Rake异常管理

Question

考虑一下这个Rake任务：

namespace :foo do
  task :bar do
    begin
      raise 'foo'
    rescue RuntimeError => ex
      raise ex.class, 'bar', ex.backtrace
    end
  end
end

其结果如下：

rake aborted!
bar
/home/vagrant/proj/lib/tasks/foo.rake:52:in `block (2 levels) in <top (required)>'
/home/vagrant/.gem/ruby/2.3.0/gems/activesupport-4.2.5/lib/active_support/dependencies.rb:268:in `load'
/home/vagrant/.gem/ruby/2.3.0/gems/activesupport-4.2.5/lib/active_support/dependencies.rb:268:in `block in load'
/home/vagrant/.gem/ruby/2.3.0/gems/activesupport-4.2.5/lib/active_support/dependencies.rb:240:in `load_dependency'
/home/vagrant/.gem/ruby/2.3.0/gems/activesupport-4.2.5/lib/active_support/dependencies.rb:268:in `load'
-e:1:in `<main>'
foo
/home/vagrant/proj/lib/tasks/foo.rake:52:in `block (2 levels) in <top (required)>'
/home/vagrant/.gem/ruby/2.3.0/gems/activesupport-4.2.5/lib/active_support/dependencies.rb:268:in `load'
/home/vagrant/.gem/ruby/2.3.0/gems/activesupport-4.2.5/lib/active_support/dependencies.rb:268:in `block in load'
/home/vagrant/.gem/ruby/2.3.0/gems/activesupport-4.2.5/lib/active_support/dependencies.rb:240:in `load_dependency'
/home/vagrant/.gem/ruby/2.3.0/gems/activesupport-4.2.5/lib/active_support/dependencies.rb:268:in `load'
-e:1:in `<main>'
Tasks: TOP => foo:bar
(See full trace by running task with --trace)

如您所见，既有原始异常(foo)，也有新异常(bar)。为什么会这样呢？我原以为foo的例外，已经得到了适当的拯救，不会出现在这里。

Answer 1

没有两个原始异常(foo)。

Rakefile：

namespace :foo do
  task :bar do
    begin
      raise "message from foo"
    rescue RuntimeError => ex
      puts "#{ex.message}. foo is rescued"             # <= 1st
      raise ex.class, 'message from bar', ex.backtrace # <= 2nd
    end
  end
end

输出：

message from foo. foo is rescued # <= 1st
rake aborted!
message from bar # <= 2nd
/home/vagrant/Rakefile:4:in `block (2 levels) in <top (required)>'
message from foo # <= from backtrace
/home/vagrant/Rakefile:4:in `block (2 levels) in <top (required)>'
Tasks: TOP => foo:bar
(See full trace by running task with --trace)