textwrap 0.1.0 → 0.1.1

data/README.txt

@@ -4,16 +4,20 @@ Copyright (c) 2007 SUGAWARA Genki <sgwr_dts@yahoo.co.jp>
 
 == Example
 
+    $KCODE = 'u'
     require 'textwrap'
     include Textwrap
-    puts Textwrap.fill(<<EOS, 40)
+
+    p fill(<<EOS, 40, :consider_linebreak_whitespace => true)
     The wrap() method is just like fill() except that it returns
     a list of strings instead of one big string with newlines to separate
     the wrapped lines.
     EOS
+    #=> "The wrap() method is just like fill()\nexcept that it returns a list of strings\ninstead of one big with newlines\nto separate the wrapped lines."
 
-    # The wrap() method is just like fill()
-    # except that it returns a list of strings
-    # instead of one big string with newlines
-    # to separate the wrapped lines.
-
+    p fill(<<EOS, 40)
+    拙者親方と申すは御立会の内に御存知の御方も御座りましょうが、御江戸を発って二十里上方、
+    相州小田原一色町を御過ぎなされて、青物町を上りへ御出でなさるれば、欄干橋虎屋藤右衛門、
+    只今では剃髪致して圓斎と名乗りまする。
+    EOS
+    #=> "拙者親方と申すは御立会の内に御存知の御方\nも御座りましょうが、御江戸を発って二十里\n上方、相州小田原一色町を御過ぎなされて、\n青物町を上りへ御出でなさるれば、欄干橋虎\n屋藤右衛門、只今では剃髪致して圓斎と名乗\nりまする。"

data/lib/textwrap.rb

@@ -1,24 +1,22 @@
-#
-# = testwrap.rb
-#
-# Copyright (c) 2007 SUGAWARA Genki <sgwr_dts@yahoo.co.jp>
-#
-# == Example
-#
+#     $KCODE = 'u'
 #     require 'textwrap'
 #     include Textwrap
-#     puts Textwrap.fill(<<EOS, 40)
+#
+#     p fill(<<EOS, 40, :consider_linebreak_whitespace => true)
 #     The wrap() method is just like fill() except that it returns
 #     a list of strings instead of one big string with newlines to separate
 #     the wrapped lines.
 #     EOS
+#     #=> "The wrap() method is just like fill()\nexcept that it returns a list of strings\ninstead of one big with newlines\nto separate the wrapped lines."
+#
+#     p fill(<<EOS, 40)
+#     拙者親方と申すは御立会の内に御存知の御方も御座りましょうが、御江戸を発って二十里上方、
+#     相州小田原一色町を御過ぎなされて、青物町を上りへ御出でなさるれば、欄干橋虎屋藤右衛門、
+#     只今では剃髪致して圓斎と名乗りまする。
+#     EOS
+#     #=> "拙者親方と申すは御立会の内に御存知の御方\nも御座りましょうが、御江戸を発って二十里\n上方、相州小田原一色町を御過ぎなされて、\n青物町を上りへ御出でなさるれば、欄干橋虎\n屋藤右衛門、只今では剃髪致して圓斎と名乗\nりまする。"
 #
-#     # The wrap() method is just like fill()
-#     # except that it returns a list of strings
-#     # instead of one big string with newlines
-#     # to separate the wrapped lines.
 #
-
 require 'strscan'
 
 module Textwrap
@@ -30,32 +28,76 @@ module Textwrap
 
     initial_indent    = options.fetch(:initial_indent, '')
     subsequent_indent = options.fetch(:subsequent_indent, '')
+    break_long_words  = options.fetch(:break_long_words, true)
+
+    code = options.fetch(:code, $KCODE)
+
+    if (m = /\A[esu]/i.match(code))
+      code = m[0].downcase
+    else
+      code = 'n'
+    end
 
     if options.fetch(:expand_tabs, true)
       text = text.gsub(/\t/, '        ') # SP x 8
     end
 
     if options.fetch(:replace_whitespace, true)
-      text = text.gsub(/\s/, ' ')
+      text = text.gsub(/\t\f/, ' ')
+    end
+
+    if options.has_key?(:consider_linebreak_whitespace)
+      consider_linebreak_whitespace = options[:consider_linebreak_whitespace]
+    elsif code == 'n'
+      consider_linebreak_whitespace = true
+    else
+      consider_linebreak_whitespace = false
     end
 
+    gauge_width = lambda {|str|
+      w = 0
+
+      str.scan(Regexp.new('.', nil, code)) do |c|
+        w += (c[0] <= 0x7f) ? 1 : 2
+      end
+
+      return w
+    }
+
     if options.fetch(:break_long_words, true)
-      regexps = [/\s{1,#{width}}/, /\S{1,#{width}}/]
+      break_word = lambda {|src, scnr|
+        return [src, scnr] if gauge_width.call(src) <= width
+        w = 0; dest = ''
+
+        src.scan(Regexp.new('.', nil, code)) do |c|
+          w += (c[0] <= 0x7f) ? 1 : 2
+
+          break if w > width
+          dest << c
+        end
+
+        new_scnr = StringScanner.new(src.slice(dest.length..-1) + scnr.rest)
+        return [dest, new_scnr]
+      }
     else
-      regexps = [/\s+/, /\S+/]
+      break_word = lambda {|str, scnr| [str, scnr]}
     end
 
-    s = StringScanner.new(text.split($/).join)
+    regexps = [/\s+/, /\S+/]
+    s = StringScanner.new(text.split(/[\r\n]+/).join(consider_linebreak_whitespace ? ' ' : ''))
     lines = []
     indent = initial_indent
-    words = [s.scan(regexps.reverse!.first) || s.scan(regexps.reverse!.first) || '']
+    word = s.scan(regexps.reverse!.first) || s.scan(regexps.reverse!.first) || ''
+    word, s = break_word.call(word, s)
+    words = [word]
 
     until s.eos?
       word = s.scan(regexps.reverse!.first) or next
-      line = indent + words.join.strip
+      word, s = break_word.call(word, s)
+      line = indent + words.join.sub(/\A\s+/, '')
 
-      if (line + word).length > width
-        lines << line
+      if gauge_width.call(line + word) > width
+        lines << line.sub(/\s+\Z/, '')
         words.clear
         indent = subsequent_indent
       end

metadata

@@ -3,8 +3,8 @@ rubygems_version: 0.9.2
 specification_version: 1
 name: textwrap
 version: !ruby/object:Gem::Version 
-  version: 0.1.0
-date: 2007-10-20 00:00:00 +09:00
+  version: 0.1.1
+date: 2007-10-28 00:00:00 +09:00
 summary: Text wrapping and filling library like Python textwrap module.
 require_paths: 
 - lib