reline 0.5.12 → 0.6.1

data/lib/reline/unicode/east_asian_width.rb

@@ -1,6 +1,6 @@
 class Reline::Unicode::EastAsianWidth
   # This is based on EastAsianWidth.txt
-  # UNICODE_VERSION = '15.1.0'
+  # UNICODE_VERSION = '16.0.0'
 
   CHUNK_LAST, CHUNK_WIDTH = [
     [0x1f, 2],
@@ -174,7 +174,7 @@ class Reline::Unicode::EastAsianWidth
     [0x82d, 0],
     [0x858, 1],
     [0x85b, 0],
-    [0x897, 1],
+    [0x896, 1],
     [0x89f, 0],
     [0x8c9, 1],
     [0x8e1, 0],
@@ -646,6 +646,8 @@ class Reline::Unicode::EastAsianWidth
     [0x261c, -1],
     [0x261d, 1],
     [0x261e, -1],
+    [0x262f, 1],
+    [0x2637, 2],
     [0x263f, 1],
     [0x2640, -1],
     [0x2641, 1],
@@ -664,6 +666,8 @@ class Reline::Unicode::EastAsianWidth
     [0x266f, -1],
     [0x267e, 1],
     [0x267f, 2],
+    [0x2689, 1],
+    [0x268f, 2],
     [0x2692, 1],
     [0x2693, 2],
     [0x269d, 1],
@@ -753,14 +757,12 @@ class Reline::Unicode::EastAsianWidth
     [0x3130, 1],
     [0x318e, 2],
     [0x318f, 1],
-    [0x31e3, 2],
+    [0x31e5, 2],
     [0x31ee, 1],
     [0x321e, 2],
     [0x321f, 1],
     [0x3247, 2],
     [0x324f, -1],
-    [0x4dbf, 2],
-    [0x4dff, 1],
     [0xa48c, 2],
     [0xa48f, 1],
     [0xa4c6, 2],
@@ -879,9 +881,11 @@ class Reline::Unicode::EastAsianWidth
     [0x10ae6, 0],
     [0x10d23, 1],
     [0x10d27, 0],
+    [0x10d68, 1],
+    [0x10d6d, 0],
     [0x10eaa, 1],
     [0x10eac, 0],
-    [0x10efc, 1],
+    [0x10efb, 1],
     [0x10eff, 0],
     [0x10f45, 1],
     [0x10f50, 0],
@@ -943,6 +947,16 @@ class Reline::Unicode::EastAsianWidth
     [0x1136c, 0],
     [0x1136f, 1],
     [0x11374, 0],
+    [0x113ba, 1],
+    [0x113c0, 0],
+    [0x113cd, 1],
+    [0x113ce, 0],
+    [0x113cf, 1],
+    [0x113d0, 0],
+    [0x113d1, 1],
+    [0x113d2, 0],
+    [0x113e0, 1],
+    [0x113e2, 0],
     [0x11437, 1],
     [0x1143f, 0],
     [0x11441, 1],
@@ -982,6 +996,8 @@ class Reline::Unicode::EastAsianWidth
     [0x116b6, 1],
     [0x116b7, 0],
     [0x1171c, 1],
+    [0x1171d, 0],
+    [0x1171e, 1],
     [0x1171f, 0],
     [0x11721, 1],
     [0x11725, 0],
@@ -1059,10 +1075,16 @@ class Reline::Unicode::EastAsianWidth
     [0x11f40, 0],
     [0x11f41, 1],
     [0x11f42, 0],
+    [0x11f59, 1],
+    [0x11f5a, 0],
     [0x1343f, 1],
     [0x13440, 0],
     [0x13446, 1],
     [0x13455, 0],
+    [0x1611d, 1],
+    [0x16129, 0],
+    [0x1612c, 1],
+    [0x1612f, 0],
     [0x16aef, 1],
     [0x16af4, 0],
     [0x16b2f, 1],
@@ -1080,7 +1102,7 @@ class Reline::Unicode::EastAsianWidth
     [0x187f7, 2],
     [0x187ff, 1],
     [0x18cd5, 2],
-    [0x18cff, 1],
+    [0x18cfe, 1],
     [0x18d08, 2],
     [0x1afef, 1],
     [0x1aff3, 2],
@@ -1116,6 +1138,10 @@ class Reline::Unicode::EastAsianWidth
     [0x1d1ad, 0],
     [0x1d241, 1],
     [0x1d244, 0],
+    [0x1d2ff, 1],
+    [0x1d356, 2],
+    [0x1d35f, 1],
+    [0x1d376, 2],
     [0x1d9ff, 1],
     [0x1da36, 0],
     [0x1da3a, 1],
@@ -1148,6 +1174,8 @@ class Reline::Unicode::EastAsianWidth
     [0x1e2ef, 0],
     [0x1e4eb, 1],
     [0x1e4ef, 0],
+    [0x1e5ed, 1],
+    [0x1e5ef, 0],
     [0x1e8cf, 1],
     [0x1e8d6, 0],
     [0x1e943, 1],
@@ -1241,15 +1269,13 @@ class Reline::Unicode::EastAsianWidth
     [0x1fa6f, 1],
     [0x1fa7c, 2],
     [0x1fa7f, 1],
-    [0x1fa88, 2],
-    [0x1fa8f, 1],
-    [0x1fabd, 2],
-    [0x1fabe, 1],
-    [0x1fac5, 2],
+    [0x1fa89, 2],
+    [0x1fa8e, 1],
+    [0x1fac6, 2],
     [0x1facd, 1],
-    [0x1fadb, 2],
-    [0x1fadf, 1],
-    [0x1fae8, 2],
+    [0x1fadc, 2],
+    [0x1fade, 1],
+    [0x1fae9, 2],
     [0x1faef, 1],
     [0x1faf8, 2],
     [0x1ffff, 1],

