From e4570039c135e13fadd93df4766b61407098577f Mon Sep 17 00:00:00 2001
From: John W Higgins <wishdev@gmail.com>
Date: Thu, 26 Jan 2023 15:07:49 -0800
Subject: [PATCH] Move the host request parsing to a separate method. (#85)

Allows for someone to override the parsing to accommodate "alternatives".
One could write the following to allow underscores in host names.

require 'webrick/httprequest'

module WEBrick
  class HTTPRequest
    private

    def parse_host_request_line(host, scheme)
      uri = URI.parse("#{scheme}://#{host}")
      [uri.host, uri.port]
    end
  end
end

Also adding the "o" option to the regex so the regex is only built once.
---
 lib/webrick/httprequest.rb | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/lib/webrick/httprequest.rb b/lib/webrick/httprequest.rb
index ff2c8a8..680ac65 100644
--- a/lib/webrick/httprequest.rb
+++ b/lib/webrick/httprequest.rb
@@ -491,8 +491,7 @@ def parse_uri(str, scheme="http")
       if @forwarded_host
         host, port = @forwarded_host, @forwarded_port
       elsif self["host"]
-        pattern = /\A(#{URI::REGEXP::PATTERN::HOST})(?::(\d+))?\z/n
-        host, port = *self['host'].scan(pattern)[0]
+        host, port = parse_host_request_line(self["host"])
       elsif @addr.size > 0
         host, port = @addr[2], @addr[1]
       else
@@ -504,6 +503,11 @@ def parse_uri(str, scheme="http")
       return URI::parse(uri.to_s)
     end
 
+    def parse_host_request_line(host)
+      pattern = /\A(#{URI::REGEXP::PATTERN::HOST})(?::(\d+))?\z/no
+      host.scan(pattern)[0]
+    end
+
     def read_body(socket, block)
       return unless socket
       if tc = self['transfer-encoding']