Pippi

Pippi is a utility for finding suboptimal Ruby class API usage.

Consider this little array:

[1, 2, 3]

Now suppose we want to find the first element in that array that's greater than one. We can use Array#select, which returns another Array, and then use Array#first:

[1, 2, 3].select { |x| x > 1 }.first

Of course that's terribly inefficient. Since we only need one element we don't need to select all elements that match the predicate. We should use Array#detect instead:

[1, 2, 3].detect { |x| x > 1 }

A change like this is a small optimization, but they can add up. More importantly, they communicate the intent of the programmer; the use of Array#detect makes it clear that we're just looking for the first item to match the predicate.

This sort of thing can be be found during a code review, or maybe when you're just poking around the code. But why not have a tool find it instead? Thus, pippi. Pippi observes code while it's running - by hooking into your test suite execution - and reports misuse of class-level APIs.

There are many nifty Ruby static analysis tools - flay, reek, flog, etc. This is not like those. It doesn't parse source code; it doesn't examine an abstract syntax tree or even sequences of MRI instructions. So it cannot find the types of issues that those tools can find. Instead, it's focused on runtime analysis; that is, method calls and method call sequences.

Here's an important caveat: pippi is not, and more importantly cannot, be free of false positives. That's because of the halting problem. Pippi finds suboptimal API usage based on data flows as driven by a project's test suite. There may be alternate data flows where this API usage is correct. For example, in the code below, if rand < 0.5 is true, then the Array will be mutated and the program cannot correctly be simplified by replacing "select followed by first" with "detect":

x = [1, 2, 3].select { |y| y > 1 }
x.reject! { |y| y > 2 } if rand < 0.5
x.first

There are various techniques that eliminate many of these false positives. For example, after flagging an issue, pippi watches subsequent method invocations and if those indicate the initial problem report was in error it'll remove the problem from the report.

Pippi is entirely dependent on the test suite to execute code in order to find problems. If a project's test code coverage is small, pippi probably won't find much.

Here's how pippi stacks up using the Aaron Quint Ruby Performance Character Profiles system:

• Specificity - very specific, finds actual detailed usages of bad code
• Impact - very impactful, slows things down lots
• Difficulty of Operator Use - easy to install, just a new gemfile entry
• Readability - results are easy to read
• Realtimedness - finds stuff right away
• Special Abilities - ?

Finally, why "pippi"? Because Pippi Longstocking was a Thing-Finder, and pippi finds things.

• Add gem 'pippi' to the test group in your project's Gemfile
• Add this to test_helper.rb just before the require 'rails/test_help' line

if ENV['USE_PIPPI'].present?
  Pippi::AutoRunner.new(:checkset => ENV['PIPPI_CHECKSET'] || "basic")
  # you can also pass in an IO:
  # Pippi::AutoRunner.new(:checkset => "basic", :io => $stdout)
end

• Run it:

USE_PIPPI=true bundle exec rake test:units && cat log/pippi.log

• You can also select a different checkset:

USE_PIPPI=true PIPPI_CHECKSET=rails bundle exec rake test:units && cat log/pippi.log

• And you can run multiple checksets:

USE_PIPPI=true PIPPI_CHECKSET=basic,rails bundle exec rake test:units && cat log/pippi.log

Here's a demo Rails application.

• Add gem 'pippi' to the test group in your project's Gemfile
• Add this to spec/spec_helper.rb or spec/rails_helper.rb, just below the require 'rspec/rails' line (if there is one):

if ENV['USE_PIPPI'].present?
  require 'pippi'
  Pippi::AutoRunner.new(:checkset => ENV['PIPPI_CHECKSET'] || "basic")
end

• Run it:

USE_PIPPI=true bundle exec rake spec && cat log/pippi.log

Dan Kohn suggests you could use something like:

if grep -v gem < log/pippi.log; then echo "$(wc -l < log/pippi.log) Pippi flaws found" && false; else echo 'No pippi flaws found'; fi

Assuming you're using bundler:

# Add this to your project's Gemfile:
gem 'pippi'
# Run 'bundle', see some output
# To run a particular check:
bundle exec pippi tmp/tmpfile.rb MapFollowedByFlatten Foo.new.bar out.txt
# Or to run all the basic Pippi checks on your code and exercise it with MyClass.new.exercise_some_code:
bundle exec ruby -rpippi/auto_runner -e "MyClass.new.exercise_some_code"

Pippi has the concept of "checksets" which are, well, sets of checks. The current checksets are listed below.

Don't use reverse followed by each; use reverse_each instead

For example, rather than doing this:

[1,2,3].reverse.each {|x| x+1 }

Instead, consider doing this:

[1,2,3].reverse_each {|x| x+1 }

Don't use select followed by any?; use any? with a block instead

For example, rather than doing this:

[1,2,3].select {|x| x > 1 }.any?

Instead, consider doing this:

[1,2,3].any? {|x| x > 1 }

Don't use select followed by empty?; use none? instead

For example, rather than doing this:

[1,2,3].select {|x| x > 1 }.empty?

