RE: Paremeterized $table_model file name

most of the discrete time (event driven or "verilog") file operators
allow this already..
we should allow string parameters, AND string localparam's  to be used
in these contexts..
(ie not just the $table_model, but in other places where names or
strings might be used..
(format strings for $strobe/$display/) 

And while we are at it, why not allow the (some) of the same format
characters as are used in $fopen.. 
for the file path, or module name? 
so the file name could be "%M_tunecurve.dat" where %M is the instance
name.. 
or %C_tunecurve.dat where %C is the module name.. (did I get that
right?)
I'll hope so cause I don't want to open any more docs tonight.


Jonathan David   Mixed-Signal IC 
jbdavid@cadence.com Ph (408)894-2646

-----Original Message-----
From: owner-verilog-ams@eda.org [mailto:owner-verilog-ams@eda.org] On
Behalf Of Muranyi, Arpad
Sent: Wednesday, August 03, 2005 1:21 PM
To: VerilogAMS Reflector
Subject: Paremeterized $table_model file name

Geoffrey, 

Yes, what you say about the macro used as file name works
in the simulators I tried.  However, I wonder whether the
usage of string parameters as file names is legal or not
according to the spec.  If not, may I make a request to
add it to the spec ASAP?  It seems that since string
parameters are constants, this should still work without
any complications.

Thanks,

Arpad
=========================================================

-----Original Message-----
From: geoffrey.coram@analog.com [mailto:geoffrey.coram@analog.com] 
Sent: Tuesday, July 26, 2005 10:14 AM
To: Muranyi, Arpad
Cc: VerilogAMS Reflector
Subject: Re: define and strings

Arpad -

> By the way, can this string literal come from a macro then?

I would expect that this would work:

`define FILENAME "mytable.txt"

I(a,b) <+ $table_model(`FILENAME, V(a,b), ...);

because the preprocessor would place "mytable.txt" inside the
$table_model call before the syntax is checked.  This does not
conflict with the earlier contention that "$table_model can only
take a double quote enclosed string literal."

> And if string parameters are treated a constants which can
> only change at compile time, why couldn't it come from a
> string parameter?

I believe that parameters can change at *elaboration* time, 
which is subtly different from compile time.  But, as pointed
out before, it is an obvious extension to allow a string
parameter as the first argument to $table_model.


-Geoffrey