X-Git-Url: http://review.tizen.org/git/?a=blobdiff_plain;f=src%2Futf8.c;h=d37ba8e301d7efb7a07dd63d2e26e29d33893836;hb=5b3774ace991a396752ff0a846fdfb5c38424551;hp=a7fa82e271f2bcdaf35c3f6d5c9abc9105145eb7;hpb=767fa86d42a5e25e7043622d189247e02a5ca379;p=platform%2Fupstream%2Flibxkbcommon.git

diff --git a/src/utf8.c b/src/utf8.c
index a7fa82e..d37ba8e 100644
--- a/src/utf8.c
+++ b/src/utf8.c
@@ -24,12 +24,19 @@
  * Author: Rob Bradford <rob@linux.intel.com>
  */
 
+#include "config.h"
+
 #include <stddef.h>
 #include <stdbool.h>
 #include <inttypes.h>
 
 #include "utf8.h"
 
+/* Conformant encoding form conversion from UTF-32 to UTF-8.
+ *
+ * See https://www.unicode.org/versions/Unicode15.0.0/ch03.pdf#G28875
+ * for further details.
+*/
 int
 utf32_to_utf8(uint32_t unichar, char *buffer)
 {
@@ -45,21 +52,20 @@ utf32_to_utf8(uint32_t unichar, char *buffer)
         length = 2;
         head = 0xc0;
     }
+    /* Handle surrogates */
+    else if (0xd800 <= unichar && unichar <= 0xdfff) {
+        goto ill_formed_code_unit_subsequence;
+    }
     else if (unichar <= 0xffff) {
         length = 3;
         head = 0xe0;
     }
-    else if (unichar <= 0x1fffff) {
+    else if (unichar <= 0x10ffff) {
         length = 4;
         head = 0xf0;
     }
-    else if (unichar <= 0x3ffffff) {
-        length = 5;
-        head = 0xf8;
-    }
     else {
-        length = 6;
-        head = 0xfc;
+        goto ill_formed_code_unit_subsequence;
     }
 
     for (count = length - 1, shift = 0; count > 0; count--, shift += 6)
@@ -69,6 +75,10 @@ utf32_to_utf8(uint32_t unichar, char *buffer)
     buffer[length] = '\0';
 
     return length + 1;
+
+ill_formed_code_unit_subsequence:
+    buffer[0] = '\0';
+    return 0;
 }
 
 bool