Modern Perl

To dzil or not to dzil...?

That is the Question!

I've wanted to like Dist::Zilla for a long time since I generally agree with the idea that the developer dependencies and management should be cleanly separated from the user installation experience.  Where I've felt let down by dzil in the past is the size of the installation and the fact that it really seems to take over the development process.

However during my last cycle of upgrading and patching modules by other authors who are using DistZilla I found that the process is significantly smoother.  Most of what I like to do I can do the way I always have, or have straightforward dzil commands.  So I'm tempted again!

For years I've tried to mimic the clean separation of developer and installer  using Module::Install, which I have a lot of experience using and in general has been pretty solid.  I minic this with a bit of custom logic that detects developer mode and including the extra bits if so.  Again, this works, but its another 'I've rolled my own' things, which generally I don't like when there are good community accepted and supported solutions available.

So, I'm open to using dzil, just wondering what other people experiences have been.  If I change some of my CPAN stuff to DistZilla, will that help/hurt/neither my contributor base?

So of the thoughts I am having so far:

1. For modules that already are somewhat dependency heavy (already need Moose for example) DistZilla is probably not going to be a big stretch for contributors
2. Conversely, for some of my Plack middlewares where I am trying hard to keep light, I might consider just continuing to do what I currently do
3. For projects in communities where DistZilla is not preferred, I might avoid it for now.

Thanks for suggestions and advice in advance!

Perl 5.14 Stable Release!

Perl 5.14 went gold over the weekend.  I used Perlbrew to easily install it on my Macbook Pro. 

Although I've been testing my modules on the various beta and release candidates, I'm looking forward to checking it all over on this newly minted Perl.  Congrats to the release team!

 

AMD Tapper

Just was glancing through AMD's github repo and CPAN namespace and it seems this is a pretty big code dump of a rather cool and modern coded Perl application.  So, if you are looking for more examples of real Catalyst and DBIC applications in the wild, now we have another example you can go review!

Just goes to show you, Perl is a heavy hitter in a lot of unexpected places.  Props to AMD for opensourcing this, and welcome to the Perl community!

Catalyst PSGI / Plack Trial Version, Fun and Games

There's currently only a handful of blockers left on the psgi Catalyst branch in the repo.  If you use Catalyst now is definitely the time to cut a clean local::lib for you application and test it out!  If you need step by step instructions on how to do that, here's what I do:

1) Create a clean local::lib with Catalyst installed (also install cpanminus to help you later)

curl -L http://cpanmin.us/ | perl - -l ./testcatalystlibs App::local::lib::helper App::cpanminus http://search.cpan.org/CPAN/authors/id/F/FL/FLORA/Catalyst-Runtime-5.89002-TRIAL.tar.gz

2) Tell Perl about you new local::lib

./testcatalystlibs/bin/localenv bash

3) Install the rest of your dependencies

cpanm Makefile.PL

And start testing!

Now, for the fun and games part.  Since Catalyst is now using Plack natively, that means you can do interesting things like add middleware directly in your Catalyst application subclass, similarly to theway you aggregate Catalyst Plugins.

around 'psgi_app', sub {
  my ($orig, $self, @args) = @_;
  my $psgi = $self->$orig(@args);
  builder {
    mount '/static', builder {
      Plack::App::Directory
        ->new(root => $self->path_to('share','static'))
        ->to_app;
    },
    mount '/', builder {
      enable_if {
        $self->debug;
      } 'Debug', panels => $self->config->{debug_panels};
      $psgi;
    }
  };
};

This makes your app.psgi super clean:

#!/usr/bin/env perl

use strict;
use warnings FATAL => 'all';
use MyApp::Web;

MyApp::Web->psgi_app;

Personally I find it more natural to add middleware this way, since in my head that middleware is part of the application.  Perhaps in the future we can have a cleaner approach than simple wrapping the psgi_app method, or making it easier to use different middlewares in different environments.

Maximize Promotional Value of the Impending Perl 5.14 Gold Release?

As I suspect many of you have, I've been trying to test out all the important CPAN projects I use on the lastest Perl 5.14 release candidate.  And while I was doing so I was thinking, "It's really great that Perl is back on a solid release track, with minor and major versions scheduled and discussed."  It leads to great press and gives enterprises at least the feeling that Perl is alive and well.

And then I got to thinking, even though its awesome that Perl core language is in such good hands, truth be told I am equally affected by the great development and advances seen in some of our more important  CPAN projects such as DBIC, Moose, Catalyst, Plack.  Unfortunately Perl branding is currently such that when a new version of Moose or DBIC comes out, it really doesn't get picked up by the relevent media sources in the same way that new versions of Perl do.  So I started to wonder how fair would it be to the members of the Perl core hackers if the press release about Perl 5.14 going gold mentioned some key advances for some of our more important Modern Perl CPAN projects since Perl 5.12 was gold?  We could even mention some other not strictly technical advancements such as the Modern Perl book (or other important book releases) stats about CPAN testers exceeding 11 million test reports, etc.

For better or worse, over the past few years I've noticed that while Perl major version releases get some press, other important projects really don't.  Although I don't want to unfairly ride on the massive effort of the Perl core team to get 5.14 out the door, perhaps for the sake of advocacy and general information spreading, we should consider bundling the 5.14 announcement with a choice addition of other major accomplishments from the Perl community?