roles Archives

Web Scraping with Zydeco

So I like to keep local copies of my blog posts as Atom entries, but noticed yesterday that I had a few gaps in my collection. The Atom feeds offered by only have the most recent articles though, so I decided to write a quick script to scrape the posts. Luckily, I managed to get a table containing the URLs for each post I needed, so I didn't need to bother with following links to find the pages; I just needed to grab the content from them.

I thought some people might find the code interesting especially for its use of lazy attributes. This is one of those "it only needs to be used once, so making the code maintainable isn't important" kinds of projects, do bear that in mind. I've cleaned up the whitespace and added comments for this blog post, but other than that, it's just a quickly hacked together script.

MooX::Pression — now much faster

The test suite for MooX::Pression used to run in 79 seconds on my laptop. Now it's at 10 seconds.

And no, I didn't cut out any tests — I switched from using Keyword::Declare to a combination of Keyword::Simple and PPR. (Keyword::Declare is a wrapper around Keyword::Simple and PPR, but I found out by using them directly, I could massively improve compile-time speed.)

MooX::Pression allows you to build classes and roles with multimethods, types, method signatures, and sweet, sweet, sugary syntax…

Announcing MooX::Press

MooX::Press is a quick way of building a bunch of Moo roles and classes in one use statement.

The most basic example would be:

  package MyApp {
    use MooX::Press class => ['Foo', 'Bar'];
  my $thing1 = MyApp::Foo->new();
  my $thing2 = MyApp->new_foo();   # alternative constructor

But do-nothing classes with a constructor and nothing else aren't very exciting. Let's define a class with some subclasses which have attributes and roles and methods and stuff.

About Toby Inkster

user-pic I'm tobyink on CPAN, IRC and PerlMonks.