Re: [perl #129283] Compile-time check for $1, $2 om RHS of s///operator

Am 16.09.2016 um 23:36 schrieb Abigail:
> On Fri, Sep 16, 2016 at 04:15:13AM -0700, Ed Avis wrote:
>> # New Ticket Created by  "Ed Avis"
>> # Please include the string:  [perl #129283]
>> # in the subject line of all future correspondence about this issue.
>> # <URL: https://rt.perl.org/Ticket/Display.html?id=129283 >
>>
>> Currently this gives a run-time warning, but no compile-time error:
>>
>>     use 5.022;
>>     $_ = 'abc';
>>     s/(a)/x$1$2/;
>>     say;
>>
>> If the regexp on the LHS of s/// is known at compile time, then the
>> number of capturing groups it contains is known.  Under 'use strict',
>> a mention in the RHS of a capture variable which doesn't have a
>> corresponding capturing group (like $2 in the above example)
>> should give a compile-time error.
>>
>
>
> "use strict" never promotes a 'Use of uninitialized value' to an
> error. I don't think perl being consistent when it comes to s///
> is a bug.

I agree.

It's not that $2 doesn't exist there; it just contains undef. 'strict' 
seems the wrong place for this check.

I would like to see it as a warning, but I don't know how hard it would 
be to implement (e.g. you'd have to check for embedded regex matches 
that could reset capture variables).

-- 
Lukas Mai <plokinom@gmail.com>

• [perl #129283] Compile-time check for $1, $2 om RHS of s/// operator by Ed Avis
• Re: [perl #129283] Compile-time check for $1, $2 om RHS of s///operator by Abigail
• Re: [perl #129283] Compile-time check for $1, $2 om RHS of s///operator by Lukas Mai
• RE: [perl #129283] Compile-time check for $1, $2 om RHS of s///operator by Ed Avis
• RE: [perl #129283] Compile-time check for $1, $2 om RHS of s///operator by demerphq