Handle unterminated strings when encoding.

If the null terminator is not present, string will be limited to the
data size of the field.

If you are still using the pb_enc_string (deprecated since 0.1.3) from
callbacks, now would be an excellent time to stop. The pb_field_t for
the callback will not contain proper data_size. Use pb_encode_string()
instead.

Update issue 68
Status: FixedInGit

1 files changed, 10 insertions, 2 deletions

diff --git a/pb_encode.c b/pb_encode.c
index 48a3c95..0e048ac 100644
--- a/pb_encode.c
+++ b/pb_encode.c
@@ -461,8 +461,16 @@ bool checkreturn pb_enc_bytes(pb_ostream_t *stream, const pb_field_t *field, con
 
 bool checkreturn pb_enc_string(pb_ostream_t *stream, const pb_field_t *field, const void *src)
 {
-    UNUSED(field);
-    return pb_encode_string(stream, (const uint8_t*)src, strlen((const char*)src));
+    /* strnlen() is not always available, so just use a for-loop */
+    size_t size = 0;
+    const char *p = (const char*)src;
+    while (size < field->data_size && *p != '\0')
+    {
+        size++;
+        p++;
+    }
+
+    return pb_encode_string(stream, (const uint8_t*)src, size);
 }
 
 bool checkreturn pb_enc_submessage(pb_ostream_t *stream, const pb_field_t *field, const void *src)