Integer with leading zeros to string

Tue Sep 28, 2010 4:59 pm   

Hi,

I have the following integer declaration:

signal test_num : integer := 0555;

Now I want to make a string out of it:

signal test_string : string (1 to 4);
begin
test_string <= integer'image(test_num);

Modelsim complains "Array lengths do not match. Left is 4 (1 to 4).
Right is 3 (1 to 3)."

How can I convert test_num (ranging from 0001 to 9999) to a string
correctly taking into account the leading zeros?

Cheers,
hssig

Tue Sep 28, 2010 5:39 pm   

The following should be displayed:


report "This should be displayed " & integer'image(test_num) severity
note;

-> This should be displayed 0555

cheers,
hssig

Tue Sep 28, 2010 7:36 pm   

On Sep 28, 2:59 pm, hssig <hs...@gmx.net> wrote:


Don't forget that the integer doesn't know about leading
zeros. You are entitled to write them in the integer literal,
but they are of course NOT stored in the integer itself. So
you need a format mechanism:

function format(
value : natural; --- the numeric value
width : positive; -- number of characters
leading : character := ' ')
return string --- guarantees to return "width" chars
is
constant img: string := integer'image(value);
variable str: string(1 to width) := (others => leading);
begin
if img'length > width then
report "Format width " & integer'image(width)
& " is too narrow for value " & img
severity warning;
str := (others => '*');
else
str(width+1-img'length to width) := img;
end if;
return str;
end;

....

---- this line should give "0055"
constant N: integer := 55;
report "value is " & format(integer'image(N), 4, '0');

Any use?

Dealing with negative integers is left as an exercise
--
Jonathan Bromley

Tue Sep 28, 2010 7:55 pm   

On Tue, 28 Sep 2010 09:36:39 -0700 (PDT), Jonathan Bromley wrote:


[...]


oops, obviously the first argument should
be N rather than integer'image(N).
Hasty end-of-working-day post.
--
Jonathan Bromley

Tue Sep 28, 2010 8:04 pm   

On Tue, 28 Sep 2010 06:59:09 -0700 (PDT), hssig wrote:


One final afterthought: take a look at
http://www.easics.com/webtools/freesics
for a more general (and very cunning) solution.
--
Jonathan Bromley

Wed Sep 29, 2010 4:23 pm   

Hi Jonathan,

I like you proposals. Thank you very much.

Cheers,
hssig

Wed Sep 29, 2010 5:44 pm   

On Sep 28, 11:36 am, Jonathan Bromley <s...@oxfordbromley.plus.com>
wrote:

VHDL integers do store leading zeroes, up to at least 32 binary bits
total. However, as you stated, standard output formats do not display
them. Just because you initialize an integer with a literal which
happens to be formatted with leading zeroes does not mean that the
display output from that same integer will automatically be formatted
with leading zeroes.

Andy

Wed Sep 29, 2010 6:02 pm   

On Sep 29, 3:44 pm, Andy <jonesa...@comcast.net> wrote:


yeah, guess I didn't say very clearly what I meant :-)

let's try again: there is no information stored in the
integer variable that indicates whether or not its
original textual representation had leading zeros (and,
indeed, no information about any other aspect of its
original textual representation except the integer's
numeric value).

Ho hum....

Jonathan