Add ASSUMEs for UTF-8 byte lengths

[khwilliamson]

The maximum number of bytes in a Perl extended UTF-8 character is 13 on ASCII platforms; 14 on EBCDIC. Yet the variable that returns that number is a Size_t. By adding these clues to these inline functions, the compiler may be able to do some optimizations.

This isn't done here on another inline function, utf8_to_uv_msgs(). That is because it currently returns the call of a non-inline function, so the ASSUME would be unreachable code. I don't know if that actually matters. Or that function's boolean result could be stored in a temporary the ASSUME done, and then utf8_to_uv_msgs() would return the temporary's value.

Opinions welcome

• This set of changes does not require a perldelta entry.

[tonycoz]

I do wonder if assuming expectlen >= 1 (etc) is useful.

[khwilliamson]

Revised to do check that the value is 1<= x <= MAX

[khwilliamson]

@tonycoz Any ideas on the second paragraph of #23680 (comment)

[tonycoz]

This isn't done here on another inline function, utf8_to_uv_msgs(). That is because it currently returns the call of a non-inline function, so the ASSUME would be unreachable code. I don't know if that actually matters. Or that function's boolean result could be stored in a temporary the ASSUME done, and then utf8_to_uv_msgs() would return the temporary's value.

I think there's some value in:

bool result = somfunc(...advancep);
ASSUME(advancep == NULL || inRANGE(...));
return result;

also, while -fanalyze -flto is impractical, -flto is practical, so ASSUME()s in non-inline functions can have some value.

[khwilliamson]

I added an ASSUME in a place where it would only help link time optimization, and changed as suggested in #23680 (comment).

I also audited the code base for other potential spots to change, and found none.