Instead, consider doing this:

[1,2,3].none? {|x| x > 1 }

Don't use select followed by first; use detect instead

For example, rather than doing this:

[1,2,3].select {|x| x > 1 }.first

Instead, consider doing this:

[1,2,3].detect {|x| x > 1 }

Don't use select followed by none?; use none? with a block instead

For example, rather than doing this:

[1,2,3].select {|x| x > 1 }.none?

Instead, consider doing this:

[1,2,3].none? {|x| x > 1 }

Don't use consecutive select blocks; use a single select instead

For example, rather than doing this:

[1,2,3].select {|x| x > 1 }.select {|x| x > 2 }

Instead, consider doing this:

[1,2,3].select {|x| x > 2 }

Don't use select followed by size; use count instead

For example, rather than doing this:

[1,2,3].select {|x| x > 1 }.size

Instead, consider doing this:

[1,2,3].count {|x| x > 1 }

Don't use assert_equal with nil as a first argument; use assert_nil instead

For example, rather than doing this:

x = nil ; assert_equal(nil, x)

Instead, consider doing this:

x = nil ; assert_nil(x)

Don't use map followed by flatten(1); use flat_map instead

For example, rather than doing this:

[1,2,3].map {|x| [x,x+1] }.flatten(1)

Instead, consider doing this:

[1,2,3].flat_map {|x| [x, x+1]}

Don't use String#strip followed by empty?; use String#blank? instead

For example, rather than doing this:

'   '.strip.empty?

Instead, consider doing this:

'   '.blank?

# unnecessary assignment since String#strip! mutates receiver
# wrong
x = x.strip!
# right
x.strip!

# Use Pathname
# wrong
File.read(File.join(Rails.root, "config", "database.yml")
# right
Rails.root.join("config", "database.yml").read

# Use Kernel#tap
# wrong
x = [1,2]
x << 3
return x
# right
[1,2].tap {|y| y << 3 }


# Rails checks

# No need to call to_i on ActiveRecord::Base methods passed to route generators
# wrong
product_path(@product.to_i)
# right
product_path(@product)

# something with replacing x.map.compact with x.select.map

This is not a good fit for Pippi because it involves a receiver usage that can be detected with static analysis.

wrong:

class Foo
  def self.bar
    Foo.new
  end
end

right:

class Foo
  def self.bar
    self.new
  end
end

You might wonder why Pippi "rails" checkset doesn't have the rule "replace "foobar".gsub(/foo/, '') with "foobar".remove(/foo/)". That's because of the behavior of global variables such as $&. This behavior is nicely explained by Frederick Cheung on this StackOverflow comment. It's also broken down by David Black here and by Aaron Patterson and others here. Due to the issue explained there, Pippi's technique of prepending a proxy method breaks code that's further downstream when used with the block form of gsub.

• Clean up this initial hacked out metaprogramming
• Finish refactoring duplicated code into MethodSequenceChecker

To see teacher output for a file tmp/baz.rb:

rm -f pippi_debug.log ; PIPPI_DEBUG=1 bundle exec pippi tmp/baz.rb DebugCheck Foo.new.bar tmp/out.txt ; cat pippi_debug.log

When trying to find issues in a project:

# in project directory (e.g., aasm)
rm -rf pippi_debug.log pippi.log .bundle/gems/pippi-0.0.1/ .bundle/cache/pippi-0.0.1.gem .bundle/specifications/pippi-0.0.1.gemspec && bundle update pippi --local && PIPPI_DEBUG=1 bundle exec ruby -rpippi/auto_runner -e "puts 'hi'" && grep -C 5 BOOM pippi_debug.log
# or to run some specs with pippi watching:
rm -rf pippi_debug.log pippi.log .bundle/gems/pippi-0.0.1/ .bundle/cache/pippi-0.0.1.gem .bundle/specifications/pippi-0.0.1.gemspec && bundle update pippi --local && PIPPI_DEBUG=1 bundle exec ruby -rpippi/auto_runner -Ispec spec/unit/*.rb

• Bump version number
• Move anything from 'training' to 'buggy' or elsewhere
• Tie off Changelog notes
• Regenerate docs with pippi:generate_docs, copy and paste that into README
• Commit, push
• Tag the release (e.g., git tag -a v0.0.8 -m 'v0.0.8' && git push origin v0.0.8)
• bundle exec gem build pippi.gemspec
• gem push pippi-x.gem
• Update pippi_demo

• Andrew Kozin: fixes to :io option
• Christopher Schramm(@cschramm) bugfixes in fault proc clearing
• Enrique Delgado: Documentation fixes
• Evan Phoenix(@evanphx) for the idea of watching method invocations at runtime using metaprogramming rather than using Tracepoint.
• Hubert Dąbrowski: Ruby 2.0.0 fixes
• Igor Kapkov(@igas) documentation fixes
• Josh Bodah: Better logging support
• LivingSocial for letting me develop and open source this utility.
• Martin Spickermann: Better output format
• Michael Bernstein(@mrb) (of CodeClimate fame) for an inspirational discussion of code anaysis in general.
• Olle Jonsson(@olleolleolle) rubocop fixes