I'm in the process of implementing a data conversion function for one compound type to another. In this particular case, the destination compound type has two additional fields and 4 that have been renamed.
Ideally, all I'd need to do is intelligently copy the data from the old field name to the new field name, but as I study the documentation it's not really clear to me if I need to account for endian-ness of each individual field. Each atomic type I use is defined as a native type.
If, for example, I write a file on a Sun where the old compound type is used, and the compound type contains native atomic types, when I read that file using a new compound type on a PC, where the new compound type is also defined using native types, do I need to account for endianness translation in the type conversion function that translates the old compound to the new compound?
Hopefully I phrased that well enough.