data/lib/reline/unicode.rb

@@ -28,12 +28,12 @@ class Reline::Unicode
     0x19 => '^Y',
     0x1A => '^Z', # C-z
     0x1B => '^[', # C-[ C-3
+    0x1C => '^\\', # C-\
     0x1D => '^]', # C-]
     0x1E => '^^', # C-~ C-6
     0x1F => '^_', # C-_ C-7
     0x7F => '^?', # C-? C-8
   }
-  EscapedChars = EscapedPairs.keys.map(&:chr)
 
   NON_PRINTING_START = "\1"
   NON_PRINTING_END = "\2"
@@ -61,7 +61,7 @@ class Reline::Unicode
 
     # This code is essentially doing the same thing as
     # `str.encode(utf8, **replace_options).encode(encoding, **replace_options)`
-    # but also avoids unneccesary irreversible encoding conversion.
+    # but also avoids unnecessary irreversible encoding conversion.
     converted.gsub(/\X/) do |c|
       c.encode(Encoding::UTF_8)
       c
@@ -262,375 +262,126 @@ class Reline::Unicode
   end
 
   def self.em_forward_word(line, byte_pointer)
-    byte_size = 0
-    while line.bytesize > (byte_pointer + byte_size)
-      size = get_next_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size, size)
-      break if mbchar.encode(Encoding::UTF_8) =~ /\p{Word}/
-      byte_size += size
-    end
-    while line.bytesize > (byte_pointer + byte_size)
-      size = get_next_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size, size)
-      break if mbchar.encode(Encoding::UTF_8) =~ /\P{Word}/
-      byte_size += size
-    end
-    byte_size
+    gcs = line.byteslice(byte_pointer..).grapheme_clusters
+    nonwords = gcs.take_while { |c| !word_character?(c) }
+    words = gcs.drop(nonwords.size).take_while { |c| word_character?(c) }
+    nonwords.sum(&:bytesize) + words.sum(&:bytesize)
   end
 
   def self.em_forward_word_with_capitalization(line, byte_pointer)
-    byte_size = 0
-    new_str = String.new
-    while line.bytesize > (byte_pointer + byte_size)
-      size = get_next_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size, size)
-      break if mbchar.encode(Encoding::UTF_8) =~ /\p{Word}/
-      new_str += mbchar
-      byte_size += size
-    end
-    first = true
-    while line.bytesize > (byte_pointer + byte_size)
-      size = get_next_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size, size)
-      break if mbchar.encode(Encoding::UTF_8) =~ /\P{Word}/
-      if first
-        new_str += mbchar.upcase
-        first = false
-      else
-        new_str += mbchar.downcase
-      end
-      byte_size += size
-    end
-    [byte_size, new_str]
+    gcs = line.byteslice(byte_pointer..).grapheme_clusters
+    nonwords = gcs.take_while { |c| !word_character?(c) }
+    words = gcs.drop(nonwords.size).take_while { |c| word_character?(c) }
+    [nonwords.sum(&:bytesize) + words.sum(&:bytesize), nonwords.join + words.join.capitalize]
   end
 
