pcre2_string_utils: avoid segfault with strlen(NULL)

pcre2(?:_dfa)?_match should return PCRE2_ERROR_NULL if the subject
is NULL, but the order that is done is incorrect, leading to crashes.

Workaround the issue by allowing strlen(NULL) to return a value
of 0, so it wouldn't segfault while trying to access a NULL subject
and therefore allowing the current check to be reached even if it is
done after the length of the subject is evaluated because it was
provided as PCRE2_ZERO_TERMINATED.

As a side effect, this also prevents crashes in pcre2_substitute
when the subject or the replacement string were NULL and the length
was provided as PCRE2_ZERO_TERMINATED and that would come out handy
if we want to be able to allow a NULL subject as valid when a length
of 0 was also provided.

doc/pcre2_substitute.3

@@ -53,7 +53,7 @@ changed to the length of the new string, excluding the trailing zero that is
 automatically added.
 .P
 The subject and replacement lengths can be given as PCRE2_ZERO_TERMINATED for
-zero-terminated strings. The options are:
+zero-terminated strings. if used with a replacement string of NULL, then it is assumed to be equivalent to the behaviour expected from a replacement string of NULL and a length of 0. The options are:
 .sp
   PCRE2_ANCHORED             Match only at the first position
   PCRE2_ENDANCHORED          Pattern can match only at end of subject

src/pcre2_string_utils.c

@@ -209,6 +209,7 @@ PCRE2_SIZE
 PRIV(strlen)(PCRE2_SPTR str)
 {
 PCRE2_SIZE c = 0;
+if (str == NULL) return 0;
 while (*str++ != 0) c++;
 return c;
 }