Default task for namespace in Rake

Given something like:

namespace :my_tasks do
  task :foo do
    do_something
  end

  task :bar do
    do_something_else
  end

  task :all => [:foo, :bar]
end

How do I make :all be the default task, so that running rake my_tasks will call it (instead of having to call rake my_tasks:all)?

Place it outside the namespace like this:

namespace :my_tasks do
  task :foo do
    do_something
  end

  task :bar do
    do_something_else
  end

end

task :all => ["my_tasks:foo", "my_tasks:bar"]

Also... if your tasks require arguments then:

namespace :my_tasks do
  task :foo, :arg1, :arg2 do |t, args|
    do_something
  end

  task :bar, :arg1, :arg2  do |t, args|
    do_something_else
  end

end

task :my_tasks, :arg1, :arg2 do |t, args|
  Rake::Task["my_tasks:foo"].invoke( args.arg1, args.arg2 )
  Rake::Task["my_tasks:bar"].invoke( args.arg1, args.arg2 )
end

Notice how in the 2nd example you can call the task the same name as the namespace, ie 'my_tasks'

Not very intuitive, but you can have a namespace and a task that have the same name, and that effectively gives you what you want. For instance

namespace :my_task do
  task :foo do
    do_foo
  end
  task :bar do
    do_bar
  end
end

task :my_task do
  Rake::Task['my_task:foo'].invoke
  Rake::Task['my_task:bar'].invoke
end

Now you can run commands like,

rake my_task:foo

and

rake my_task

I suggest you to use this if you have lots of tasks in the namespace.

task :my_tasks do
  Rake.application.in_namespace(:my_tasks){|namespace| namespace.tasks.each(&:invoke)}
end

And then you can run all tasks in the namespace by:

rake my_tasks

With this, you don't need to worry to change your :all task when you add new tasks into that namespace.

I use this Rakefile for cucumber:

require 'cucumber'
require 'cucumber/rake/task'

namespace :features do
  Cucumber::Rake::Task.new(:fast) do |t|
    t.profile = 'fast'
  end

  Cucumber::Rake::Task.new(:slow) do |t|
    t.profile = 'slow'
  end

  task :ci => [:fast, :slow]
end

task :default => "features:ci"

Then if I type just:

rake

It runs the default task, which runs both fast and slow tests.

I learned this from Cheezy's blog.

The way I'm reading obvio171's question is that he is asking¹⁾ for a systematic way to invoke a certain task in a namespace by invoking the namespace as a task.

I've frequently encountered the same need. I like to logically group tasks into namespaces. Often that grouping resembles a hierarchy. Hence the desire to invoke the group makes very much sense to me.

Here's my take:

module Rake::DSL
  def group(name, &block)
    ns = namespace name, &block
    default = ns[:default]
    task name => "#{name}:default" if default
    ns
  end
end

group :foo do
  task :foo1 do |t| puts t.name end
  task :foo2 do |t| puts t.name end
  task :default => [:foo1, :foo2]
end

task :default => :foo

---

¹⁾...or was asking, years ago. Nonetheless a still interesting question.

Add the following task outside of the namespace:

desc "Run all my tasks"
task :my_tasks => ["my_tasks:all"]

Keep in mind, that you can have a task with the same name as the namespace.

And hier a bigger example, that shows, how you can make use of tasks, which have the same name as the namespace, even when nesting namespaces:

namespace :job1 do
  task :do_something1 do
        puts "job1:do_something1"
    end

  task :do_something2 do
        puts "job1:do_something2"
    end
  task :all => [:do_something1, :do_something2]
end

desc "Job 1"
task :job1 => ["job1:all"]

# You do not need the "all"-task, but it might be handier to have one.
namespace :job2 do
  task :do_something1 do
        puts "job2:do_something1"
    end

  task :do_something2 do
        puts "job2:do_something2"
    end
end

desc "Job 2"
task :job2 => ["job2:do_something1", "job2:do_something2"]

namespace :superjob do
    namespace :job1 do
        task :do_something1 do
            puts "superjob:job1:do_something1"
        end

        task :do_something2 do
            puts "superjob:job1:do_something2"
        end
    end

    desc "Job 1 in Superjob"
    task :job1 => ["job1:do_something1", "job1:do_something2"]

    namespace :job2 do
        task :do_something1 do
            puts "superjob:job2:do_something1"
        end

        task :do_something2 do
            puts "superjob:job2:do_something2"
        end
    end

    desc "Job 2 in Superjob"
    task :job2 => ["job2:do_something1", "job2:do_something2"]
end

desc "My Super Job"
task :superjob => ["superjob:job1", "superjob:job2"]

# Do them all just by calling "$ rake"
task :default => [:job1, :job2, :superjob]

Just copy it and try it out.

Based on Rocky's solution https://stackoverflow.com/questions/1579231/default-task-for-namespace-in-rake/17006916#17006916

And this dexter's answer https://stackoverflow.com/questions/7363798/is-there-a-way-to-know-the-current-rake-task/7365507#7365507

namespace :root do
  namespace :foo do
  end

  namespace :target do
    task :all do |task_all|
      Rake.application.in_namespace(task_all.scope.path) do |ns|
        ns.tasks.each { |task| task.invoke unless task.name == task_all.name } 
      end
    end

    task :one do
    end

    task :another do
    end
  end
end

Combining Szymon Lipiński's and Shyam Habarakada's answers, here is what I think is the most idiomatic and consise answer:

namespace :my_tasks do
  task :foo do
    do_something
  end

  task :bar do
    do_something_else
  end

end

task :my_tasks => ["my_tasks:foo", "my_tasks:bar"]

allows you to do rake my_tasks while avoiding cumbersome invocation of the subtasks.