Commit 006548da91bfe05375ae0e786c6c13e9bad85a40

Edward Thomson 2015-05-29T16:07:51

git__strcasecmp: treat input bytes as unsigned Treat input bytes as unsigned before doing arithmetic on them, lest we look at some non-ASCII byte (like a UTF-8 character) as a negative value and perform the comparison incorrectly.

diff --git a/src/util.c b/src/util.c
index 6bb7d03..b8d6cf5 100644
--- a/src/util.c
+++ b/src/util.c
@@ -173,7 +173,7 @@ int git__strcasecmp(const char *a, const char *b)
 {
 	while (*a && *b && tolower(*a) == tolower(*b))
 		++a, ++b;
-	return (tolower(*a) - tolower(*b));
+	return ((unsigned char)tolower(*a) - (unsigned char)tolower(*b));
 }
 
 int git__strcasesort_cmp(const char *a, const char *b)
@@ -193,7 +193,7 @@ int git__strcasesort_cmp(const char *a, const char *b)
 	}
 
 	if (*a || *b)
-		return tolower(*a) - tolower(*b);
+		return (unsigned char)tolower(*a) - (unsigned char)tolower(*b);
 
 	return cmp;
 }
diff --git a/tests/core/string.c b/tests/core/string.c
index ec95756..c6c2d95 100644
--- a/tests/core/string.c
+++ b/tests/core/string.c
@@ -39,3 +39,41 @@ void test_core_string__2(void)
 	cl_assert(git__strcasesort_cmp("BAR", "foo") < 0);
 	cl_assert(git__strcasesort_cmp("fooBar", "foobar") < 0);
 }
+
+void test_core_string__strcmp(void)
+{
+	cl_assert(git__strcmp("", "") == 0);
+	cl_assert(git__strcmp("foo", "foo") == 0);
+	cl_assert(git__strcmp("Foo", "foo") < 0);
+	cl_assert(git__strcmp("foo", "FOO") > 0);
+	cl_assert(git__strcmp("foo", "fOO") > 0);
+
+	cl_assert(strcmp("rt\303\202of", "rt dev\302\266h") > 0);
+	cl_assert(strcmp("e\342\202\254ghi=", "et") > 0);
+	cl_assert(strcmp("rt dev\302\266h", "rt\303\202of") < 0);
+	cl_assert(strcmp("et", "e\342\202\254ghi=") < 0);
+
+	cl_assert(git__strcmp("rt\303\202of", "rt dev\302\266h") > 0);
+	cl_assert(git__strcmp("e\342\202\254ghi=", "et") > 0);
+	cl_assert(git__strcmp("rt dev\302\266h", "rt\303\202of") < 0);
+	cl_assert(git__strcmp("et", "e\342\202\254ghi=") < 0);
+}
+
+void test_core_string__strcasecmp(void)
+{
+	cl_assert(git__strcasecmp("", "") == 0);
+	cl_assert(git__strcasecmp("foo", "foo") == 0);
+	cl_assert(git__strcasecmp("foo", "Foo") == 0);
+	cl_assert(git__strcasecmp("foo", "FOO") == 0);
+	cl_assert(git__strcasecmp("foo", "fOO") == 0);
+
+	cl_assert(strcasecmp("rt\303\202of", "rt dev\302\266h") > 0);
+	cl_assert(strcasecmp("e\342\202\254ghi=", "et") > 0);
+	cl_assert(strcasecmp("rt dev\302\266h", "rt\303\202of") < 0);
+	cl_assert(strcasecmp("et", "e\342\202\254ghi=") < 0);
+
+	cl_assert(git__strcasecmp("rt\303\202of", "rt dev\302\266h") > 0);
+	cl_assert(git__strcasecmp("e\342\202\254ghi=", "et") > 0);
+	cl_assert(git__strcasecmp("rt dev\302\266h", "rt\303\202of") < 0);
+	cl_assert(git__strcasecmp("et", "e\342\202\254ghi=") < 0);
+}