...
>
> super_ids = nil # assign to declare
> ...
> print super_id, "\n"
>
> If you run this you'll get
> NameError: undefined local variable or method `super_id' for main:Object


Which basically means that in this "language" there is NO WAY to declare
a variable. No way to say, "hey dude, I want a new (censored) x in
here". All you get is a RUNTIME error if you print (or compute with
or...) a variable sooner than you assign to it. Which is pretty useless.
And which means that if I do a

coll.each { |i| blah blah blah}

there is no telling whether it will use a new local i or some i
"declared" outside. And there is no way to make sure it does one or the
other.

And whenever I need to loop through something I have to make sure i use
a different variable than any that could ever be "declared" in the same
scope. Don't you think it's a bit stupid? You probably don't, you
program in Ruby.

--
Posted via http://www.ruby-....

Rob Biedenharn

2/28/2007 2:50:00 PM

0


On Feb 28, 2007, at 8:10 AM, Jenda Krynicky wrote:

> Logan Capaldo wrote:
>> On Jan 20, 2006, at 3:05 PM, Tom Allison wrote:
>>
>>> my $super_ids;
>>> ...
>>> print $super_id,"\n";
>>
>> In ruby...
>>
>> super_ids = nil # assign to declare
super_ids (plural)
>> ...
>> print super_id, "\n"
super_id (singular)
>>
>> If you run this you'll get
>> NameError: undefined local variable or method `super_id' for
>> main:Object
>
>
> Which basically means that in this "language" there is NO WAY to
> declare
> a variable. No way to say, "hey dude, I want a new (censored) x in
> here". All you get is a RUNTIME error if you print (or compute with
> or...) a variable sooner than you assign to it. Which is pretty
> useless.
> And which means that if I do a
>
> coll.each { |i| blah blah blah}
>
> there is no telling whether it will use a new local i or some i
> "declared" outside. And there is no way to make sure it does one or
> the
> other.
>
> And whenever I need to loop through something I have to make sure i
> use
> a different variable than any that could ever be "declared" in the
> same
> scope. Don't you think it's a bit stupid? You probably don't, you
> program in Ruby.

No, you just need to know if YOU created such a variable within the
current lexical scope. Your 'i' in the block will always exist
INSIDE the block. In Ruby 1.8, that reuses the 'i' that already
exists in the current scope, but 1.9/2.0 changes this (I believe;
haven't personally checked this out).

Of course, if you can't consistently spell your variables the same
way, the "language" isn't likely to know what you meant.

-Rob

Rob Biedenharn http://agileconsult...
Rob@AgileConsultingLLC.com