[ruby/reline] Improve the performance of get_mbchar_width

It is about three times faster to use one big regexp instead of sequential matching. ruby/reline@e36f6c0707
ruby · Dec 4, 2020 · 76cac4c · 76cac4c
1 parent b3e0db8
commit 76cac4c
Show file tree

Hide file tree

Showing 2 changed files with 38 additions and 28 deletions.
diff --git a/lib/reline/unicode.rb b/lib/reline/unicode.rb
@@ -72,20 +72,32 @@ def self.escape_for_print(str)
     }.join
   end
 
+  require 'reline/unicode/east_asian_width'
+
+  MBCharWidthRE = /
+    (?<width_2_1>
+      [#{ EscapedChars.map {|c| "\\x%02x" % c.ord }.join }] (?# ^ + char, such as ^M, ^H, ^[, ...)
+    )
+  | (?<width_3>^\u{2E3B}) (?# THREE-EM DASH)
+  | (?<width_0>^\p{M})
+  | (?<width_2_2>
+      #{ EastAsianWidth::TYPE_F }
+    | #{ EastAsianWidth::TYPE_W }
+    )
+  | (?<width_1>
+      #{ EastAsianWidth::TYPE_H }
+    | #{ EastAsianWidth::TYPE_NA }
+    | #{ EastAsianWidth::TYPE_N }
+    )
+  /x
+
   def self.get_mbchar_width(mbchar)
-    case mbchar.encode(Encoding::UTF_8)
-    when *EscapedChars # ^ + char, such as ^M, ^H, ^[, ...
-      2
-    when /^\u{2E3B}/ # THREE-EM DASH
-      3
-    when /^\p{M}/
-      0
-    when EastAsianWidth::TYPE_A
-      Reline.ambiguous_width
-    when EastAsianWidth::TYPE_F, EastAsianWidth::TYPE_W
-      2
-    when EastAsianWidth::TYPE_H, EastAsianWidth::TYPE_NA, EastAsianWidth::TYPE_N
-      1
+    m = mbchar.encode(Encoding::UTF_8).match(MBCharWidthRE)
+    case
+    when m[:width_2_1], m[:width_2_2] then 2
+    when m[:width_3] then 3
+    when m[:width_0] then 0
+    when m[:width_1] then 1
     else
       nil
     end
@@ -591,5 +603,3 @@ def self.vi_first_print(line)
     [byte_size, width]
   end
 end
-
-require 'reline/unicode/east_asian_width'
diff --git a/lib/reline/unicode/east_asian_width.rb b/lib/reline/unicode/east_asian_width.rb
@@ -1,27 +1,27 @@
 class Reline::Unicode::EastAsianWidth
   # This is based on EastAsianWidth.txt
-  # http://www.unicode.org/Public/13.0.0/ucd/EastAsianWidth.txt
+  # EastAsianWidth.txt
 
   # Fullwidth
-  TYPE_F = /^([#{ %W(
+  TYPE_F = /^[#{ %W(
     \u{3000}
     \u{FF01}-\u{FF60}
     \u{FFE0}-\u{FFE6}
-  ).join }])/
+  ).join }]/
 
   # Halfwidth
-  TYPE_H = /^([#{ %W(
+  TYPE_H = /^[#{ %W(
     \u{20A9}
     \u{FF61}-\u{FFBE}
     \u{FFC2}-\u{FFC7}
     \u{FFCA}-\u{FFCF}
     \u{FFD2}-\u{FFD7}
     \u{FFDA}-\u{FFDC}
     \u{FFE8}-\u{FFEE}
-  ).join }])/
+  ).join }]/
 
   # Wide
-  TYPE_W = /^([#{ %W(
+  TYPE_W = /^[#{ %W(
     \u{1100}-\u{115F}
     \u{231A}-\u{231B}
     \u{2329}-\u{232A}
@@ -136,21 +136,21 @@ class Reline::Unicode::EastAsianWidth
     \u{1FAD0}-\u{1FAD6}
     \u{20000}-\u{2FFFD}
     \u{30000}-\u{3FFFD}
-  ).join }])/
+  ).join }]/
 
   # Narrow
-  TYPE_NA = /^([#{ %W(
+  TYPE_NA = /^[#{ %W(
     \u{0020}-\u{007E}
     \u{00A2}-\u{00A3}
     \u{00A5}-\u{00A6}
     \u{00AC}
     \u{00AF}
     \u{27E6}-\u{27ED}
     \u{2985}-\u{2986}
-  ).join }])/
+  ).join }]/
 
   # Ambiguous
-  TYPE_A = /^([#{ %W(
+  TYPE_A = /^[#{ %W(
     \u{00A1}
     \u{00A4}
     \u{00A7}-\u{00A8}
@@ -330,10 +330,10 @@ class Reline::Unicode::EastAsianWidth
     \u{E0100}-\u{E01EF}
     \u{F0000}-\u{FFFFD}
     \u{100000}-\u{10FFFD}
-  ).join }])/
+  ).join }]/
 
   # Neutral
-  TYPE_N = /^([#{ %W(
+  TYPE_N = /^[#{ %W(
     \u{0000}-\u{001F}
     \u{007F}-\u{00A0}
     \u{00A9}
@@ -1160,5 +1160,5 @@ class Reline::Unicode::EastAsianWidth
     \u{1FBF0}-\u{1FBF9}
     \u{E0001}
     \u{E0020}-\u{E007F}
-  ).join }])/
+  ).join }]/
 end