Stop possible underrun when processing a string

Calling String::utf8("Unicode String", -1) assumes that the string will be NULL terminated.
However, the length parameter is always used to find the end of the string.  So there is the
chance the character before th start of the string is read.

Making the pointer NULL in the case where it's out of range, still allows the following
to work as expected
while (ptrtmp != ptrtmp_limit && *ptrtmp)
....

1 files changed, 2 insertions, 2 deletions

diff --git a/core/string/ustring.cpp b/core/string/ustring.cpp
index a7e12138f2..d0d600007b 100644
--- a/core/string/ustring.cpp
+++ b/core/string/ustring.cpp
@@ -1822,7 +1822,7 @@ Error String::parse_utf8(const char *p_utf8, int p_len, bool p_skip_cr) {
 	bool decode_failed = false;
 	{
 		const char *ptrtmp = p_utf8;
-		const char *ptrtmp_limit = &p_utf8[p_len];
+		const char *ptrtmp_limit = p_len >= 0 ? &p_utf8[p_len] : nullptr;
 		int skip = 0;
 		uint8_t c_start = 0;
 		while (ptrtmp != ptrtmp_limit && *ptrtmp) {
@@ -2099,7 +2099,7 @@ Error String::parse_utf16(const char16_t *p_utf16, int p_len) {
 	bool decode_error = false;
 	{
 		const char16_t *ptrtmp = p_utf16;
-		const char16_t *ptrtmp_limit = &p_utf16[p_len];
+		const char16_t *ptrtmp_limit = p_len >= 0 ? &p_utf16[p_len] : nullptr;
 		uint32_t c_prev = 0;
 		bool skip = false;
 		while (ptrtmp != ptrtmp_limit && *ptrtmp) {