shared/utf8: add utf8_is_valid_n()

Sometimes we need to check strings without the terminating NUL. Add a variant
that does that.

src/basic/utf8.c

@@ -150,18 +150,22 @@ bool utf8_is_printable_newline(const char* str, size_t length, bool allow_newlin
         return true;
 }
 
-char *utf8_is_valid(const char *str) {
-        const char *p;
+char *utf8_is_valid_n(const char *str, size_t len_bytes) {
+        /* Check if the string is composed of valid utf8 characters. If length len_bytes is given, stop after
+         * len_bytes. Otherwise, stop at NUL. */
 
         assert(str);
 
-        p = str;
-        while (*p) {
+        for (const char *p = str; len_bytes != (size_t) -1 ? (size_t) (p - str) < len_bytes : *p != '\0'; ) {
                 int len;
 
-                len = utf8_encoded_valid_unichar(p, (size_t) -1);
-                if (len < 0)
-                        return NULL;
+                if (_unlikely_(*p == '\0') && len_bytes != (size_t) -1)
+                        return NULL; /* embedded NUL */
+
+                len = utf8_encoded_valid_unichar(p,
+                                                 len_bytes != (size_t) -1 ? len_bytes - (p - str) : (size_t) -1);
+                if (_unlikely_(len < 0))
+                        return NULL; /* invalid character */
 
                 p += len;
         }

src/basic/utf8.h

@@ -14,7 +14,10 @@
 
 bool unichar_is_valid(char32_t c);
 
-char *utf8_is_valid(const char *s) _pure_;
+char *utf8_is_valid_n(const char *str, size_t len_bytes) _pure_;
+static inline char *utf8_is_valid(const char *s) {
+        return utf8_is_valid_n(s, (size_t) -1);
+}
 char *ascii_is_valid(const char *s) _pure_;
 char *ascii_is_valid_n(const char *str, size_t len);