Actually, a better test would be to simply try replicating. Here is the deal. There are a lot of data structures already in native datatypes that are subject to byte swapping - e.g. even INT....but definitely datetime, numeric, money, float, etc. Some are not (e.g. varchar/varbinary) - will explain in a minute. As part of the standard OCS locales/charsets/platform support we can replicate datatypes between platform without issue. Part of this is because SRS generally translates params into string form for the SQL - but if using an fstring with output RPC - much as if making a client side RPC call (a common situation for fully prepared statements and proc calls), the client translates the string datavalues into the appropriate datatype in binary form. OCS & ASE together can handle the endianess or otherwise we would have a real mess with PC clients & Unix DBMS servers.
However, a key factor of byte swapping in endianess is based on full bytes vs. swapping bit positions. In other words, in a two byte sequence such as 0x0A, becomes 0xA0 when swapped vs. swapping the bits within the bytes. Generally speaking varchar/varbinary are simply strings of bytes - a single byte for each character. As a result, no byte swapping is required. Even for some multi-byte, things are the same as UTF-8 using multiple single bytes for each character. The only one I am not sure about is 16-bit character sets.....however, like I said, I am pretty sure the OCS layer handles this.