   def self.em_backward_word(line, byte_pointer)
-    byte_size = 0
-    while 0 < (byte_pointer - byte_size)
-      size = get_prev_mbchar_size(line, byte_pointer - byte_size)
-      mbchar = line.byteslice(byte_pointer - byte_size - size, size)
-      break if mbchar.encode(Encoding::UTF_8) =~ /\p{Word}/
-      byte_size += size
-    end
-    while 0 < (byte_pointer - byte_size)
-      size = get_prev_mbchar_size(line, byte_pointer - byte_size)
-      mbchar = line.byteslice(byte_pointer - byte_size - size, size)
-      break if mbchar.encode(Encoding::UTF_8) =~ /\P{Word}/
-      byte_size += size
-    end
-    byte_size
+    gcs = line.byteslice(0, byte_pointer).grapheme_clusters.reverse
+    nonwords = gcs.take_while { |c| !word_character?(c) }
+    words = gcs.drop(nonwords.size).take_while { |c| word_character?(c) }
+    nonwords.sum(&:bytesize) + words.sum(&:bytesize)
   end
 
   def self.em_big_backward_word(line, byte_pointer)
-    byte_size = 0
-    while 0 < (byte_pointer - byte_size)
-      size = get_prev_mbchar_size(line, byte_pointer - byte_size)
-      mbchar = line.byteslice(byte_pointer - byte_size - size, size)
-      break if mbchar =~ /\S/
-      byte_size += size
-    end
-    while 0 < (byte_pointer - byte_size)
-      size = get_prev_mbchar_size(line, byte_pointer - byte_size)
-      mbchar = line.byteslice(byte_pointer - byte_size - size, size)
-      break if mbchar =~ /\s/
-      byte_size += size
-    end
-    byte_size
+    gcs = line.byteslice(0, byte_pointer).grapheme_clusters.reverse
+    spaces = gcs.take_while { |c| space_character?(c) }
+    nonspaces = gcs.drop(spaces.size).take_while { |c| !space_character?(c) }
+    spaces.sum(&:bytesize) + nonspaces.sum(&:bytesize)
   end
 
