Use cheaper string constructors

* src/coding.c (make_string_from_utf8):
* src/fileio.c (file_name_directory):
* src/lread.c (Fintern):
* src/gnutls.c (Fgnutls_format_certificate):
Use `make_multibyte_string` and `make_unibyte_string` instead of
more expensive `make_specified_string` and `make_string_from_bytes`.

src/coding.c

@@ -6365,7 +6365,7 @@ make_string_from_utf8 (const char *text, ptrdiff_t nbytes)
   /* If TEXT is a valid UTF-8 string, we can convert it to a Lisp
      string directly.  Otherwise, we need to decode it.  */
   if (chars == nbytes || bytes == nbytes)
-    return make_specified_string (text, chars, nbytes, true);
+    return make_multibyte_string (text, chars, nbytes);
   else
     {
       struct coding_system coding;

src/fileio.c

@@ -523,7 +523,7 @@ file_name_directory (Lisp_Object filename)
   else
     {
       dostounix_filename (beg);
-      tem_fn = make_specified_string (beg, -1, p - beg, 0);
+      tem_fn = make_unibyte_string (beg, p - beg);
     }
   SAFE_FREE ();
   return tem_fn;

src/gnutls.c

@@ -1646,8 +1646,7 @@ string representation.  */)
 	     emacs_gnutls_strerror (err));
     }
 
-  Lisp_Object result = make_string_from_bytes ((char *) out.data, out.size,
-					       out.size);
+  Lisp_Object result = make_unibyte_string ((char *) out.data, out.size);
   gnutls_free (out.data);
   gnutls_x509_crt_deinit (crt);
 

src/lread.c

@@ -5025,8 +5025,8 @@ it defaults to the value of `obarray'.  */)
     {
       if (longhand)
 	{
-	  tem = intern_driver (make_specified_string (longhand, longhand_chars,
-						      longhand_bytes, true),
+	  tem = intern_driver (make_multibyte_string (longhand, longhand_chars,
+						      longhand_bytes),
 			       obarray, tem);
 	  xfree (longhand);
 	}