Regexp's source's encoding is wrong

[nirvdrum]

The docs for the Regexp class are a bit misleading, stating for the source method "Returns the original string of the pattern." However, it appears MRI updates the source to match the encoding from encoding.

Here's an example:

MRI 2.2.2:

> ruby -v -e "r = Regexp.new('Hi', nil, 'e'); p r.encoding; p r.source.encoding"
ruby 2.2.2p95 (2015-04-13 revision 50295) [x86_64-linux]
-e:1: warning: encoding option is ignored - e
#<Encoding:US-ASCII>
#<Encoding:US-ASCII>

MRI 1.9.3:

> ruby -v -e "r = Regexp.new('Hi', nil, 'e'); p r.encoding; p r.source.encoding"
ruby 1.9.3p551 (2014-11-13 revision 48407) [x86_64-linux]
-e:1: warning: encoding option is ignored - e
#<Encoding:US-ASCII>
#<Encoding:US-ASCII>

JRuby 9k:

> bin/jruby -v -e "@method = :new; r = Regexp.send(@method, 'Hi', nil, 'e'); p r.encoding; p r.source.encoding"
jruby 9.0.0.0-SNAPSHOT (2.2.2) 2015-06-03 0430387 Java HotSpot(TM) 64-Bit Server VM 25.45-b02 on 1.8.0_45-b14 +jit [linux-amd64]
-e:1: warning: encoding option is ignored - e
#<Encoding:US-ASCII>
#<Encoding:UTF-8>

JRuby 1.7.20:

> ruby -v -e "r = Regexp.new('Hi', nil, 'e'); p r.encoding; p r.source.encoding"
jruby 1.7.20 (1.9.3p551) 2015-05-04 3086e6a on Java HotSpot(TM) 64-Bit Server VM 1.8.0_45-b14 +jit [linux-amd64]
-e:1 warning: encoding option is ignored - e
#<Encoding:US-ASCII>
#<Encoding:UTF-8>

[enebo]

This is fixed in 9k and 1.7 is EOL. Resolving against 9.2.0.0 so people see it