Re: proposed addition to perlvar

On 9 June 2010 23:08, Abigail <[email protected]> wrote:
> On Wed, Jun 09, 2010 at 02:52:29PM +0000, Matt Sergeant wrote:
>> On Tue, 8 Jun 2010, Chas. Owens wrote:
>>
>>> How about:
>>>
>>> =item $#
>>>
>>> C<$#> used to be a variable that could be used to format printed number.
>>> After a deprecation cycle, its magic was removed in 5.10 and using it
>>> now triggers a severe warning: C<$# is no longer supported>.
>>>
>>> C<$#> is also used as sigil, which, when prepended on the name of an
>>> array gives the index of the last element in that array.
>>>
>>>    my @a            = ("a", "b", "c");
>>>    my $last_index   = $#a;     # $last_index is 2
>>>    my $last_element = $a[$#a]; # $last_element is "c"
>>
>> Do we want to "encourage" this kind of use or prefer to encourage the
>> [-1] format?
>
>
> Just because a construct is documented does not imply it's encouraged.
> It's important for documentation to be complete. In a section showing
> an example of '$#', the example should preferable have as few other
> things as possible - I would prefer the above over any foreach construct.
> These lines are about $#. Not foreach. Not ranges. And it is code that
> people will encounter - and consult the documentation for.
>
> Let's stay clear of a -1 vs. $#a preference debate. This part of the
> documentation is not the place.

I agree, but I think it would be the place to note that using $#arry
unguarded by an array size assertion may result in a fatal exception.

I suppose one might argue that we shouldnt throw a fatal exception
when we ask for the last element of an empty array anymore than we
should throw an exception when they ask for the first element, which
of course we dont.

cgeersm
Yves



-- 
perl -Mre=debug -e "/just|another|perl|hacker/"