Re: gEDA-user: Verilog 2001

[Hagen Sankowski <office-at-NOSPAManydroid_dot_de>]

Hello Matt and Stephen!

Am Donnerstag, 13.02.03 um 01:48 Uhr schrieb Stephen Williams:

> Some of Verilog-2001 is already supported, including signed arithmetic,
> declaration assignments, port declaration lists, and other items I 
> forget.

> Arrays of primitive gates (i.e. bufif drv [31:0]) already work fine.
> Arrays of modules are harder, but someday I'll get to them. Generate
> statements would not be far behind module instance arrays, but they
> are not on my to-do list for 0.8. I agree that working generate
> syntax would be extremely kool, and is not as hard as it looks.
>
> I'm also keeping an eye on SystemVerilog developments, and will be
> bringing in some bits soon. Integral types in particular are next on
> my to-do list.

Nice deja vu.
Just yesterday I feed icarus with Verilog 2001 files to see how well it 
would be supported. My last attempt was with 0.6 and horrible. Now, 
with your 0.7 for MacOS X, I see just a working Verilog 1995 Edition 
and your -g switch on command line. Why you run far away to 
SystemVerilog with your Integral types and lay still down needed 
Verilog 2001 support in case of configuration / generate?

I'm a hardware design engineer and working with gates. I'm not a 
software loving engineer with need of sophisticated types. My focus is 
real hardware stuff near to silicon. In my opinion it's important to 
finish one standard edition well before starting the next one. Sorry 
but your nice -g switch is missleading I didn't know if one feature 
from edition will supported or not.

How about a documentation on features and best usage in Icarus Verilog? 
Did you need any help I could give? My implementation style stresses 
Icarus very well, but how many guys working on icarus software to 
follow the hints?

Regards
Hagen Sankowski