   def self.ed_transpose_words(line, byte_pointer)
-    right_word_start = nil
-    size = get_next_mbchar_size(line, byte_pointer)
-    mbchar = line.byteslice(byte_pointer, size)
-    if size.zero?
-      # ' aaa bbb [cursor]'
-      byte_size = 0
-      while 0 < (byte_pointer + byte_size)
-        size = get_prev_mbchar_size(line, byte_pointer + byte_size)
-        mbchar = line.byteslice(byte_pointer + byte_size - size, size)
-        break if mbchar.encode(Encoding::UTF_8) =~ /\p{Word}/
-        byte_size -= size
-      end
-      while 0 < (byte_pointer + byte_size)
-        size = get_prev_mbchar_size(line, byte_pointer + byte_size)
-        mbchar = line.byteslice(byte_pointer + byte_size - size, size)
-        break if mbchar.encode(Encoding::UTF_8) =~ /\P{Word}/
-        byte_size -= size
-      end
-      right_word_start = byte_pointer + byte_size
-      byte_size = 0
-      while line.bytesize > (byte_pointer + byte_size)
-        size = get_next_mbchar_size(line, byte_pointer + byte_size)
-        mbchar = line.byteslice(byte_pointer + byte_size, size)
-        break if mbchar.encode(Encoding::UTF_8) =~ /\P{Word}/
-        byte_size += size
-      end
-      after_start = byte_pointer + byte_size
-    elsif mbchar.encode(Encoding::UTF_8) =~ /\p{Word}/
-      # ' aaa bb[cursor]b'
-      byte_size = 0
-      while 0 < (byte_pointer + byte_size)
-        size = get_prev_mbchar_size(line, byte_pointer + byte_size)
-        mbchar = line.byteslice(byte_pointer + byte_size - size, size)
-        break if mbchar.encode(Encoding::UTF_8) =~ /\P{Word}/
-        byte_size -= size
-      end
-      right_word_start = byte_pointer + byte_size
-      byte_size = 0
-      while line.bytesize > (byte_pointer + byte_size)
-        size = get_next_mbchar_size(line, byte_pointer + byte_size)
-        mbchar = line.byteslice(byte_pointer + byte_size, size)
-        break if mbchar.encode(Encoding::UTF_8) =~ /\P{Word}/
-        byte_size += size
-      end
-      after_start = byte_pointer + byte_size
-    else
-      byte_size = 0
-      while (line.bytesize - 1) > (byte_pointer + byte_size)
-        size = get_next_mbchar_size(line, byte_pointer + byte_size)
-        mbchar = line.byteslice(byte_pointer + byte_size, size)
-        break if mbchar.encode(Encoding::UTF_8) =~ /\p{Word}/
-        byte_size += size
-      end
-      if (byte_pointer + byte_size) == (line.bytesize - 1)
-        # ' aaa bbb [cursor] '
-        after_start = line.bytesize
-        while 0 < (byte_pointer + byte_size)
-          size = get_prev_mbchar_size(line, byte_pointer + byte_size)
-          mbchar = line.byteslice(byte_pointer + byte_size - size, size)
-          break if mbchar.encode(Encoding::UTF_8) =~ /\p{Word}/
-          byte_size -= size
-        end
-        while 0 < (byte_pointer + byte_size)
-          size = get_prev_mbchar_size(line, byte_pointer + byte_size)
-          mbchar = line.byteslice(byte_pointer + byte_size - size, size)
-          break if mbchar.encode(Encoding::UTF_8) =~ /\P{Word}/
-          byte_size -= size
-        end
-        right_word_start = byte_pointer + byte_size
-      else
-        # ' aaa [cursor] bbb '
-        right_word_start = byte_pointer + byte_size
-        while line.bytesize > (byte_pointer + byte_size)
-          size = get_next_mbchar_size(line, byte_pointer + byte_size)
-          mbchar = line.byteslice(byte_pointer + byte_size, size)
-          break if mbchar.encode(Encoding::UTF_8) =~ /\P{Word}/
-          byte_size += size
-        end
-        after_start = byte_pointer + byte_size
-      end
-    end
-    byte_size = right_word_start - byte_pointer
-    while 0 < (byte_pointer + byte_size)
-      size = get_prev_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size - size, size)
-      break if mbchar.encode(Encoding::UTF_8) =~ /\p{Word}/
-      byte_size -= size
-    end
-    middle_start = byte_pointer + byte_size
-    byte_size = middle_start - byte_pointer
-    while 0 < (byte_pointer + byte_size)
-      size = get_prev_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size - size, size)
-      break if mbchar.encode(Encoding::UTF_8) =~ /\P{Word}/
-      byte_size -= size
+    gcs = line.byteslice(0, byte_pointer).grapheme_clusters
+    pos = gcs.size
+    gcs += line.byteslice(byte_pointer..).grapheme_clusters
+    pos += 1 while pos < gcs.size && !word_character?(gcs[pos])
+    if pos == gcs.size # 'aaa  bbb [cursor] '
+      pos -= 1 while pos > 0 && !word_character?(gcs[pos - 1])
+      second_word_end = gcs.size
+    else # 'aaa  [cursor]bbb'
+      pos += 1 while pos < gcs.size && word_character?(gcs[pos])
+      second_word_end = pos
+    end
+    pos -= 1 while pos > 0 && word_character?(gcs[pos - 1])
+    second_word_start = pos
+    pos -= 1 while pos > 0 && !word_character?(gcs[pos - 1])
+    first_word_end = pos
+    pos -= 1 while pos > 0 && word_character?(gcs[pos - 1])
+    first_word_start = pos
+
+    [first_word_start, first_word_end, second_word_start, second_word_end].map do |idx|
+      gcs.take(idx).sum(&:bytesize)
     end
-    left_word_start = byte_pointer + byte_size
-    [left_word_start, middle_start, right_word_start, after_start]
   end
 
   def self.vi_big_forward_word(line, byte_pointer)
-    byte_size = 0
-    while (line.bytesize - 1) > (byte_pointer + byte_size)
-      size = get_next_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size, size)
-      break if mbchar =~ /\s/
-      byte_size += size
-    end
-    while (line.bytesize - 1) > (byte_pointer + byte_size)
-      size = get_next_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size, size)
-      break if mbchar =~ /\S/
-      byte_size += size
-    end
-    byte_size
+    gcs = line.byteslice(byte_pointer..).grapheme_clusters
+    nonspaces = gcs.take_while { |c| !space_character?(c) }
+    spaces = gcs.drop(nonspaces.size).take_while { |c| space_character?(c) }
+    nonspaces.sum(&:bytesize) + spaces.sum(&:bytesize)
   end
 
