autocorrect hangs for fetchers.rb in Ruby cassandra driver

[stamhankar999]

rubocop -a -d cluster/schema/fetchers.rb
For /Users/sandeeptamhankar/ruby-driver/lib/cassandra: configuration from /Users/sandeeptamhankar/ruby-driver/.rubocop.yml
Inheriting configuration from /Users/sandeeptamhankar/ruby-driver/.rubocop_todo.yml
Default configuration from /Users/sandeeptamhankar/.rbenv/versions/1.9.3-p551/lib/ruby/gems/1.9.1/gems/rubocop-0.36.0/config/default.yml
Inheriting configuration from /Users/sandeeptamhankar/.rbenv/versions/1.9.3-p551/lib/ruby/gems/1.9.1/gems/rubocop-0.36.0/config/enabled.yml
Inheriting configuration from /Users/sandeeptamhankar/.rbenv/versions/1.9.3-p551/lib/ruby/gems/1.9.1/gems/rubocop-0.36.0/config/disabled.yml
Inspecting 1 file
Scanning /Users/sandeeptamhankar/ruby-driver/lib/cassandra/cluster/schema/fetchers.rb

CPU is pegged, so I suspect a tight loop. Clone the ruby-driver repo (https://github.com/datastax/ruby-driver.git) and try this and you should see the same behavior. NOTE: without auto-correct, rubocop does terminate.

[alexdowad]

RC's autocorrect is incredibly slow in some cases. I am working on speeding it up.

Let's see how it works for you after that.

[stamhankar999]

Well, just to be clear, the autocorrect was working through several of my files and then it got stuck on fetchers.rb for many minutes. To narrow things down, I reverted client.rb (which is about 900+ lines of code) and ran rubocop on just that file with auto-correct, and it finished in about 15 seconds. Do the same with fetchers.rb (which is about 1100+ lines), and it sits for minutes before I abort it.

So I don't think this is just a performance issue. Now, fetchers.rb is a fairly complex file, so it may be slower because of that, but orders of magnitude slower sounds very suspicious.

[jonas054]

When I run auto-correction on that file I get the following error after about 90 seconds:

/home/jonas/.rvm/gems/ruby-2.3.0/gems/rubocop-0.36.0/lib/rubocop/runner.rb:184:in `check_for_infinite_loop': Infinite loop detected in /home/jonas/dev/ruby-driver/lib/cassandra/cluster/schema/fetchers.rb. (RuboCop::Runner::InfiniteCorrectionLoop)
    from /home/jonas/.rvm/gems/ruby-2.3.0/gems/rubocop-0.36.0/lib/rubocop/runner.rb:159:in `block in do_inspection_loop'
    from /home/jonas/.rvm/gems/ruby-2.3.0/gems/rubocop-0.36.0/lib/rubocop/runner.rb:158:in `loop'
    from /home/jonas/.rvm/gems/ruby-2.3.0/gems/rubocop-0.36.0/lib/rubocop/runner.rb:158:in `do_inspection_loop'
    from /home/jonas/.rvm/gems/ruby-2.3.0/gems/rubocop-0.36.0/lib/rubocop/runner.rb:87:in `process_file'
    from /home/jonas/.rvm/gems/ruby-2.3.0/gems/rubocop-0.36.0/lib/rubocop/runner.rb:59:in `block in inspect_files'
    from /home/jonas/.rvm/gems/ruby-2.3.0/gems/rubocop-0.36.0/lib/rubocop/runner.rb:57:in `each'
    from /home/jonas/.rvm/gems/ruby-2.3.0/gems/rubocop-0.36.0/lib/rubocop/runner.rb:57:in `inspect_files'
    from /home/jonas/.rvm/gems/ruby-2.3.0/gems/rubocop-0.36.0/lib/rubocop/runner.rb:35:in `run'
    from /home/jonas/.rvm/gems/ruby-2.3.0/gems/rubocop-0.36.0/lib/rubocop/cli.rb:30:in `run'
    from /home/jonas/.rvm/gems/ruby-2.3.0/gems/rubocop-0.36.0/bin/rubocop:14:in `block in <top (required)>'
    from /home/jonas/.rvm/rubies/ruby-2.3.0/lib/ruby/2.3.0/benchmark.rb:308:in `realtime'
    from /home/jonas/.rvm/gems/ruby-2.3.0/gems/rubocop-0.36.0/bin/rubocop:13:in `<top (required)>'
    from /home/jonas/.rvm/gems/ruby-2.3.0/bin/rubocop:23:in `load'
    from /home/jonas/.rvm/gems/ruby-2.3.0/bin/rubocop:23:in `<main>'
    from /home/jonas/.rvm/gems/ruby-2.3.0/bin/ruby_executable_hooks:15:in `eval'
    from /home/jonas/.rvm/gems/ruby-2.3.0/bin/ruby_executable_hooks:15:in `<main>'

I can try to find the cause of the problem soon.

[jonas054]

I've narrowed it down to

def f
  X.map do |(a,
             b)|
  end
end

It also causes an infinite auto-correction loop.

[alexdowad]

It looks like MultilineBlockLayout and BlockEndNewline are fighting.

[alexdowad]

I've got a fix.

[alexdowad]

I don't have time to write specs right now, but I'll push first.