Re: What happens to DEC Ada?

["Marin David Condic" <marin.condic.auntie.spam@pacemicro.com>]

In my experience, a body of code has to get *pretty large* before a
recompile of the entire thing starts to become painful. (Especially with
today's high speed processors.) It has to get even larger than that before
the customary Gnat timestamp method of determining what needs to be
recompiled becomes painful. IMHO, effort expended to avoid a few more files
that might not actually need to be recompiled isn't worth it. I'd
concentrate on bigger problems and count on hardware to keep pace with the
size of the project. :-)

MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com
Web:      http://www.mcondic.com/


"Robert Dewar" <dewar@gnat.com> wrote in message
news:5ee5b646.0108031100.2be1c7d6@posting.google.com...
>
> Well I guess this can seem counter-cultural if you are not familiar
> with GNAT technology, but in fact the notion of smart recompilation
> fits perfectly well. Indeed GNAT *has* a (very limited) notion of
> smart recompilation (you can use the appropriate option, look in the
> documentation, to request a mode in which reformatting, addition or
> deletion of comments etc, is ignored and does not cause recompiles).
>
> It would be technically feasible to do more, but simply not technically
> worthwhile. Smart recompilation adds basic overhead to all compiles,
> so overall it would not be a win. There are of course special cases
> where it might help, but overall, our analysis is that it is a loss.