Re: improved QR & Cholesky updating

[Jaroslav Hajek]

On Wed, Jan 21, 2009 at 8:41 PM, John W. Eaton <address@hidden> wrote:
> On 21-Jan-2009, Jaroslav Hajek wrote:
>
> | True, but it doesn't catch a warning, does it?
>
> Are there cases where missing functions only generate warnings?
>

Not really missing. What I meant is that after I fix this issue,
without qrupdate the functions will be present but sub-optimal. Maybe
some code will want to detect the situation. Now that I think of it,
how does "testif" detect that HAVE_QRUPDATE has been defined? Can this
be done directly from an m-file? If yes, then the above last argument
is dead...

> | OK, so where should the replacements go?
> | Putting them in the QR & Cholesky classes means writing replacement
> | code for 4 * 8 + 4 * 5 = 52 methods (OK, In fact only quarter and then
> | copy-paste-search-replace-adjust). I suppose writing generic
> | replacements directly in the DLD funcs would be simpler. Also, I'm
> | thinking of issuing a warning the first time any of the updating
> | function is run, only once.
> |
> | What do you think?
>
> I think the best place is in the QR::update method (for example),
> because that way, any code that might use that class won't have to
> work around missing functionality either.  Can templates or macros
> help here?
>

OK. I originally considered rebasing QR & CHOL classes on a common
templated ancestor, but eventually I realized there was really not
that much code to share, as most methods are relatively thin wrappers
for the Fortran ones.
Still I can write a .cc file containing the replacement templates and
use that even in the current code.

cheers

-- 
RNDr. Jaroslav Hajek
computing expert
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz