release 1.220 on tuesday, 2014-11-11

[wchristian]

mainly a not for myself

[wchristian]

And 1.219_001 is on CPAN for this.

Summary:

• incompatible behavior fixes on PPI::Statement::Sub->prototype
• improved parsing of various syntax elements
• code quality improvements

Details:

• {} is now recognized as anonymous hash constructor instead of a code block after these operators: &&= //= || && // ? : (GitHub #36) (MOREGAN)
• regex capture variables greater than $9 are now parsed completely, instead of being parsed as single digit captures with numbers after them (GitHub #38) (MOREGAN)
• DESTROY and AUTOLOAD subs are now parsed even without the sub keyword (GitHub #39) (MOREGAN)
• PPI::Statement::Sub->prototype behavior now matches its documentation, instead of returning the prototype string unchanged and still including the parens (GitHub #56) (MOREGAN)
• PPI::Statement::Sub->prototype now returns undef on subs without a prototype, instead of returning an empty string (GitHub #56) (MOREGAN)
• list of keywords which are not parsed as packages when followed by the Perl4 package separator ' has been increased (GitHub #77) (MOREGAN)
• application of a number of Perl::Critic policies and documentation fixes (GitHub #53) (MOREGAN, MITHALDU)
• automation of README.md generation for git (GitHub #86) (COWENS)

[jeffreykegler]

The bleeding edge Github version on Marpa::R2 tests all OK.

[thaljef]

With respect to Perl-Critic/Perl-Critic#192 and Perl-Critic/Perl-Critic#390, this release does parse certain hash constructors better, but it still misses in some cases, like this: %{$a_hash_ref}

[thaljef]

I can confirm that it fixes Perl-Critic/Perl-Critic#455

[wchristian]

@jeffreykegler : Thanks. :)

@thaljef : Can you give me a concrete failure example, i.e. code i can run and debug? I see that %{$a_hash_ref} parses as a cast, followed by a block containing a variable, which looks correct to me. What should it be instead?

[thaljef]

%{$a_hash_ref}

Isn't that an anonymous hash constructor instead of a block?

[thaljef]

I can also confirm that this fixes Perl-Critic/Perl-Critic#451

[wchristian]

This script implies strongly to me that it is indeed a block:

use strict;
use warnings;
use Data::Dumper;
my $r = { 1, 2, 3, 4 };
my %thing = %{
    print qq[blah1\n];
      $r;
};
print Dumper( \%thing );

__END__
blah1
$VAR1 = {
          '1' => 2,
          '3' => 4
        };

Edit: Though amusingly Perl::Tidy gets a little confused by it.

Editedit: And thanks for the confirmations, good to know. :)

[thaljef]

You're right, it is a block. One of my Perl::Critic test cases may be bogus. I'll investigate more tomorrow.

[wchristian]

Neat. Enlightenment all around. :D

[epa]

The title of this bug is "release 1.120" but did you mean 1.220?

[wchristian]

Doh, you're right. Early typo. :)

[thaljef]

@wchristian Perl-Critic is good to go with this release of PPI.

Thanks for all your effort to keep this vital project moving forward.

[genehack]

+1 on my PPI-using module still passing all tests.

[kevindawson]

ok here :) info. follows

$ cpanm https://cpan.metacpan.org/authors/id/M/MI/MITHALDU/PPI-1.219_001.tar.gz
--> Working on https://cpan.metacpan.org/authors/id/M/MI/MITHALDU/PPI-1.219_001.tar.gz
Fetching https://cpan.metacpan.org/authors/id/M/MI/MITHALDU/PPI-1.219_001.tar.gz ... OK
Configuring PPI-1.219_001 ... OK
Building and testing PPI-1.219_001 ... OK
Successfully installed PPI-1.219_001
1 distribution installed

$ prove -l
t/autoprereq.t .................. ok     
t/scan-module.t ................. ok   
t/scan_prereqs.t ................ ok   
t/scanner-eval.t ................ ok     
t/scanner-test_requires.t ....... ok    
t/scanner-use_module.t .......... ok    
t/scanner-useok.t ............... ok    
t/scanner-version-assertions.t .. ok    
All tests successful.
Files=8, Tests=558,  6 wallclock secs ( 0.17 usr  0.01 sys +  5.17 cusr  0.24 csys =  5.59 CPU)
Result: PASS

$ perl -v
This is perl 5, version 18, subversion 2 (v5.18.2) built for i686-linux-thread-multi

[moregan]

I've been travelling and unable to chip in. Thank you, @wchristian , for your continuing efforts.

[wchristian]

@moregan Sure thing, you did plenty already. :)

If you find a little time, it would help me if you could remove some of the merged branches from your repo.

[wchristian]

1.220 is on CPAN, now to close all the related issues. :)