Inconsistent encoding when calling #to_s on arrays containing hashes with non us-ascii characters #6748
Description
Environment Information
- jruby-9.2.19.0 installed using rvm
- Linux tester 5.4.0-48-generic #52-Ubuntu SMP Thu Sep 10 10:58:49 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux
- Darwin macbook.local 19.6.0 Darwin Kernel Version 19.6.0: Thu Oct 29 22:56:45 PDT 2020; root:xnu-6153.141.2.2~1/RELEASE_X86_64 x86_64
Other relevant info you may wish to add:
- Discovered while tracking down a Logstash bug (Plugin can crash if fingerprinted data includes an array of hashes including non US-ASCII characters logstash-plugins/logstash-filter-fingerprint#63)
Expected Behavior
This is the behaviour from MRI:
ubuntu@tester:~$ irb
2.5.7 :001 >extended = [{"non_ascii"=>"ง"}, {"ascii" => "two"}]
=> [{"non_ascii"=>"ง"}, {"ascii"=>"two"}]
2.5.7 :002 > extended.to_s.encoding
=> #<Encoding:UTF-8>
2.5.7 :003 > "#{extended}".encoding
=> #<Encoding:UTF-8>
2.5.7 :004 > extended = [{"ascii" => "two"}, {"non_ascii"=>"ง"}]
=> [{"ascii"=>"two"}, {"non_ascii"=>"ง"}]
2.5.7 :005 > "#{extended}".encoding
=> #<Encoding:UTF-8>
2.5.7 :006 > extended.to_s.encoding
=> #<Encoding:UTF-8>
2.5.7 :007 > exit
Actual Behavior
ubuntu@tester:~$ irb
jruby-9.2.19.0 :001 > extended = [{"non_ascii"=>"ง"}, {"ascii" => "two"}]
=> [{"non_ascii"=>"ง"}, {"ascii"=>"two"}]
jruby-9.2.19.0 :002 > extended.to_s.encoding
=> #<Encoding:US-ASCII>
jruby-9.2.19.0 :003 > "#{extended}".encoding
=> #<Encoding:US-ASCII>
jruby-9.2.19.0 :004 > extended = [{"ascii" => "two"}, {"non_ascii"=>"ง"}]
=> [{"ascii"=>"two"}, {"non_ascii"=>"ง"}]
jruby-9.2.19.0 :005 > extended.to_s.encoding
=> #<Encoding:UTF-8>
jruby-9.2.19.0 :006 > "#{extended}".encoding
=> #<Encoding:UTF-8>