regexp_mem.rb

puts RUBY_DESCRIPTION
require 'benchmark'

N = Integer(ARGV[0] || 50_000_000)

content = "Accept-Language: fr-FR, en-US#{"0" * N}"

def print_size(context, sike_kb)
  puts "#{context}: #{(sike_kb / 1024.0).round(1)} MB"
end

print_size("baseline", `ps ax -o pid,rss | grep -E "^[[:space:]]*#{$$}"`.strip.split.map(&:to_i).last)

print_size("content", content.bytesize / 1024)

REGEXP_UNION = Regexp.union(
  "Accept-Language:",
  "Thread-Topic:",
).freeze
reg = /.*#{REGEXP_UNION}/
p reg

GC.start
print_size("before match?", `ps ax -o pid,rss | grep -E "^[[:space:]]*#{$$}"`.strip.split.map(&:to_i).last)
allocated = GC.stat(:total_allocated_objects)

p Benchmark.realtime {
  p content.match(reg)
}

p GC.stat(:total_allocated_objects) - allocated
GC.start
print_size("after match?", `ps ax -o pid,rss | grep -E "^[[:space:]]*#{$$}"`.strip.split.map(&:to_i).last)
p "---"