SQLRPGLE precompiler - allow usage of large VARCHARS by interpreting the same as SQLTYPE(CLOB)

Currently you cannot use large RPGLE VARCHARs (length >= 32705) in EXEC SQL statements.

I suggest updating the SQL precompiler and SQL processing so that they can be.

This would remove the need for defining SQLTYPE(CLOB) in new programs, VARCHARs could be used instead. And as a result, CLOB data would become interoperable with all the regular RPGLE BIFs such as %trim, %scan, %xlate, etc. with no conversion, and no need to reference different variants of the variable name in different circumstances.

A VARCHAR with a 4 byte prefix (either explicitly with VARCHAR(x:4), or implicit with VARCHAR(x) where x > 65535) has an identical structure in memory to the SQLTYPE DS (4 bytes unsigned int for length, then x bytes character), and the SQL precompiler, when it converts SQL statements, references with with %ADDR so I think this would be a pure precompiler change?

Supporting VARCHARs with a 2 byte prefix would I guess be more complicated as the SQL processing that references the address must know if it has a 2 byte or 4 byte prefix and act accordingly.

This would apply to the other LOB types too, not just CLOB.

This is Intended as an improvement on this other IDEA: https://ibm-power-systems.ideas.ibm.com/ideas/IBMI-I-3732 which suggests the ability to optionally have the SQL precompiler convert SQLTYPE(CLOB) to a VARCHAR(x:4) instead of to the UNS(10)/CHAR(x) DS. The difference being that IDEA might only be SQL Precompiler changes, whereas this one is more complicated.

Thanks to Scott Klement, and JonBoy of https://code400.com/forum, for the discussion - sorry I forgot about it for so long!
https://code400.com/forum/forum/iseries-programming-languages/rpg-rpgle/151378-safe-to-pointer-overlay-a-varchar-onto-a-clob