attachments 0.0.1 → 0.0.2

data/VERSION

@@ -1 +1 @@
-0.0.1
+0.0.2

data/lib/attachments/extract.rb

@@ -2,6 +2,7 @@ require 'rubygems'
 require 'uuidtools'
 require 'fileutils'
 require 'mail'
+require 'iconv'
 
 
 module Attachments
@@ -76,7 +77,7 @@ module Attachments
 
     def parse_part mail
       # Filter parts with a type that is not of interest
-      ct = mail.content_type.split(/;/, 2)[0] if mail.content_type
+      ct, charset = mail.content_type.split(/;/, 2) if mail.content_type
       unless(mail.multipart? || (ct && @include_types.include?(ct)))
         return
       end
@@ -105,7 +106,13 @@ module Attachments
                    m.encode("utf-8")
                  rescue
                    # Ruby 1.8 doesn't know encode
-                   m
+                   if charset && charset.match(/charset=/i)
+                     charset.gsub!(/charset=/i, "")
+                     charset.strip!
+                     Iconv.conv("utf-8", charset, m)
+                   else
+                     m
+                   end
                  end
                else
                  mail.body.decoded

data/test/test_attachments.rb

@@ -1,4 +1,5 @@
 require 'helper'
+require 'iconv'
 
 class TestAttachments < Test::Unit::TestCase
   def setup
@@ -30,4 +31,14 @@ class TestAttachments < Test::Unit::TestCase
     @extract.close
     assert !File::exists?(tmpfile)
   end
+
+
+  should "save as text/plain as utf-8" do
+    tmpfile = @extract.files[0][:tmpfile]
+    m = File.read(tmpfile)
+    assert_nothing_raised do
+      # Provoke an exception if not valid utf-8
+      Iconv.conv("ISO-8859-1", "utf-8", m)
+    end
+  end
 end

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: attachments
 version: !ruby/object:Gem::Version 
-  hash: 29
+  hash: 27
   prerelease: false
   segments: 
   - 0
   - 0
-  - 1
-  version: 0.0.1
+  - 2
+  version: 0.0.2
 platform: ruby
 authors: 
 - Rune Myrland