[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [f-cpu] Re: genereic_adder patch


Michael Riepe wrote:
> On Sun, Sep 16, 2001 at 02:14:35AM +0200, Yann Guidon wrote:
> > Michael Riepe wrote:
> > > On Sun, Sep 16, 2001 at 01:43:56AM +0200, Yann Guidon wrote:
> > > > Here is a tiny modification to Michael's generic adder package.
> > > > The newest version of simili doesn't like the way the agregate
> > > > was used, but it is not too difficult to work around it.
> > > Yep... I had to do this several times, too.
> > i'm waiting for your updated "MR official" version :-)
> That will take a while.  I'm quite busy at the moment (next month's iX
> issue has to be finished), and I wanted to do a little more than just
> change one line before I make a new release.

so i assume that my patched version will hold until you release another.
by the way, wouldn't it be great if you separated the testbench
from the ciadd package ? So if you can't compile the testbench
at least you could compile the package...

> [...]
> > > > As written before, Simili accepts this :
> > > >   Signal s : std_ulogic_vector(N downto M) := random(s);
> > > You mean, you use `s' before the definition is finished?
> > > That looks broken, somehow.
> > that #should# work. i don't remember why and how, or where i have
> > read it. At least Simili does it.
> And what is it supposed to do?
Since at that point the compiler knows the new name and type,
   Signal s : std_ulogic_vector(N downto M) := random(s);
means that you send a std_ulogic_vector(N downto M) to random().
So you don't have to repeat the declarations.
Unfortunately there is no way that i know for a function
to read the type and range of the final recipient.

> [...]
> > mmmmm overloading comes to rescue us :-)
> > that's what i call irony :-P
> > [i mean : it is sad that a complex and misleading feature comes
> >  to relieve you from the non-implementation of a simple feature]
> Isn't it great? ;)
a pity.

but who am i to say that, i'm only a newbie.

>  Michael "Tired" Riepe <Michael.Riepe@stud.uni-hannover.de>

PS : back at sound compression/processing, i am dusting off
the 8-tap real Winograd FP FFT code i wrote 3 months ago.
I am curious to see what i can do when i put a JPEG-like DCT
on top of it :*)
To unsubscribe, send an e-mail to majordomo@seul.org with
unsubscribe f-cpu       in the body. http://f-cpu.seul.org/