On Fri, 16 Dec 2005, Mark Taylor wrote:
> 1. Will datGetVC() be forthcoming? If not that's OK, I'll just drop
> it from the java interface, it's not currently used anywhere.
>
> 2. I think the third argument of datRef() should be a size_t *
> and not a size_t?
Leave those bits for Tim.
> 3. DAT__NOLOC and DAT__ROOT are still char *s (#defined as character
> literals). Shouldn't they be HDSLocs?
No. Don't use these values in C. Both these are now the same as a NULL
HDSLoc locator.
> 4. What include files am I supposed to be using? I presume only
> hds.h and hds_types.h.
That's right.
> However, doing this I can't allocate
> an HDSLoc, since it doesn't know what sizeof(HDSLoc) is.
> That is defined in dat1.h, but dat1.h is not installed under
> /star/include.
Actual HDSLocs are created and returned (and freed) by HDS, you should
just need to deal in pointers to these, there's no other storage needed
now.
> dat1.h also contains the types _BYTE, _WORD etc,
> which I want to use to find the sizes of the various mapped HDS
> types.
Don't we just now work with the idea that _BYTE = char,_WORD = short,
_REAL = float, _INTEGER = int etc? Might be a fuzzy area (there are some
old compilers with int == 2 bytes)
Cheers,
Peter.
|