From: Kenichi Handa <handa@gnu.org>
Date: Tue, 20 Aug 2013 14:56:03 +0000 (+0300)
Subject: src/character.c (string_char): Improve commentary.
X-Git-Url: https://git.hcoop.net/bpt/emacs.git/commitdiff_plain/3f246b657225c786c460b22d774ab0b2e7488b55

src/character.c (string_char): Improve commentary.
---

diff --git a/src/ChangeLog b/src/ChangeLog
index e21d82bdc0..997459f01c 100644
--- a/src/ChangeLog
+++ b/src/ChangeLog
@@ -1,3 +1,7 @@
+2013-08-20  Kenichi Handa  <handa@gnu.org>
+
+	* character.c (string_char): Improve commentary.
+
 2013-08-20  Paul Eggert  <eggert@cs.ucla.edu>
 
 	* image.c (SIGNATURE_DIGESTSIZE): Remove.
diff --git a/src/character.c b/src/character.c
index b2caaa290a..1bde2364e3 100644
--- a/src/character.c
+++ b/src/character.c
@@ -174,11 +174,14 @@ string_char (const unsigned char *p, const unsigned char **advanced, int *len)
 
   if (*p < 0x80 || ! (*p & 0x20) || ! (*p & 0x10))
     {
+      /* 1-, 2-, and 3-byte sequences can be handled by the macro.  */
       c = STRING_CHAR_ADVANCE (p);
     }
   else if (! (*p & 0x08))
     {
-      c = ((((p)[0] & 0xF) << 18)
+      /* A 4-byte sequence of this form:
+	 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx  */
+      c = ((((p)[0] & 0x7) << 18)
 	   | (((p)[1] & 0x3F) << 12)
 	   | (((p)[2] & 0x3F) << 6)
 	   | ((p)[3] & 0x3F));
@@ -186,7 +189,14 @@ string_char (const unsigned char *p, const unsigned char **advanced, int *len)
     }
   else
     {
-      c = ((((p)[1] & 0x3F) << 18)
+      /* A 5-byte sequence of this form:
+
+	 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
+
+	 Note that the top 4 `x's are always 0, so shifting p[1] can
+	 never exceed the maximum valid character codepoint. */
+      c = (/* (((p)[0] & 0x3) << 24) ... always 0, so no need to shift. */
+	   (((p)[1] & 0x3F) << 18)
 	   | (((p)[2] & 0x3F) << 12)
 	   | (((p)[3] & 0x3F) << 6)
 	   | ((p)[4] & 0x3F));