From 623787d240c601ba5d11ff68b8b6f5bb6bcce3df Mon Sep 17 00:00:00 2001
From: yugui <yugui@b2dd03c8-39d4-4d8f-98ff-823fe69b080e>
Date: Mon, 29 Dec 2008 09:26:34 +0000
Subject: * lib/erb.rb (ERB): m17n of ERB. adds rdoc.   fixes #712. c.f.
 [ruby-dev:37516].

* lib/erb.rb (ERB::Compiler#compile): recognizes magic comments.
  returns a pair of compiled script and its script encoding.

* lib/erb.rb (ERB#set_eoutvar): make generated scripts return a
  string in correct encoding.

* lib/erb.rb (ERB#def_method): use Kernel#eval for encoding-awareness
  of the evaluated string.

* bin/erb.rb (ERB::Main.run): adds -E and -U options.
  String is no longer Enumerable.

* man/erb.1: new manapage.

* test/erb/test_erb_m17n.rb: new test case for m17n features.

git-svn-id: http://svn.ruby-lang.org/repos/ruby/trunk@21170 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
---
 lib/erb.rb | 51 +++++++++++++++++++++++++++++++++++++++++++++------
 1 file changed, 45 insertions(+), 6 deletions(-)

(limited to 'lib/erb.rb')

diff --git a/lib/erb.rb b/lib/erb.rb
index c87994128..c5194817b 100644
--- a/lib/erb.rb
+++ b/lib/erb.rb
@@ -58,6 +58,24 @@
 #
 # See the ERB.new and ERB#result methods for more detail.
 #
+# == Character encodings
+#
+# ERB (or ruby code generated by ERB) returns a string in the same
+# character encoding as the input string.  When the input string has
+# a magic comment, however, it returns a string in the encoding specified
+# by the magic comment.
+#
+#   # -*- coding: UTF-8 -*-
+#   require 'erb'
+#
+#   template = ERB.new <<EOF
+#   <%#-*- coding: Big5 -*-%>
+#     \_\_ENCODING\_\_ is <%= \_\_ENCODING\_\_ %>.
+#   EOF
+#   puts template.result
+#
+# <em>Prints:</em> \_\_ENCODING\_\_ is Big5.
+#
 #
 # == Examples
 #
@@ -517,10 +535,14 @@ class ERB
     end
 
     def compile(s)
-      out = Buffer.new(self, s.encoding)
+      enc = s.encoding
+      raise ArgumentError, "#{enc} is not ASCII compatible" if enc.dummy?
+      s = s.dup.force_encoding("ASCII-8BIT") # don't use constant Enoding::ASCII_8BIT for miniruby
+      enc = detect_magic_comment(s) || enc
+      out = Buffer.new(self, enc)
 
       content = ''
-      scanner = make_scanner(s.dup.force_encoding("ASCII-8BIT"))
+      scanner = make_scanner(s) 
       scanner.scan do |token|
 	if scanner.stag.nil?
 	  case token
@@ -560,6 +582,7 @@ class ERB
 	    when '<%='
 	      out.push("#{@insert_cmd}((#{content}).to_s)")
 	    when '<%#'
+	      # content = content.force_encoding(@enc)
 	      # out.push("# #{content.dump}")
 	    end
 	    scanner.stag = nil
@@ -573,7 +596,7 @@ class ERB
       end
       out.push("#{@put_cmd} #{content.dump}") if content.size > 0
       out.close
-      out.script
+      return out.script, enc
     end
 
     def prepare_trim_mode(mode)
@@ -613,6 +636,18 @@ class ERB
     end
     attr_reader :percent, :trim_mode
     attr_accessor :put_cmd, :insert_cmd, :pre_cmd, :post_cmd
+
+    private
+    def detect_magic_comment(s)
+      if /\A<%#(.*)%>/ =~ s or (@percent and /\A%#(.*)/ =~ s)
+	comment = $1
+	comment = $1 if comment[/-\*-\s*(.*?)\s*-*-$/]
+	if %r"coding\s*[=:]\s*([[:alnum:]\-_]+)" =~ comment
+	  enc = $1.sub(/-(?:mac|dos|unix)/i, '')
+	  enc = Encoding.find(enc)
+	end
+      end
+    end
   end
 end
 
@@ -688,7 +723,7 @@ class ERB
     @safe_level = safe_level
     compiler = ERB::Compiler.new(trim_mode)
     set_eoutvar(compiler, eoutvar)
-    @src = compiler.compile(str)
+    @src, @enc = *compiler.compile(str)
     @filename = nil
   end
 
@@ -714,7 +749,7 @@ class ERB
     compiler.pre_cmd = cmd
 
     cmd = []
-    cmd.push(eoutvar)
+    cmd.push("#{eoutvar}.force_encoding(__ENCODING__)")
 
     compiler.post_cmd = cmd
   end
@@ -745,7 +780,11 @@ class ERB
   end
 
   def def_method(mod, methodname, fname='(ERB)')  # :nodoc:
-    mod.module_eval("def #{methodname}\n" + self.src + "\nend\n", fname, -1)
+    src = self.src
+    magic_comment = "#coding:#{@enc}\n"
+    mod.module_eval do
+      eval(magic_comment + "def #{methodname}\n" + src + "\nend\n", binding, fname, -2)
+    end
   end
 
   def def_module(methodname='erb')  # :nodoc:
-- 
cgit