Re: ANN: Ada 2005 Math Extensions, 20100810 release

["Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de>]

On Sun, 15 Aug 2010 18:06:51 +0100, Simon Wright wrote:

> In this case I had to use Complex_Arrays.Real_Arrays.Real because of the
> sort of problems you indicate.

One part of that problem was that such names became mysteriously hidden
upon instantiation in the generic bodies. The compiler claimed that Real
was not declared in Real_Arrays, or that Real_Arrays was not declared in
Complex_Arrays, or that Real does not match the type Real of Real_Arrays
etc. It was the worst kind of error I ever had. You have 20+ generic units.
Every body compiles. Then in a chain of 20 generic instantiations #15
fails. You try another variant, refactor this, that, rewrite parent's
children to their parameters or reverse. Two weeks later #17 fails. You
cannot test anything before a modification cycle is complete, because first
10 instantiations are always fine. Basically this is how we are accustomed
to work with C++ templates, but not with Ada. It is stable now, but anxiety
flows through my bones each time I try a new GNAT on it...

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de