junklib: Fix the regression caused by the previous commit / failing test.

Added a new utf8 string copy function `u8_strnbcpy_size`, that can handle zero-separated multi-value strings.

Author: Oleksiy-Yakovenko

src/junklib.c

@@ -552,12 +552,8 @@ ddb_iconv (const char *cs_out, const char *cs_in, char *out, int outlen, const c
     // to utf8 branch
     if (!strcasecmp (cs_out, UTF8_STR)) {
         if (!strcasecmp (cs_in, UTF8_STR)) {
-            int result = u8_strnbcpy(out, in, outlen-1);
-            out[result] = 0;
-            int valid = u8_valid (out, result, NULL);
-            if (valid) {
-                len = result;
-            }
+            int result = u8_strnbcpy_size(out, in, inlen, outlen);
+            len = result;
         }
         else if (!strcasecmp (cs_in, "cp1251")) {
             len = cp1251_to_utf8 (in, inlen, out, outlen);

src/utf8.c

@@ -278,6 +278,40 @@ int u8_strnbcpy (char *dest, const char* src, int num_bytes) {
     return nb - num_bytes;
 }
 
+/* Copies as many characters from src as would not exceed num_bytes in the destination */
+/* Uses input_size as the input size. */
+/* \0 characters are copied as-is. */
+/* returns the number of bytes copied. */
+int u8_strnbcpy_size (char *dest, const char* src, int input_size, int num_bytes) {
+    int32_t prev_index = 0;
+    int32_t index = 0;
+    int32_t nb = num_bytes;
+
+    while (index <= input_size && num_bytes > 0 && dest) {
+        if (src[index] == 0) {
+            *dest++ = 0;
+            num_bytes--;
+            prev_index = ++index;
+            continue;
+        }
+        u8_inc (src, &index);
+        if (index > input_size) {
+            break;
+        }
+        int32_t charlen = index - prev_index;
+        if (charlen > num_bytes) {
+            break;
+        }
+        memcpy (dest, &src[prev_index], charlen);
+        prev_index = index;
+        dest += charlen;
+        num_bytes -= charlen;
+    }
+    return nb - num_bytes;
+}
+
+
+
 /* copies a character from src to dest provided it does not exceed num_bytes */
 /* returns the number of bytes copied, not counting a null terminator, which is not written */
 int u8_charcpy (char *dest, const char *src, int num_bytes) {

src/utf8.h

@@ -66,6 +66,13 @@ int u8_strncpy (char *dest, const char* src, int num_chars);
  */
 int u8_strnbcpy (char *dest, const char* src, int num_bytes);
 
+/* Copies as many characters from src as would not exceed num_bytes in the destination;
+ Uses input_size as the input size.
+ \0 characters are copied as-is.
+ returns the number of bytes copied.
+*/
+int u8_strnbcpy_size (char *dest, const char* src, int input_size, int num_bytes);
+
 /* copy single utf8 character of up to num_bytes bytes large, only if num_bytes is large enough;
   return number of bytes copied
  */