Profile memory usage from processing HTML contents

.github/workflows/actions/munge_gemfile.rb

@@ -0,0 +1,8 @@
+# frozen_string_literal: true
+
+CONTENTS = <<~RUBY
+ source 'https://rubygems.org'
+ gemspec
+RUBY
+
+File.open('Gemfile', 'wb') { |f| f.puts CONTENTS }

.github/workflows/memprof.yml

@@ -0,0 +1,38 @@
+name: Memory Profile
+on:
+ push:
+ branches:
+ - master
+ pull_request:
+ branches:
+ - master
+
+jobs:
+ memory_profile:
+ if: "!contains(github.event.commits[0].message, '[ci skip]')"
+ runs-on: "ubuntu-latest"
+ steps:
+ - uses: actions/checkout@master
+ - uses: actions/setup-ruby@v1
+ with:
+ ruby-version: '2.7'
+ - name: Set up Dependencies Cache
+ uses: actions/cache@v1
+ with:
+ path: vendor/bundle
+ key: ubuntu-latest-gems
+ restore-keys: |
+ ubuntu-latest-gems-
+ - name: Install dependencies
+ run: |
+ bundle config path vendor/bundle
+ bundle install -j8
+ bundle exec rake gem:spec
+ bundle exec ruby .github/workflows/actions/munge_gemfile.rb
+ bundle install -j8
+ - name: Compile Nokogiri
+ run: bundle exec rake compile
+ - name: Memory Profile of parsing and serializing HTML Document
+ run: bundle exec rake memprof:html_doc
+ - name: Memory Profile of parsing and serializing HTML Document Fragment
+ run: bundle exec rake memprof:html_doc_frag

Gemfile

@@ -13,6 +13,7 @@ gem "hoe-debugging", "~>2.0", :group => [:development, :test]
 gem "hoe-gemspec", "~>1.0", :group => [:development, :test]
 gem "hoe-git", "~>1.6", :group => [:development, :test]
 gem "hoe-markdown", "~>1.1", :group => [:development, :test]
+gem "memory_profiler", "~>0.9", :group => [:development, :test]
 gem "minitest", "~>5.8", :group => [:development, :test]
 gem "racc", "~>1.5.0", :group => [:development, :test]
 gem "rake", "~>13.0", :group => [:development, :test]

Rakefile

@@ -54,6 +54,7 @@ HOE = Hoe.spec 'nokogiri' do
  ["hoe-gemspec", "~> 1.0"],
  ["hoe-git", "~> 1.6"],
  ["hoe-markdown", "~> 1.1"],
+ ["memory_profiler", "~> 0.9"],
  ["minitest", "~> 5.8"],
  ["racc", "~> 1.5.0"],
  ["rake", "~> 13.0"],
@@ -80,6 +81,7 @@ require_relative "tasks/debug"
 require_relative "tasks/docker"
 require_relative "tasks/docs-linkify"
 require_relative "tasks/rubocop"
+require_relative "tasks/memprof"
 require_relative "tasks/set-version-to-timestamp"
 
 # work around Hoe's inflexibility about the default tasks

tasks/memprof.rb

@@ -0,0 +1,34 @@
+# frozen_string_literal: true
+
+namespace :memprof do
+ require "memory_profiler"
+ module NokoMemProf
+ def self.profile abs_path
+ puts "\nTesting with Nokogiri v" + Nokogiri::VERSION.to_s
+ puts
+ contents = File.read abs_path
+ MemoryProfiler.report do
+ 1000.times { yield contents }
+ end.pretty_print(scale_bytes: true, normalize_paths: true)
+ end
+ end
+
+
+ task :html_doc do
+ require "nokogiri"
+
+ abs_path = File.expand_path("../test/files/fixture.html", __dir__)
+ NokoMemProf.profile(abs_path) do |contents|
+ Nokogiri::HTML::Document.parse(contents).to_html
+ end
+ end
+
+ task :html_doc_frag do
+ require "nokogiri"
+
+ abs_path = File.expand_path("../test/files/fixture_body.html", __dir__)
+ NokoMemProf.profile(abs_path) do |contents|
+ Nokogiri::HTML::DocumentFragment.parse(contents).to_html
+ end
+ end
+end