+Converts a potentially UTF-8 encoded string C<s> of length C<*lenp> into native
+byte encoding. On input, the boolean C<*is_utf8p> gives whether or not C<s> is
+actually encoded in UTF-8.
+
+Unlike L</utf8_to_bytes> but like L</bytes_to_utf8>, this is non-destructive of
+the input string.
+
+Do nothing if C<*is_utf8p> is 0, or if there are code points in the string
+not expressible in native byte encoding. In these cases, C<*is_utf8p> and
+C<*lenp> are unchanged, and the return value is the original C<s>.
+
+Otherwise, C<*is_utf8p> is set to 0, and the return value is a pointer to a
+newly created string containing a downgraded copy of C<s>, and whose length is
+returned in C<*lenp>, updated. The new string is C<NUL>-terminated.
+
+Upon successful return, the number of variants in the string can be computed by
+having saved the value of C<*lenp> before the call, and subtracting the
+after-call value of C<*lenp> from it.
+
+=cut
+
+There is a macro that avoids this function call, but this is retained for
+anyone who calls it with the Perl_ prefix */
+
+U8 *
+Perl_bytes_from_utf8(pTHX_ const U8 *s, STRLEN *lenp, bool *is_utf8p)
+{
+ PERL_ARGS_ASSERT_BYTES_FROM_UTF8;
+ PERL_UNUSED_CONTEXT;
+
+ return bytes_from_utf8_loc(s, lenp, is_utf8p, NULL);
+}
+
+/*
+No = here because currently externally undocumented
+for apidoc bytes_from_utf8_loc
+
+Like C<L</bytes_from_utf8>()>, but takes an extra parameter, a pointer to where
+to store the location of the first character in C<"s"> that cannot be
+converted to non-UTF8.
+
+If that parameter is C<NULL>, this function behaves identically to
+C<bytes_from_utf8>.
+
+Otherwise if C<*is_utf8p> is 0 on input, the function behaves identically to
+C<bytes_from_utf8>, except it also sets C<*first_non_downgradable> to C<NULL>.
+
+Otherwise, the function returns a newly created C<NUL>-terminated string
+containing the non-UTF8 equivalent of the convertible first portion of
+C<"s">. C<*lenp> is set to its length, not including the terminating C<NUL>.
+If the entire input string was converted, C<*is_utf8p> is set to a FALSE value,
+and C<*first_non_downgradable> is set to C<NULL>.
+
+Otherwise, C<*first_non_downgradable> set to point to the first byte of the
+first character in the original string that wasn't converted. C<*is_utf8p> is
+unchanged. Note that the new string may have length 0.
+
+Another way to look at it is, if C<*first_non_downgradable> is non-C<NULL> and
+C<*is_utf8p> is TRUE, this function starts at the beginning of C<"s"> and
+converts as many characters in it as possible stopping at the first one it
+finds that can't be converted to non-UTF-8. C<*first_non_downgradable> is
+set to point to that. The function returns the portion that could be converted
+in a newly created C<NUL>-terminated string, and C<*lenp> is set to its length,
+not including the terminating C<NUL>. If the very first character in the
+original could not be converted, C<*lenp> will be 0, and the new string will
+contain just a single C<NUL>. If the entire input string was converted,
+C<*is_utf8p> is set to FALSE and C<*first_non_downgradable> is set to C<NULL>.
+
+Upon successful return, the number of variants in the converted portion of the
+string can be computed by having saved the value of C<*lenp> before the call,
+and subtracting the after-call value of C<*lenp> from it.