[win32/registry] Fallback to UTF-8 for unknown codepages

	There are some codepages like cp708 for which no ruby encoding exists:

	    $ ruby -e "Encoding.find('cp708')"
	    Traceback (most recent call last):
	    	1: from -e:1:in `<main>'
	    -e:1:in `find': unknown encoding name - cp708 (ArgumentError)

	win32/registry uses ENCODING to transcode error messages and expand environment
	variables from UTF-16LE, so using UTF-8 seems like the best choice and is better
	than a hard failure.

	This should resolve [Bug #13831]

Merge JSON 2.7.2

	[Bug #20649] Allow `nil` as 2nd argument of `assign_error`

	Fallback to the last token element in that case, for the backward
	compatibilities.

When the registerred unblock function is called, it should retry
the cancelled blocking function if possible after checkints.

For example, `SIGCHLD` can cancel this method, but it should not
raise any exception if there is no trap handlers.

The following is repro-code:

```ruby
require 'socket'
PN = 10_000

1000000.times{
  p _1
  PN.times{
    fork{
      sleep rand(0.3)
    }
  }
  i = 0
  while i<PN
    cpid = Process.wait -1, Process::WNOHANG
    if cpid
      # p [i, cpid]
      i += 1
    end

    begin
      TCPServer.new(nil, 0).close
    rescue
      p $!
      exit!
    end
  end
}
```

	[Bug #20592] Fix segfault when sending NULL to freeaddrinfo

	On alpine freeaddrinfo does not accept NULL pointer

	Fix dangling `else`

	Revert "Rollback to released version numbers of stringio and strscan"

	This reverts commit 6a79e53823e328281b9e9eee53cd141af28f8548.

	[ruby/strscan] StringScanner#captures: Return nil not "" for unmached capture (https://github.com/ruby/strscan/pull/72)

	fix https://github.com/ruby/strscan/issues/70
	If there is no substring matching the group (s[3]), the behavior is
	different.

	If there is no substring matching the group, the corresponding element
	(s[3]) should be nil.

	```
	s = StringScanner.new('foobarbaz') #=> #<StringScanner 0/9 @ "fooba...">
	s.scan /(foo)(bar)(BAZ)?/  #=> "foobar"
	s[0]           #=> "foobar"
	s[1]           #=> "foo"
	s[2]           #=> "bar"
	s[3]           #=> nil
	s.captures #=> ["foo", "bar", ""]
	s.captures.compact #=> ["foo", "bar", ""]
	```

	```
	s = StringScanner.new('foobarbaz') #=> #<StringScanner 0/9 @ "fooba...">
	s.scan /(foo)(bar)(BAZ)?/  #=> "foobar"
	s[0]           #=> "foobar"
	s[1]           #=> "foo"
	s[2]           #=> "bar"
	s[3]           #=> nil
	s.captures #=> ["foo", "bar", nil]
	s.captures.compact #=> ["foo", "bar"]
	```

	https://docs.ruby-lang.org/ja/latest/method/MatchData/i/captures.html
	```
	/(foo)(bar)(BAZ)?/ =~ "foobarbaz" #=> 0
	$~.to_a        #=> ["foobar", "foo", "bar", nil]
	$~.captures #=> ["foo", "bar", nil]
	$~.captures.compact #=> ["foo", "bar"]
	```

	* StringScanner#captures is not yet documented.
	https://docs.ruby-lang.org/ja/latest/class/StringScanner.html

	https://github.com/ruby/strscan/commit/1fbfdd3c6f

	[ruby/strscan] Bump version

	https://github.com/ruby/strscan/commit/d6f97ec102

	[ruby/zlib] In Zlib::GzipReader#eof? check if we're actually at eof

	Only consider it eof if we read ahead and something fills the buf.
	If not, we may only have empty blocks and the footer.

	Fixes https://github.com/ruby/zlib/pull/56

	https://github.com/ruby/zlib/commit/437bea8003

	[Bug #20322] Fix rb_enc_interned_str_cstr null encoding

	The documentation for `rb_enc_interned_str_cstr` notes that `enc` can be
	a null pointer, but this currently causes a segmentation fault when
	trying to autoload the encoding. This commit fixes the issue by checking
	for NULL before calling `rb_enc_autoload`.