Dynamic regex seems to return incorrect, cached regex object (in --dev) #6992
Description
Environment Information
Provide at least:
- JRuby version (
jruby -v) and command line (flags, JRUBY_OPTS, etc)jruby 9.2.16.0 (2.5.7) 2021-03-03 f82228dc32 OpenJDK 64-Bit Server VM 25.292-b10 on 1.8.0_292-b10 +jit [darwin-x86_64]JRUBY_OPTS=--dev
- Operating system and platform (e.g.
uname -a)Darwin OT-MBP.local 20.6.0 Darwin Kernel Version 20.6.0: Wed Nov 10 22:23:07 PST 2021; root:xnu-7195.141.14~1/RELEASE_X86_64 x86_64
Expected Behavior
File: jruby_dynamic_regex_bug.rb
# frozen_string_literal: true
require "bundler/inline"
gemfile(true) do
source "https://rubygems.org"
gem 'concurrent-ruby'
end
alpha = "a-z"
digit = "0-9"
def keep_chars(component, character_class)
component.gsub(/[^#{character_class}]/, '')
end
n_success = Concurrent::AtomicFixnum.new(0)
n_failed = Concurrent::AtomicFixnum.new(0)
threads = 50.times.map do
Thread.new do
1_000.times do
hex = SecureRandom.hex
only_alpha = keep_chars(hex, alpha)
only_digit = keep_chars(hex, digit)
if only_digit.match?(/\A[0-9]+\z/) && only_alpha.match?(/\A[a-z]+\z/i)
n_success.increment
else
puts "[#{Thread.current.object_id}] [FAILED] hex = #{hex} only_alpha = #{only_alpha} only_digit = #{only_digit}"
n_failed.increment
end
sleep 0.01
end
end
end
threads.each(&:join)
puts
puts "*" * 80
puts [n_success.value, n_failed.value].inspect
puts "*" * 80$ jruby jruby_dynamic_regex_bug.rb
[...]
********************************************************************************
[50000, 0]
********************************************************************************
Note: without --dev we get the expected result.
Actual Behavior
$ jruby --dev jruby_dynamic_regex_bug.rb
[...]
********************************************************************************
[49675, 325]
********************************************************************************
Note: the exact results vary but should be [50000, 0].
If the method is changed to
def keep_chars(component, character_class)
component.gsub(Regexp.new("[^#{character_class}]"), '')
endit works fine.