Jeweler maintenance has now shifted to Fred Mitchell. I am now maintaining both Jeweler and its fork, juwelier. I will keep Jeweler at least functional with the latest Ruby releases, but put new features in Juwelier. Your input on this is more than welcome.
Jeweler provides the noble ruby developer with two primary features:
- a library for managing and releasing RubyGem projects
- a scaffold generator for starting new RubyGem projects
PLEASE NOTE that if you are starting afresh, please use the successor Juwelier I (Fred Mitchell, flajann2) will be maintaining both Jeweler and Juwelier, but will be adding new features to Juwelier, and eventually "merge" this one into Juwelier after some namespace issues are dealt with.
Use RubyGems to install the heck out of jeweler to get started:
$ gem install jeweler
With jeweler installed, you can use the jeweler command to generate a new project. For the most basic use, just give it a name:
$ jeweler hello-gem
This requires some Git configuration (like name, email, GitHub account, etc), but jeweler will prompt along the way.
Your new hello-gem gem is ready in the hello-gem directory. Take a peek, and you'll see several files and directories
Rakefilesetup for jeweler, running tests, generating documentation, and releasing to rubygems.orgREADME.rdocwith contribution guidelines and copyright info crediting youLICENSEwith the MIT licensed crediting youGemfilewith development dependencies filled inlib/hello-gem.rbwaiting for you to codetest/containing a (failing) shoulda test suite shoulda
The jeweler command supports a lot of options. Mostly, they are for generating baked in support for this test framework, or that.
Check out jeweler --help for the most up to date options.
Beyond just editing source code, you'll be interacting with your gem using rake a lot. To see all the tasks available with a brief description, you can run:
$ rake -T
You'll need a version before you can start installing your gem locally. The easiest way is with the version:write Rake task. Let's imagine you start with 0.1.0
$ rake version:write MAJOR=0 MINOR=1 PATCH=0
You can now go forth and develop, now that there's an initial version defined. Eventually, you should install and test the gem:
$ rake install
The install rake task builds the gem and gem installs it. You're all set if you're using RVM, but you may need to run it with sudo if you have a system-installed ruby:
$ sudo rake install
At last, it's time to ship it! Make sure you have everything committed and pushed, then go wild:
$ rake release
This will automatically:
- Generate
hello-gem.gemspecand commit it - Use
gitto tagv0.1.0and push it - Build
hello-gem-0.1.0.gemand push it to rubygems.org
rake release accepts REMOTE(default: origin), LOCAL_BRANCH(default: master), REMOTE_BRANCH(default: master) and BRANCH(default: master)as options.
$ rake release REMOTE=upstream LOCAL_BRANCH=critical-security-fix REMOTE_BRANCH=v3
This will tag and push the commits on your local branch named critical-security-fix to branch named v3 in remote named upstream (if you have commit rights
on upstream) and release the gem.
$ rake release BRANCH=v3
If both remote and local branches are the same, use BRANCH option to simplify.
This will tag and push the commits on your local branch named v3 to branch named v3 in remote named origin (if you have commit rights
on origin) and release the gem.
It feels good to release code. Do it, do it often. But before that, bump the version. Then release it. There's a few ways to update the version:
# version:write like before
$ rake version:write MAJOR=0 MINOR=3 PATCH=0
# bump just major, ie 0.1.0 -> 1.0.0
$ rake version:bump:major
# bump just minor, ie 0.1.0 -> 0.2.0
$ rake version:bump:minor
# bump just patch, ie 0.1.0 -> 0.1.1
$ rake version:bump:patch
Then it's the same release we used before:
$ rake release
If you've been following along so far, your gem is just a blank slate. You're going to need to make it colorful and full of metadata.
You can customize your gem by updating your Rakefile. With a newly generated project, it will look something like this:
require 'jeweler'
Jeweler::Tasks.new do |gem|
# gem is a Gem::Specification... see http://guides.rubygems.org/specification-reference/ for more options
gem.name = "whatwhatwhat"
gem.summary = %Q{TODO: one-line summary of your gem}
gem.description = %Q{TODO: longer description of your gem}
gem.email = "josh@technicalpickles.com"
gem.homepage = "http://github.com/technicalpickles/whatwhatwhat"
gem.authors = ["Joshua Nichols"]
end
Jeweler::RubygemsDotOrgTasks.new
It's crucial to understand the gem object is just a Gem::Specification. You can read up about it at guides.rubygems.org/specification-reference. This is the most basic way of specifying a gem, Jeweler-managed or not. Jeweler just exposes this to you, in addition to providing some reasonable defaults, which we'll explore now.
