[DRE-commits] [SCM] bsfilter.git branch, master, updated. debian/1%1.0.17-2-15-g15ef7f6

[Taku YASUI]

The following commit has been merged in the master branch:
commit a6181296af9a55cdea0f9c7e8d27d28ee4023d5f
Author: Taku YASUI <tach at debian.org>
Date:   Sun Jun 17 22:20:36 2012 +0900

    Add debian/patches/010_disable_chasen not to use chasen.

diff --git a/debian/changelog b/debian/changelog
index 3b0cd17..6ac9c23 100644
--- a/debian/changelog
+++ b/debian/changelog
@@ -5,6 +5,7 @@ bsfilter (1:1.0.17-3) UNRELEASED; urgency=low
   * debian/copyright: Convert to the Debian copyright format version 1.0.
   * debian/docs: Add HTML document to the package.
   * debian/rules: Add to run test suite.
+  * Add debian/patches/010_disable_chasen not to use chasen.
 
  -- Taku YASUI <tach at debian.org>  Sun, 17 Jun 2012 21:54:15 +0900
 
diff --git a/debian/patches/debian-changes-1:1.0.17-1 b/debian/patches/000_debian-changes-1:1.0.17-1
similarity index 100%
rename from debian/patches/debian-changes-1:1.0.17-1
rename to debian/patches/000_debian-changes-1:1.0.17-1
diff --git a/debian/patches/010_disable_chasen b/debian/patches/010_disable_chasen
new file mode 100644
index 0000000..7902124
--- /dev/null
+++ b/debian/patches/010_disable_chasen
@@ -0,0 +1,112 @@
+--- a/bsfilter/bsfilter
++++ b/bsfilter/bsfilter
+@@ -1,4 +1,4 @@
+-#! /usr/bin/env ruby
++#! /usr/bin/ruby1.8
+ ## -*-Ruby-*- $Id: bsfilter,v 1.86 2010/11/20 10:07:45 nabeken Exp $
+ ## Copyright (C) 2003, 2004, 2005, 2006 NABEYA Kenichi
+ ##
+@@ -1057,9 +1057,6 @@
+         else
+           @m = MeCab::Tagger.new([$0, "-Ochasen"])
+         end
+-      when "chasen"
+-        Chasen.getopt("-F", '%H %m\n', "-j")
+-        @method = Proc::new {|s| chasen(s)}
+       when "kakasi"
+         @method = Proc::new {|s| kakasi(s)}
+       else
+@@ -1123,31 +1120,6 @@
+       return array
+     end
+     
+-    def chasen(str)
+-      str = str.gsub(/[\x00-\x7f]/, ' ')
+-      if (str =~ /\A +\z/)
+-        return []
+-      end
+-      array = Array::new
+-      Chasen.sparse(str).split("\n").each do |hinshi_token|
+-        if (hinshi_token =~ /(.*) (.*)/)
+-          hinshi = $1
+-          token = $2
+-          if (hinshi == "\xcc\xbe\xbb\xec")
+-            if ((token =~ Reg_kanji_katakana) || (token.length > 2))
+-              array.push(token)
+-            end
+-          else
+-            token.gsub!(Reg_not_kanji_katakana, '')
+-            if ((token =~ Reg_kanji) || (token.length > 2))
+-              array.push(token)
+-            end
+-          end
+-        end
+-      end
+-      return array
+-    end
+-    
+     def block(str)
+       tokens = str.scan(Reg_kanji)
+       tokens.concat(str.scan(Reg_katakana))
+@@ -1960,7 +1932,7 @@
+ 		specify the name of database type
+ 		"sdbm" by default
+ 
+-        --jtokenizer|-j bigram|block|mecab|chasen|kakasi
++        --jtokenizer|-j bigram|block|mecab|kakasi
+ 		specify algorithm of a tokenizer for Japanese language
+ 		"bigram" by default
+ 
+@@ -3144,8 +3116,6 @@
+     when "block"
+     when "mecab"
+       require 'MeCab'
+-    when "chasen"
+-      require 'chasen.o'
+     when "kakasi"
+       require 'kakasi'
+     else
+--- a/test/test.rb
++++ b/test/test.rb
+@@ -217,14 +217,9 @@
+     @bsfilter2.setup($default_options + ["--jtokenizer", "bigram"])
+     @bsfilter2.use_dummyfh
+ 
+-    @bsfilter3 = Bsfilter::new
+-    @bsfilter3.setup($default_options + ["--jtokenizer", "chasen"])
+-    @bsfilter3.use_dummyfh
+-
+     @bsfilter0.run(@files)
+     @bsfilter1.run(@files)
+     @bsfilter2.run(@files)
+-    @bsfilter3.run(@files)
+ 
+     assert_equal(1, @bsfilter0.count_message(/tokenizer ja body Ä«´é/),  "@bsfilter0 2letters")
+     assert_equal(0, @bsfilter0.count_message(/tokenizer ja body ¸þÆü°ª/), "@bsfilter0 3letters")
+@@ -234,9 +229,6 @@
+ 
+     assert_equal(1, @bsfilter2.count_message(/tokenizer ja body Ä«´é/), "@bsfilter2 2letters")
+     assert_equal(0, @bsfilter2.count_message(/tokenizer ja body ¸þÆü°ª/), "@bsfilter2 3letters")
+-
+-    assert_equal(1, @bsfilter3.count_message(/tokenizer ja body Ä«´é/), "@bsfilter3 2letters")
+-    assert_equal(1, @bsfilter3.count_message(/tokenizer ja body ¸þÆü°ª/), "@bsfilter3 3letters")
+   end
+ 
+   def teardown
+@@ -325,16 +317,6 @@
+     end
+     @bsfilter.use_dummyfh
+     @bsfilter.run(@files)
+-    assert_equal(1, @bsfilter.count_message(/tokenizer ja body Ä«´é/), "2 letters")
+-    assert_equal(1, @bsfilter.count_message(/tokenizer ja body ¸þÆü°ª/), "3 letters")
+-  end
+-
+-  def test_chasen
+-    assert_nothing_raised('Warning: ignore this test if chasen is NOT installed') do
+-      @bsfilter.setup($default_options + ["--jtokenizer", "chasen"])
+-    end
+-    @bsfilter.use_dummyfh
+-    @bsfilter.run(@files)
+     assert_equal(1, @bsfilter.count_message(/tokenizer ja body Ä«´é/), "2 letters")
+     assert_equal(1, @bsfilter.count_message(/tokenizer ja body ¸þÆü°ª/), "3 letters")
+   end
diff --git a/debian/patches/series b/debian/patches/series
index 29ff1be..778ff65 100644
--- a/debian/patches/series
+++ b/debian/patches/series
@@ -1 +1,2 @@
-debian-changes-1:1.0.17-1
+000_debian-changes-1:1.0.17-1
+010_disable_chasen

-- 
bsfilter.git