> > To answer yesterday's question, what uses GENERIC? Those that invoke
> > GENERIC in the makefile:
> >
> > CCDPACK, DATACUBE, GENERIC (DCV), IMG, JCMTDR, POLPACK
> >
>
> JCMTDR surprises me. Especially since I thought I had the source code and
It was just a quick grep through makefiles, and not a comprehensive
list. The JCMTDR makefile defines GENERIC conversions but doesn't
actually use it. If I'd missed packages, people could then `chip in'.
> SURF doesn't (or at least it may have done but John Lightfoot never told
> me about it if he did!).
I did put a query against it. It does access PRIMDAT generic
definitions like NUM_DEC_UW.
> NDF definitely does [there are some wacky warnings when you compile NDF
> because it has lines like 'if ('C' .EQ. 'C')' in it (e.g. ndf_xgt0x).]
This was an accidental omission. I didn't find it in the grep, but I
did know that NDF has type dependent routines. Rod bundled all the
instantiations into ndf_<whatever>x.f.
> I've already put ARY in CVS so if you have the .gen files then we should
> put them in. I assume that the final tar balls (even the source code tar
> balls) will have the code, not the .gen files).
OK, now I see where you're coming from. You want the actual source,
not the files after they've been through GENERIC. Whether or not a
package comes with .gen files or expanded versions is package
dependent, and the supporting programmers should have the .gen files.
If not, we may have tarballs lurking around RAL, say if we have backups
of Rod's and Alan's files. I do have older .gen files for KAPPA and
CONVERT. Many of these won't have changed.
> applications/generic sounds good to me. [it's no different to messgen]
OK will do.
> Adding hdstrace to CVS should be fairly straightforward and a good
> learning experience.
For you or me?
Malcolm
|