   def self.vi_big_forward_end_word(line, byte_pointer)
-    if (line.bytesize - 1) > byte_pointer
-      size = get_next_mbchar_size(line, byte_pointer)
-      byte_size = size
-    else
-      return 0
-    end
-    while (line.bytesize - 1) > (byte_pointer + byte_size)
-      size = get_next_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size, size)
-      break if mbchar =~ /\S/
-      byte_size += size
-    end
-    prev_byte_size = byte_size
-    while line.bytesize > (byte_pointer + byte_size)
-      size = get_next_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size, size)
-      break if mbchar =~ /\s/
-      prev_byte_size = byte_size
-      byte_size += size
-    end
-    prev_byte_size
+    gcs = line.byteslice(byte_pointer..).grapheme_clusters
+    first = gcs.shift(1)
+    spaces = gcs.take_while { |c| space_character?(c) }
+    nonspaces = gcs.drop(spaces.size).take_while { |c| !space_character?(c) }
+    matched = spaces + nonspaces
+    matched.pop
+    first.sum(&:bytesize) + matched.sum(&:bytesize)
   end
 
   def self.vi_big_backward_word(line, byte_pointer)
-    byte_size = 0
-    while 0 < (byte_pointer - byte_size)
-      size = get_prev_mbchar_size(line, byte_pointer - byte_size)
-      mbchar = line.byteslice(byte_pointer - byte_size - size, size)
-      break if mbchar =~ /\S/
-      byte_size += size
-    end
-    while 0 < (byte_pointer - byte_size)
-      size = get_prev_mbchar_size(line, byte_pointer - byte_size)
-      mbchar = line.byteslice(byte_pointer - byte_size - size, size)
-      break if mbchar =~ /\s/
-      byte_size += size
-    end
-    byte_size
+    gcs = line.byteslice(0, byte_pointer).grapheme_clusters.reverse
+    spaces = gcs.take_while { |c| space_character?(c) }
+    nonspaces = gcs.drop(spaces.size).take_while { |c| !space_character?(c) }
+    spaces.sum(&:bytesize) + nonspaces.sum(&:bytesize)
   end
 
   def self.vi_forward_word(line, byte_pointer, drop_terminate_spaces = false)
-    if line.bytesize > byte_pointer
-      size = get_next_mbchar_size(line, byte_pointer)
-      mbchar = line.byteslice(byte_pointer, size)
-      if mbchar =~ /\w/
-        started_by = :word
-      elsif mbchar =~ /\s/
-        started_by = :space
+    gcs = line.byteslice(byte_pointer..).grapheme_clusters
+    return 0 if gcs.empty?
+
+    c = gcs.first
+    matched =
+      if word_character?(c)
+        gcs.take_while { |c| word_character?(c) }
+      elsif space_character?(c)
+        gcs.take_while { |c| space_character?(c) }
       else
-        started_by = :non_word_printable
-      end
-      byte_size = size
-    else
-      return 0
-    end
-    while line.bytesize > (byte_pointer + byte_size)
-      size = get_next_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size, size)
-      case started_by
-      when :word
-        break if mbchar =~ /\W/
-      when :space
-        break if mbchar =~ /\S/
-      when :non_word_printable
-        break if mbchar =~ /\w|\s/
+        gcs.take_while { |c| !word_character?(c) && !space_character?(c) }
       end
-      byte_size += size
-    end
-    return byte_size if drop_terminate_spaces
-    while line.bytesize > (byte_pointer + byte_size)
-      size = get_next_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size, size)
-      break if mbchar =~ /\S/
-      byte_size += size
-    end
-    byte_size
+
+    return matched.sum(&:bytesize) if drop_terminate_spaces
+
+    spaces = gcs.drop(matched.size).take_while { |c| space_character?(c) }
+    matched.sum(&:bytesize) + spaces.sum(&:bytesize)
   end
 
