RE: EXTERNAL: Re: [vhdl-200x] A compromise about modular type, boolean operations, integers...

From: <whygee@f-cpu.org>
Date: Fri Oct 31 2014 - 04:06:56 PDT
Hello,

Trying to empty my "to reply" list...

Le 2014-10-27 02:38, Jones, Andy D a écrit :
>>> Here we have i4 + i4. As long as the + integer operator does not
>>> overflow, the result can be correctly reduced to mod4 during 
>>> assignation.
> 
>> Sorry, but not.  The above code is valid VHDL, and will produce an 
>> error
>> during simulation if overflow occurs.  I don't think we want to change 
>> the
>> semantic of existing VHDL design.
> 
> I believe the point was that the overflow does not occur in "+", but
> in the assignment to another i4.  If it had been assigned to an i8,
> then there would be no overflow.

there is another point I have not seen mde (and not directly related to 
Andy's post) :

even when using vanilla integer, the "+" operator must know when it 
overflows
because the result must be arithmetically valid.
So there is already a built-in capability to handle this, which can be 
tuned
to different values to restrict the range for example.
That's what I'd like to reuse.

> The "+" operator only knows the type (integer) of the operands and
> result. It does not know the subtypes.

That's a grey area for me, an interesting point too.
Too much implementation-dependent but I don't know the language.

yg

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Received on Fri Oct 31 04:07:07 2014

This archive was generated by hypermail 2.1.8 : Fri Oct 31 2014 - 04:08:03 PDT