This fixes an error when output is redirected:
```
$ run_ruby -rreline -e '$stderr.puts Reline::Unicode.calculate_width("\u221a").inspect' </dev/null >/dev/null
/home/jeremy/tmp/ruby/lib/reline/ansi.rb:189:in `raw': Operation not supported by device (Errno::ENODEV)
```
The @@encoding -> defined?(@@encoding) changes is necessary because
without that part of the commit, the following error would be raised
by the above command:
```
/home/jeremy/tmp/reline/lib/reline/general_io.rb:10:in `encoding': uninitialized class variable @@encoding in Reline::GeneralIO (NameError)
```
Problem reported and initial patch for Windows provided by
Richard Sharman.
I tested this only on OpenBSD, but hopefully it works for other
operating systems.
Fixes [Bug #17493]
c001971bb3
In Japan, so many programmers used EUC-JP to write text files that contain
Japanese. Many .inputrc files which contain EUC-JP are still being copied and
used. This commit supports the whole encoding of what user set including UTF-8.
ref. https://github.com/ruby/reline/pull/2800b45022e16
IRB uses Reline's 3 dynamic real-time callbacks with calling Ripper;
output_modifier_proc, prompt_proc, and auto_indent_proc. These processing
times make the paste time too long.
beec3399a8
The command prompt on Windows always uses Unicode to take input and print
output but most Reline implementation depends on Encoding.default_external.
This commit introduces an abstracted structure about the encoding of Reline.
