From 91343a8328b0b13cdc111443fe7ace2dd1765166 Mon Sep 17 00:00:00 2001
From: Colin Patrick McCabe <cmccabe@alumni.cmu.edu>
Date: Thu, 14 Jul 2011 15:07:47 -0700
Subject: [PATCH] Add control character detection

Signed-off-by: Colin McCabe <colin.mccabe@dreamhost.com>
---
 src/common/utf8.c | 17 +++++++++++++++++
 src/common/utf8.h |  9 +++++++++
 src/test/utf8.cc  | 17 +++++++++++++++++
 3 files changed, 43 insertions(+)

diff --git a/src/common/utf8.c b/src/common/utf8.c
index 11957b3143564..f6a1b07b5956f 100644
--- a/src/common/utf8.c
+++ b/src/common/utf8.c
@@ -164,3 +164,20 @@ int check_utf8_cstr(const char *buf)
 {
 	return check_utf8(buf, strlen(buf));
 }
+
+int check_for_control_characters(const char *buf, int len)
+{
+	int i;
+	for (i = 0; i < len; ++i) {
+		unsigned char c = (unsigned char)buf[i];
+		if (((c != 0U) && (c < 0x20U)) || (c == 0x7fU)) {
+ 			return i + 1;
+		}
+	}
+	return 0;
+}
+
+int check_for_control_characters_cstr(const char *buf)
+{
+	return check_for_control_characters(buf, strlen(buf));
+}
diff --git a/src/common/utf8.h b/src/common/utf8.h
index e1c891cef5926..e2b25b94c33d3 100644
--- a/src/common/utf8.h
+++ b/src/common/utf8.h
@@ -31,6 +31,15 @@ int check_utf8(const char *buf, int len);
  */
 int check_utf8_cstr(const char *buf);
 
+/* Checks if a buffer contains control characters.
+ * We do count newline as a control character, but not NULL.
+ */
+int check_for_control_characters(const char *buf, int len);
+
+/* Checks if a null-terminated string contains control characters.
+ */
+int check_for_control_characters_cstr(const char *buf);
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/src/test/utf8.cc b/src/test/utf8.cc
index 7c1a4de24eb64..8e22e524ec17f 100644
--- a/src/test/utf8.cc
+++ b/src/test/utf8.cc
@@ -47,3 +47,20 @@ TEST(IsValidUtf8, InvalidUtf8) {
   uint8_t invalid2[] = { 0xc3, 0x28 };
   ASSERT_NE(0, check_utf8((char*)invalid2, sizeof(invalid2)));
 }
+
+TEST(HasControlChars, HasControlChars1) {
+  uint8_t has_control_chars[] = { 0x41, 0x01, 0x00 };
+  ASSERT_NE(0, check_for_control_characters_cstr((const char*)has_control_chars));
+  uint8_t has_control_chars2[] = { 0x7f, 0x41, 0x00 };
+  ASSERT_NE(0, check_for_control_characters_cstr((const char*)has_control_chars2));
+
+  char has_newline[] = "blah   blah\n";
+  ASSERT_NE(0, check_for_control_characters_cstr(has_newline));
+
+  char no_control_chars[] = "blah   blah";
+  ASSERT_EQ(0, check_for_control_characters_cstr(no_control_chars));
+
+  uint8_t validutf[] = { 0x66, 0xd1, 0x86, 0xd1, 0x9d, 0xd2, 0xa0, 0xd3,
+		       0xad, 0xd3, 0xae, 0x0 };
+  ASSERT_EQ(0, check_for_control_characters_cstr((const char*)validutf));
+}
-- 
2.39.5