Re: ANN: Ada/SPARK ASN.1 implementation version 0.0.01

[Shark8 <onewingedshark@gmail.com>]

On Monday, August 1, 2016 at 1:47:39 PM UTC-6, Jeffrey R. Carter wrote:
> On 08/01/2016 01:15 AM, Shark8 wrote:
> > 
> > I would certainly appropriate comments, criticism, and most especially contributions.
> 
> I took a look at the big-number pkg, and have the following suggestions:
> 
> If you want people to contribute to this effort, I'd suggest you use a coding
> style more similar to that used in the ARM. You might also want to learn how to
> spell "negative".

Corrected spelling.
Also, "coding style" should **NOT** be an issue -- it is a testament to the CS industry being stunted that it is an issue. (The 'primary' store of source should be within meaningful structures, likely in a DB, not plain=text.)

> I don't see why it's desirable to specify the representation of the non-variable
> parts of Big_Number.
> 
> I'd think it would be a good idea for code like this to be portable, in which
> case it should avoid optional types such as Long_Long_Integer, and certainly not
> assume it knows that it's 64 bits.

Agreed -- however this *is* the 0.0.01 release, and not intended to be the be-all/end-all of source code. (Again, collaboration is welcome.)

> While it's unlikely to be an issue on modern processors, given your constant
> 
> Base : constant := 2 ** System.Word_Size;
> 
> there's no guarantee that declarations
> 
> type Digit is range 0 .. Base;
> 
> type Double_Digit is mod Base ** 2;
> 
> will be accepted. A more portable approach, and one which maximizes the size of
> the digits used, is
> 
> type Double_Digit is mod System.Max_Binary_Modulus;
> 
> Digit_Size : constant := Double_Digit'Size / 2;
> 
> type Digit is mod Digit_Size;
> 
> This is the approach used by PragmARC.Unbounded_Integers.
> 
> HTH.


Thank you for the pointers.