Front page | perl.perl5.porters |
Postings from December 2010
On 2010-12-02 10:33, Abigail wrote:
> On Wed, Dec 01, 2010 at 07:38:30PM +0100, Dr.Ruud wrote:
>> On 2010-12-01 00:24, Abigail wrote:
>>> On Tue, Nov 30, 2010 at 10:34:36PM +0100, Dr.Ruud wrote:
>>>> On 2010-11-29 10:22, Dave Mitchell wrote:
>>>>> On Sun, Nov 28, 2010 at 05:16:28PM -0800, Father Chrysostomos wrote:
>>>>>> I know this is what constant.pm used to use, but it causes a bug:
>>>>>
>>>>> What do you think is the bug? It seems to be behaving the way I expect.
>>>>>
>>>>>> $ perl -MO=Deparse -e' BEGIN{my $x = 5; *foo = sub(){$x}; $x=6} print foo'
>>>>>> sub BEGIN {
>>>>>> my $x = 5;
>>>>>> *foo = sub () {
>>>>>> $x;
>>>>>> }
>>>>>> ;
>>>>>> $x = 6;
>>>>>> }
>>>>>> print 5;
>>>>
>>>> The compiler should patch the constsub at the '$x = 6' line.
>>>
>>> If the compiler can detect that afterwards $x won't change - but you
>>> need to be able to solve the halting problem to do that in general.
>>> For instance,
>>>
>>> BEGIN {
>>> my $x = 5;
>>> *foo = sub () {$x};
>>> *bar = sub {$x = shift};
>>> }
>>>
>>> the compiler could only patch the constsub if it knows bar() won't
>>> be called.
>>
>> And that is not '$x = CONSTANT', so that should then be a compile error.
>
> You're proposing a sub of the form 'sub {$x = shift;}' should be
> a compile error?
Indeed. The $x there, is (an alias for) a constant, and the compiler
know that. The other option is to lose the constantification of foo(),
which is unwanted.
A 'const' keyword would help.
--
Ruud
Thread Previous
|
Thread Next