gem.name = "whatwhatwhat"
Every gem has a name. Among other things, the gem name is how you are able to gem install it. Reference
gem.summary = %Q{TODO: one-line summary of your gem}
This is a one line summary of your gem. This is displayed, for example, when you use gem list --details or view it on rubygems.org.
gem.description = %Q{TODO: longer description of your gem}
Description is a longer description. Scholars ascertain that knowledge of where the description is used was lost centuries ago.
gem.email = "josh@technicalpickles.com"
This should be a way to get a hold of you regarding the gem.
gem.homepage = "http://github.com/technicalpickles/whatwhatwhat"
The homepage should have more information about your gem. The jeweler generator guesses this based on the assumption your code lives on GitHub, using your Git configuration to find your GitHub username. This is displayed by gem list --details and on rubygems.org.
gem.authors = ["Joshua Nichols"]
Hey, this is you, the author (or me in this case). The jeweler generator also guesses this from your Git configuration. This is displayed by gem list --details and on rubygems.org.
The quickest way to add more files is to git add them. Jeweler uses your Git repository to populate your gem's files by including added and committed and excluding .gitignored. In most cases, this is reasonable enough.
If you need to tweak the files, that's cool. Jeweler populates gem.files as a Rake::FileList. It's like a normal array, except you can include and exclude file globs:
gem.files.exclude 'tmp' # exclude temporary directory
gem.files.include 'lib/foo/bar.rb' # explicitly include lib/foo/bar.rb
If that's not enough, you can just set gem.files outright
gem.files = Dir.glob('lib/**/*.rb')
Dependencies let you define other gems that your gem needs to function. gem install your-gem will install your-gem's dependencies along with it, and when you use your-gem in an application, the dependencies will be made available. Use gem.add_dependency to register them. Reference
gem.add_dependency 'nokogiri'
This will ensure a version of nokogiri is installed, but it doesn't require anything more than that. You can provide extra args to be more specific:
gem.add_dependency 'nokogiri', '= 1.2.1' # exactly version 1.2.1
gem.add_dependency 'nokogiri', '>= 1.2.1' # greater than or equal to 1.2.1, ie, 1.2.1, 1.2.2, 1.3.0, 2.0.0, etc
gem.add_dependency 'nokogiri', '>= 1.2.1', '< 1.3.0' # greater than or equal to 1.2.1, but less than 1.3.0
gem.add_dependency 'nokogiri', '~> 1.2.1' # same thing, but more concise
When specifying which version is required, there's a bit of the condunrum. You want to allow the most versions possible, but you want to be sure they are compatible. Using >= 1.2.1 is fine most of the time, except until the point that 2.0.0 comes out and totally breaks backwards the API. That's when it's good to use ~> 1.2.1, which requires any version in the 1.2 family, starting with 1.2.1.
Executables let your gem install shell commands. Just put any executable scripts in the bin/ directory, make sure they are added using git, and Jeweler will take care of the rest.
When you need more finely grained control over it, you can set it yourself:
gem.executables = ['foo'] # note, it's the file name relative to `bin/`, not the project root
We discussed earlier how to bump the version. The rake tasks are really just convience methods for manipulating the VERSION file. It just contains a version string, like 1.2.3.
VERSION is a convention used by Jeweler, and is used to populate gem.version. You can actually set this yourself, and Jeweler won't try to override it:
gem.version = '1.2.3'
A common pattern is to have this in a version constant in your library. This is convenient, because users of the library can query the version they are using at runtime.
# in lib/foo/version.rb
class Foo
module Version
MAJOR = 1
MINOR = 2
PATCH = 3
BUILD = 'pre3'
STRING = [MAJOR, MINOR, PATCH, BUILD].compact.join('.')
end
end
# in Rakefile
require 'jeweler'
require './lib/foo/version.rb'
Jeweler::Tasks.new do |gem|
# snip
gem.version = Foo::Version::STRING
end
Jeweler lives inside of Rake. As a result, they are dear friends. But, that friendship doesn't interfere with typical Rake operations.
That means you can define your own namespaces, tasks, or use third party Rake libraries without cause for concern.
- Check out the latest master to make sure the feature hasn't been implemented or the bug hasn't been fixed yet
- Ask on the mailing list for feedback on your proposal, to see if somebody else has done it.
- Check out the issue tracker to make sure someone already hasn't requested it and/or contributed it
- Fork the project
- Start a feature/bugfix branch
- Commit and push until you are happy with your contribution
- Make sure to add tests for the feature/bugfix. This is important so I don't break it in a future version unintentionally.
- Please try not to mess with the Rakefile, version, or history. If you want to have your own version, or is otherwise necessary, that is fine, but please isolate it to its own commit so I can cherry-pick around it.
Copyright (c) 2008-2010 Josh Nichols. Copyright (c) 2016 Fred Mitchell. See LICENSE for details.