Re: Finding unused hdf data

On 04/13/09 Culley Harrelson uttered the following other thing:
> On Mon, 13 Apr 2009, David Jeske wrote:
> 
> > 
> > It seems like it might be practical to have a CS evaluation mode that hits a
> > callback for every HDF variable accessed. It would need to evaluate all
> > conditional paths. Even then it would have issues like I mentioned above,
> > however, this may be minimal in your case. This concept sounds vaguely familiar
> > to me, but I'm not sure why. Any thoughts Brandon?
> > 
> 
> Sounds very useful to me!
> 
> I was planning to do this at the file level-- basically something like
> grepping through the cs file for each item.  This would probably be enough
> for my case because I have decent naming conventions for hdf items.  I have
> some big hdf files that are used globally and they are much more
> problematic...  An evaluation mode would be more thorough.

Its the every conditional part thats challenging, but perhaps... it
would certainly be easy enough to walk the parse tree and find nodes
which use raw hdf names, though the challenge is in handling things like
each which would need to have HDF nodes to walk, nodes which are likely
specific to the page and not just static HDF.

Brandon
-- 
  "The difference between death and taxes is death doesn't get worse 
   every time Congress meets." -- Will Rogers
                                           http://www.fiction.net/blong/