cm_utf8: add an is_valid function

2 files changed, 23 insertions, 0 deletions

diff --git a/Source/cm_utf8.c b/Source/cm_utf8.c
index d41d097..62e7e8c 100644
--- a/Source/cm_utf8.c
+++ b/Source/cm_utf8.c
@@ -2,6 +2,8 @@
    file Copyright.txt or https://cmake.org/licensing for details.  */
 #include "cm_utf8.h"
 
+#include <string.h>
+
 /*
   RFC 3629
   07-bit: 0xxxxxxx
@@ -85,3 +87,20 @@ const char* cm_utf8_decode_character(const char* first, const char* last,
     return first;
   }
 }
+
+int cm_utf8_is_valid(const char* s)
+{
+  if (!s) {
+    return 0;
+  }
+
+  const char* last = s + strlen(s);
+  const char* pos = s;
+  unsigned int pc;
+
+  while (pos != last && (pos = cm_utf8_decode_character(pos, last, &pc))) {
+    /* Nothing to do. */
+  }
+
+  return pos == last;
+}
diff --git a/Source/cm_utf8.h b/Source/cm_utf8.h
index fcb43e0..27dc559 100644
--- a/Source/cm_utf8.h
+++ b/Source/cm_utf8.h
@@ -13,6 +13,10 @@ extern "C" {
 const char* cm_utf8_decode_character(const char* first, const char* last,
                                      unsigned int* pc);
 
+/** Returns whether a C string is a sequence of valid UTF-8 encoded Unicode
+    codepoints.  Returns non-zero on success. */
+int cm_utf8_is_valid(const char* s);
+
 #ifdef __cplusplus
 } /* extern "C" */
 #endif