> > (Although I see from all the CVS commits that Malcolm is doing a really
> > great job on dragging the FIGARO monster into the 64bit age which must
> > be a herculean task...).
There were more. Some of the nightly logs didn't appear. Yes Herculean
it is. It's going to take a fair bit more. The heterogeneous code,
even within sub-packages, makes for a lot more work.
> Yes, it actually turns out that FIGARO is also broken using the
> DYNAMIC_MEMORY tricks too, these also require sorting out.
I hope Tim is going to help us with the testing.
> > So we would need to replace GROWXY with something else (NDFCOPY??) first.
> Actually, I cannot think of an easy way to create an NDF cube from a stack
> of images (using KAPPA commands). All I came up with is the following
> method:
In answer to Quick Q217, I wrote a C-shell script to form a cube from a
series of same-sized 2-d NDFs. In essence it's the same as Peter
suggests. It pastes in groups of ten NDFs and uses MANIC to create the
cube rather than SETBOUND. It also uses a temporary NDF with bounds
matching the primary NDF (via SETBOUND). It doesn't use SETORIGIN.
I suppose one could use CHAIN and RESHAPE, but it would mess up the WCS.
Malcolm
|