Shadowcat Systems Limited

sufficiently advanced technology

Telephone 01524 842155
Email: info(at)shadowcat.co.uk

Sat Jul 14 16:00:00 2012

Community Diary - 2012-07-14

I guess "something vaguely like a week" was wildly optimistic, though while I'd expected my usual degree of procrastination over blogging I hadn't really expected to break my hip. In any case, onwards.

French Perl Workshop

I spent June 29-30 at the French Perl Workshop in Strasbourg.

It was awesome.

Most of the talks were in french, but as usual for me I spent the time I wasn't talking (or frantically trying to finish slides) walking the hallways talking to people, so it made very little difference to me :)

The socialisation was great, we started off at a nice local bar and then when it closed moved on to another nearby one recommended by one of the staff at the first bar (whoever you are, you were awesome, thanks!). I got to sample several local beers, but the highlight of my trip was a belgian beer called Black Mamba. It was ... like what would happen if you crossbred a good stout and a triple hop American IPA. It tried to eat my face. I loved it.

The French perl community seems to be doing all sorts of fascinating things and contain all sorts of interesting people - I spoke to a few relative newcomers to the community and had great fun doing so, notably a guy who'd been an actor and director up to about two years ago who I think will be hugely valuable as a human to geek translation layer once he stops worrying about not being good enough.

I gave two talks - the first about Moo, covering its now complete interoperability with Moose and status as the right answer to the question "I'm not sure if I can use Moose for this?", the second entitled Postmodern Deconstructionism, trying to explain why I think Plack has been such a success and how we can take those ideas into other spaces to enable more and better sharing of code. The latter got scheduled as the keynote, although I only found this out at the last minute so I'm not entirely convinced it was quite keynote quality; hopefully we'll get the video and you can draw your own conclusions.

The conference dinner was excellent, unsurprisingly I guess for a workshop in France - at one point my tablemates apologised for having been talking in French for the previous few minutes (the only thing I can manage to say in French these days is "je suis desolee mais j'ai oubliee le plus part de la langue francaise que j'ai etudie") and I replied by telling them I hadn't even noticed because I'd been too busy enjoying the food!

The inestimable BooK brought Chartreuse, resulting in late night drinking in a park and by the canal; interesting how conversations spread out differently outdoors than inside a pub. Josette from O'Reilly was there and delightful as always, and Liz and Wendy (inevitably) showed up and were lovely company. If you're not sure who these people are, I strongly recommend trying to meet them at YAPC::EU - a short description in a blog post can't do any of them justice, really.

Happy perl hacking ... next installment in something vaguely like a week.

The toolchain list and lists.perl.org

There was a little confusion over the past week over which mailing list we should be using to discuss toolchain work - things like Module::Build, ExtUtils::MakeMaker, CPAN.pm and associated code - to clear things up, the right place is the cpan-workers list hosted at the perl.org NOC. We're only just getting back into using it though so for quicker response #toolchain on irc.perl.org is still the place to go.

A couple of people had trouble getting their subscription to do what they wanted due to ezmlm's defaulting to the envelope sender of a subscription request rather than the From: header, unless you use a slightly different syntax to subscribe - which they'd not been able to find. A quick chat with Robert Spier of the perl.org NOC later and you'll see that the above page now has an 'advanced instructions' section which explains this - so thanks very much to Robert, and to everybody else please keep telling me if you run into trouble with perl.org lists and I'll see what we can do to make your experience smoother still.

A brief followup on Dancer 2

Last time I asked for people to help us test out the new code - if you aren't sure how to get it, the answer is from the PerlDancer github organisation.

It's still very rough around the edges and if you try and switch a Dancer 1 app across it will almost certainly break - but every breakage you can turn into a failing test against Dancer 2 is a breakage we can fix much more easily, so I'd implore you to at least give it a go. After all, this is how CPAN happens.

CPAN roundup

I've shipped a bunch of stuff to CPAN this week, mostly other people's work.

App::FatPacker got a new release with a refactored trace function and documentation for the module that provides it.

Role::Tiny got a test fix (a first time contributor wrote me an excellent test but accidentally used strictures.pm in it, which isn't permitted for Role::Tiny - unlike pretty much everything else I've worked on recently) - and a fix for runtime role application to fix a bug in does() calls (thanks to frew of the DBIx::Class team for this one).

Moo was actually released just before FPW, but the notable change for me is that we're now sufficiently good at interoperating with Moose that Stevan gave me permission to make #moose the channel for Moo user questions - although development is still discussed in #web-simple.

YAML got three releases, the first two to eliminate warnings under 5.16 (thanks very much to our pumpking rjbs for helping me track these down) and the third one to fix a stupid mistake I'd made doing the release the first two time (thanks very much to ingy for his infinite patience figuring out what had gone wrong :).

strictures got two new releases, the first switching the "am I part of a checkout?" extra testing to only fire on modules inside your checkout rather than strictures-using modules outside of it as well, and the second expanding the checkout discovery code to handle Dist::Zilla correctly and memoizing the checks so that it only has to check the filesystem once to achieve this (thanks to Karen Etheridge for implementing both of these and shipping the release afterwards!)

Finally, I wrote some skeleton documentation for Object::Remote and shipped an initial release. It still needs a lot more explanation, although my YAPC::NA lightning talk "The END of Everything" isn't a bad surrogate to give you the general idea. Thanks to phaylon (Robert Sedlacek) of the Shadowcat team for helping me build it, and thanks to Socialflow for sponsoring part of the development work.

Why mst has to write some perl6

During the French Perl Workshop, the organisers mentioned that they hadn't got quite as much sponsorship as they'd expected and said they'd be very greatful if people bought a T-shirt.

Unfortunately, the T-shirts were white.

So instead, I did something slightly different - before the keynote, I made an offer to the conference: For every 40 EUR they donated by midnight french time that night, I would spend an hour working on the perl6 code of jnthn (Jonathan Worthington) and masak (Carl Masak)'s choice.

You guys are AWESOME. We got 440 euros towards this year's (and next year's) FPW budget.

Which means that one weekend soon, I'll be spending eleven hours writing some perl6; the current front runner project idea is an Object::Remote connector for rakudo so you can do:

use Object::Remote;
use Object::Rmeote::Connector::Rakudo;

my $p6obj = Some::Perl6::Class->new::on(v6);

This weekend, however, is my first weekend out of hospital so I'm going to spend it catching up with other things - so, with that:

-- mst, out.