Hello,

I try to write my code in a generic fashion,
and this causes that there are cases in which I have signals with null 
range.
for example, if I have a signal :
SIGNAL a : std_logic_vector(X-1 downto 0);
then if X is 0, this is a null range vector, and this is ok.
when I use this way of writing in altera-quartus and modelsim, there is 
no problem and I simply get a null range warning.
but it appears that mentor-precision doesn't support this.

do you know if null range is a vhdl feature that must be supported?


Thanks
Ravid

Maybe something like that could help:

1

signal M : integer range 15 downto 1;

Where "15" is just an example to have a complete statement.
Your vector then has a minimal length of 1. Smaller vectors (besides 
from an academic point of view) make not much sense, and you'll find a 
way to design around them.