   def self.vi_forward_end_word(line, byte_pointer)
-    if (line.bytesize - 1) > byte_pointer
-      size = get_next_mbchar_size(line, byte_pointer)
-      mbchar = line.byteslice(byte_pointer, size)
-      if mbchar =~ /\w/
-        started_by = :word
-      elsif mbchar =~ /\s/
-        started_by = :space
-      else
-        started_by = :non_word_printable
-      end
-      byte_size = size
-    else
-      return 0
-    end
-    if (line.bytesize - 1) > (byte_pointer + byte_size)
-      size = get_next_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size, size)
-      if mbchar =~ /\w/
-        second = :word
-      elsif mbchar =~ /\s/
-        second = :space
-      else
-        second = :non_word_printable
-      end
-      second_byte_size = size
-    else
-      return byte_size
-    end
-    if second == :space
-      byte_size += second_byte_size
-      while (line.bytesize - 1) > (byte_pointer + byte_size)
-        size = get_next_mbchar_size(line, byte_pointer + byte_size)
-        mbchar = line.byteslice(byte_pointer + byte_size, size)
-        if mbchar =~ /\S/
-          if mbchar =~ /\w/
-            started_by = :word
-          else
-            started_by = :non_word_printable
-          end
-          break
-        end
-        byte_size += size
-      end
-    else
-      case [started_by, second]
-      when [:word, :non_word_printable], [:non_word_printable, :word]
-        started_by = second
-      else
-        byte_size += second_byte_size
-        started_by = second
-      end
-    end
-    prev_byte_size = byte_size
-    while line.bytesize > (byte_pointer + byte_size)
-      size = get_next_mbchar_size(line, byte_pointer + byte_size)
-      mbchar = line.byteslice(byte_pointer + byte_size, size)
-      case started_by
-      when :word
-        break if mbchar =~ /\W/
-      when :non_word_printable
-        break if mbchar =~ /[\w\s]/
-      end
-      prev_byte_size = byte_size
-      byte_size += size
-    end
-    prev_byte_size
+    gcs = line.byteslice(byte_pointer..).grapheme_clusters
+    return 0 if gcs.empty?
+    return gcs.first.bytesize if gcs.size == 1
+
+    start = gcs.shift
+    skips = [start]
+    if space_character?(start) || space_character?(gcs.first)
+      spaces = gcs.take_while { |c| space_character?(c) }
+      skips += spaces
+      gcs.shift(spaces.size)
+    end
+    start_with_word = word_character?(gcs.first)
+    matched = gcs.take_while { |c| start_with_word ? word_character?(c) : !word_character?(c) && !space_character?(c) }
+    matched.pop
+    skips.sum(&:bytesize) + matched.sum(&:bytesize)
   end
 
   def self.vi_backward_word(line, byte_pointer)
-    byte_size = 0
-    while 0 < (byte_pointer - byte_size)
-      size = get_prev_mbchar_size(line, byte_pointer - byte_size)
-      mbchar = line.byteslice(byte_pointer - byte_size - size, size)
-      if mbchar =~ /\S/
-        if mbchar =~ /\w/
-          started_by = :word
-        else
-          started_by = :non_word_printable
-        end
-        break
-      end
-      byte_size += size
-    end
-    while 0 < (byte_pointer - byte_size)
-      size = get_prev_mbchar_size(line, byte_pointer - byte_size)
-      mbchar = line.byteslice(byte_pointer - byte_size - size, size)
-      case started_by
-      when :word
-        break if mbchar =~ /\W/
-      when :non_word_printable
-        break if mbchar =~ /[\w\s]/
-      end
-      byte_size += size
-    end
-    byte_size
+    gcs = line.byteslice(0, byte_pointer).grapheme_clusters.reverse
+    spaces = gcs.take_while { |c| space_character?(c) }
+    gcs.shift(spaces.size)
+    start_with_word = word_character?(gcs.first)
+    matched = gcs.take_while { |c| start_with_word ? word_character?(c) : !word_character?(c) && !space_character?(c) }
+    spaces.sum(&:bytesize) + matched.sum(&:bytesize)
   end
 
   def self.common_prefix(list, ignore_case: false)
@@ -647,15 +398,18 @@ class Reline::Unicode
   end
 
   def self.vi_first_print(line)
-    byte_size = 0
-    while (line.bytesize - 1) > byte_size
-      size = get_next_mbchar_size(line, byte_size)
-      mbchar = line.byteslice(byte_size, size)
-      if mbchar =~ /\S/
-        break
-      end
-      byte_size += size
-    end
-    byte_size
+    gcs = line.grapheme_clusters
+    spaces = gcs.take_while { |c| space_character?(c) }
+    spaces.sum(&:bytesize)
+  end
+
+  def self.word_character?(s)
+    s.encode(Encoding::UTF_8).match?(/\p{Word}/) if s
+  rescue Encoding::UndefinedConversionError
+    false
+  end
+
+  def self.space_character?(s)
+    s.match?(/\s/) if s
   end
 end

data/lib/reline/version.rb

@@ -1,3 +1,3 @@
 module Reline
-  VERSION = '0.5.12'
+  VERSION = '0.6.1'
 end