Ruby core extensions repository (not a gem!)
In the core of apple, there are apple seeds. In your core_ext.rb should
be RubySeeds.
It is not unusual for Ruby developers to have own and favourite set of
ruby core classes extensions (typically lying somewhere in project tree in
core_ext.rb
), dragged from project to project, growing, becoming cluttered
yet beloved.
Other developers rely on some external thing like ActiveSupport or Ruby
Facets to extend their core classes. And if it is not there, its nowhere.
RubySeeds is different.
RubySeeds is not a gem. On my strong opinion, there is no good in
mixing together everything somebody can need in core classes. The same
way, there’s no good to make a small gem of each useful idea.
RubySeeds is just an easily navigable list of code snippets
with clean and simple explanations. Those snippets are thought to be just
droppable in your projects lib/core_ext.rb
(with possible renamings
and improvements).
Snippets selected for RubySeeds posess those common qualities:
There’s one huge, yet (supposedly) easily-navigable page with all seeds
described: RubySeeds.md.
(In future will be separate page with collapsible code.)
You just browse it, select what you like, expand Code section and
copy-paste it into your core_ext.rb
or wherever you want.
Since version 2.0, Ruby has refinements
which allows core (and other) classes to be extended not globally, but in
context of some module.
Before refinements:
# in core_ext.rb
class Hash
def symbolize_keys
# ...
end
end
# everywhere in code
require 'core_ext'
# now ANYWHERE Hash#symbolize_keys refers to your implementation... unless
# you require something else, redefining it.
With refinements:
# in core_ext.rb
module HashSymbolize
refine Hash do
def symbolize_keys
# ...
end
end
end
# in some concrete place:
require 'core_ext'
class MyClass
using HashSymbolize
# now Hash#symbolize_keys can be used inside your class only
end
Somebody can say that adding functionality to core classess (even through
refinements) is a bad practice.
It may be (or maybe not) true for large
enterprise-y systems with tens of juniors fixing your code here and there
and some contractor just running through it wildly…
Yet for many small-to-medium data-processing tasks you just want to be
comfortable with code, having it clean and looking like chain of data
processing calls. So, you just want
load(something).do_this.and_that.and_shit
instead of multiple
My::Deeply::Nested::Namespace.very_concrete_processing_method_name(data)
calls. And you have it.
Also, take a look at Hal Fultons post about “monkey patching”.
And here is my own latest post on the topic.
There was rubyseeds organization on github,
something about learning Ruby, I guess?..
Current project have no relation to this group, and, as
the site seems to be dead, and
I’ve already thinked out and loved the name when I found the organization,
I allowed myself to keep the name.
If you are like me (= understand, while we sometimes add features to our
beloved language to make code clear and rational), you may like also those
libraries:
Time
tasks like “next day”, “prev day”, “todays midnight”,Just usual fork-change-push-make pull request process. Note (if you haven’t
already), that RubySeeds is highly opionated set of extensions, so,
there’s no guarantee that pull requests with new features will be accepted.
NB: never try to update RubySeeds.md file – it is built automatically from
code in lib/**/*.rb
, with rake build
.
Those are code snippets barely dozen lines length. Let’s say “public
domain”, ok?