From 8f991575ea44e073698f45290fa469c781fe8146 Mon Sep 17 00:00:00 2001
From: g <g@dbf4865f-1ec6-0310-8412-f61adeb1ccb1>
Date: Mon, 1 Sep 2003 23:04:59 +0000
Subject: [PATCH] Use the "magic" version that does not assume that the string
 is in ASCII.

git-svn-id: file:///srv/svn/gliv/trunk@72 dbf4865f-1ec6-0310-8412-f61adeb1ccb1

committer: g <g@dbf4865f-1ec6-0310-8412-f61adeb1ccb1>
---
 src/str_utils.c | 34 ++++++++++++++++++++--------------
 1 file changed, 20 insertions(+), 14 deletions(-)

diff --git a/src/str_utils.c b/src/str_utils.c
index 1692664..9becb79 100644
--- a/src/str_utils.c
+++ b/src/str_utils.c
@@ -78,7 +78,14 @@ const gchar *add_mnemonic(const gchar * str)
     return *result;
 }
 
-static void init_magic(gulong * himagic, gulong * lomagic)
+/*
+ * The magic to find a '\0' in a long int is taken from the glibc.
+ * See sysdeps/generic/strlen.c in the glibc sources to have the
+ * explanation.
+ * Actually it also finds 0x80, that's why we double check when
+ * HAS_ZERO() finds something.
+ */
+static void init_magic(gulong * himagic, gulong * lomagic, gulong * magic_bits)
 {
     *himagic = 0x80808080L;
     *lomagic = 0x01010101L;
@@ -86,9 +93,13 @@ static void init_magic(gulong * himagic, gulong * lomagic)
         /* 64-bit */
         *himagic = ((*himagic << 16) << 16) | *himagic;
         *lomagic = ((*lomagic << 16) << 16) | *lomagic;
-    }
+        *magic_bits = ((0x7efefefeL << 16) << 16) | 0xfefefeffL;
+    } else
+        *magic_bits = 0x7efefeffL;
 }
 
+#define HAS_ZERO(longint) ((((longint) + magic_bits) ^ ~longint) & ~magic_bits)
+
 #define CHECK_BYTE(ptr)   \
     do {                  \
         if (*ptr == '\0') \
@@ -100,14 +111,9 @@ static void init_magic(gulong * himagic, gulong * lomagic)
         ptr++;            \
     } while (0)
 
-/*
- * The magic to find a '\0' in a long int is taken from the glibc.
- * See sysdeps/generic/strlen.c in the glibc sources to have the
- * explanation.
- */
 static gboolean str_is_ascii(const gchar * str)
 {
-    gulong himagic, lomagic;
+    gulong himagic, lomagic, magic_bits;
     gulong *long_ptr, long_int;
     gulong mask;
 
@@ -126,7 +132,7 @@ static gboolean str_is_ascii(const gchar * str)
     }
 
     long_ptr = (gulong *) str;
-    init_magic(&himagic, &lomagic);
+    init_magic(&himagic, &lomagic, &magic_bits);
     mask = 0x80808080;
     if (sizeof(gulong) > 4)
         /* 64-bit */
@@ -134,7 +140,7 @@ static gboolean str_is_ascii(const gchar * str)
 
     for (;;) {
         long_int = *long_ptr;
-        if ((long_int - lomagic) & himagic) {
+        if (HAS_ZERO(long_int)) {
             /* A '\0' has been detected. */
             const gchar *char_ptr = (const gchar *) long_ptr;
             CHECK_BYTE(char_ptr);
@@ -219,7 +225,7 @@ const gchar *filename_to_utf8(const gchar * str)
 gint count_underscores(const gchar * str)
 {
     gint nb_underscores = 0;
-    gulong himagic, lomagic;
+    gulong himagic, lomagic, magic_bits;
     gulong *long_ptr, long_int;
     gulong mask;
 
@@ -238,7 +244,7 @@ gint count_underscores(const gchar * str)
     }
 
     long_ptr = (gulong *) str;
-    init_magic(&himagic, &lomagic);
+    init_magic(&himagic, &lomagic, &magic_bits);
 
     /* '_' == 0x5F */
     mask = 0x5F5F5F5F;
@@ -248,7 +254,7 @@ gint count_underscores(const gchar * str)
 
     for (;;) {
         long_int = *long_ptr;
-        if ((long_int - lomagic) & himagic) {
+        if (HAS_ZERO(long_int)) {
             /* A '\0' has been detected. */
             const gchar *char_ptr = (const gchar *) long_ptr;
             CHECK_UNDERSCORE(char_ptr);
@@ -266,7 +272,7 @@ gint count_underscores(const gchar * str)
             gulong masked;
 
             masked = long_int ^ mask;
-            if ((masked - lomagic) & himagic) {
+            if (HAS_ZERO(masked)) {
                 /* A '_' has been detected. */
                 gchar *char_ptr = (gchar *) & mask;
                 nb_underscores += (*char_ptr++ == 0);
-- 
2.11.4.GIT