NOTE: The netcdf-hdf mailing list is no longer active. The list archives are made available for historical reasons.
> Quincey Koziol <koziol@xxxxxxxxxxxxx> writes:
>
> >> 1 - Why is H5Tenum_create not used?
> >
> > It certainly could have been, there's no particular advantage one way or
> > the other (currently). In the future, if we allow enums to have non-integer
> > base types (like an enum of floating-point values, with PI, E, etc.
> > defined),
> > we'll encourage more people to use H5Tcreate_enum().
>
> I have tested, and, as you say, either function works. I will use
> H5Tenum_create(). (That's what you meant right? There's not *another*
> function called H5Tcreate_enum, is there? ;-)
:-) No, there's only one function.
> >> 2 - In the above case, how does HDF5 choose the underlying integer
> >> type? Obviously it will be a short, but signed or unsigned? Big,
> >> little, or native endian? Or does it not matter?
> >
> > If you use H5Tcreate(), it uses the smallest "native" [signed] integer
> > that
> > will hold the number of bytes passed in for the size.
>
> Since I am going to use H5Tcreate_enum I will be explicitly providing
> the type, so I don't have to worry about signed and unsigned
> confusion.
Sounds good.
Quincey
netcdf-hdf archives: