Contents

require 'rdoc/task'
require 'rake/contrib/sshpublisher'

task :default => :test

# We test this way because of what this library does.
# The tests wrap and load C++ wrapper code constantly.
# When running all the tests at once, we very quickly run 
# into problems where Rice crashes because 
# a given C++ class is already wrapped, or glibc doesn't like our 
# unorthodox handling of it's pieces. So we need to run the
# tests individually
desc "Run the tests"
task :test do
  require 'rbconfig'
  FileList["test/*_test.rb"].each do |file|
    # To allow multiple ruby installs (like a multiruby test suite), I need to get
    # the exact ruby binary that's linked to the ruby running the Rakefile. Just saying
    # "ruby" will find the system's installed ruby and be worthless
    ruby = File.join(Config::CONFIG["bindir"], Config::CONFIG["RUBY_INSTALL_NAME"])
    sh "#{ruby} -S rspec -Itest #{file}"
  end
end

Rake::RDocTask.new do |rd|
  rd.main = "README"
  rd.rdoc_files.include("README", "lib/**/*.rb")
  rd.rdoc_files.exclude("**/jamis.rb")
  rd.template = File.expand_path(File.dirname(__FILE__) + "/lib/jamis.rb")
  rd.options << '--line-numbers' << '--inline-source'
end

RUBYFORGE_USERNAME = "jameskilton"
PROJECT_WEB_PATH = "/var/www/gforge-projects/rbplusplus"

namespace :web do
  desc "Build website"
  task :build => :rdoc do |t|
    unless File.directory?("publish")
      mkdir "publish"
    end

    sh "jekyll --pygment website publish/"
    sh "cp -r html/* publish/rbplusplus/"
  end

  desc "Update the website" 
  task :upload => "web:build"  do |t|
    Rake::SshDirPublisher.new("#{RUBYFORGE_USERNAME}@rubyforge.org", PROJECT_WEB_PATH, "publish").upload
  end

  desc "Clean up generated website files" 
  task :clean do
    rm_rf "publish"
  end
end

Version data entries

1 entries across 1 versions & 1 rubygems

Version	Path
rbplusplus-1.0.3	Rakefile