Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.c++ > #4635 > unrolled thread

literals for an int128 class

Back to article view | Back to comp.lang.c++

Contents

  literals for an int128 class cps <chrissusie8@gmail.com> - 2011-05-03 13:39 -0700
    Re: literals for an int128 class Victor Bazarov <v.bazarov@comcast.invalid> - 2011-05-03 16:51 -0400
    Re: literals for an int128 class Ian Collins <ian-news@hotmail.com> - 2011-05-04 08:53 +1200
    Re: literals for an int128 class Pete Becker <pete@versatilecoding.com> - 2011-05-03 17:17 -0400
    Re: literals for an int128 class Michael Doubez <michael.doubez@free.fr> - 2011-05-04 01:38 -0700
      Re: literals for an int128 class Ian Collins <ian-news@hotmail.com> - 2011-05-04 21:15 +1200
        Re: literals for an int128 class Michael Doubez <michael.doubez@free.fr> - 2011-05-06 01:03 -0700
          Re: literals for an int128 class Pete Becker <pete@versatilecoding.com> - 2011-05-06 08:00 -0400
            Re: literals for an int128 class Victor Bazarov <v.bazarov@comcast.invalid> - 2011-05-06 09:30 -0400
            Re: literals for an int128 class Michael Doubez <michael.doubez@free.fr> - 2011-05-06 08:00 -0700

#4635 — literals for an int128 class

I am considering writing an int128 class as a programming exercise.
Before I start (and I have a fair idea of how to do it) I was
wondering what I can do about literals and the assignment operator.  I
would like to be able to use integer literals with the assignment and
other operators with literals that are sized at 128 bits.

E.g. I'd like to be able to do:

     uint128 myInt128 = 184467440737095516160;


     Note that the foregoing literal would produce an error (correctly
so) if I attempted to assign it to an uint64 because it is (2^64) *
10.

     Is there any way I can set it up to deal with large literals?

    Or am I going to have to use strings and parse them?

[toc] | [next] | [standalone]

#4636

On 5/3/2011 4:39 PM, cps wrote:
> I am considering writing an int128 class as a programming exercise.
> Before I start (and I have a fair idea of how to do it) I was
> wondering what I can do about literals and the assignment operator.

Judging from the excerpt below, you're talking about the symbol '=' used 
for initialization of objects at the time of their definition, and *NOT* 
about the assignment operator.

 >  I
> would like to be able to use integer literals with the assignment and
> other operators with literals that are sized at 128 bits.
>
> E.g. I'd like to be able to do:
>
>       uint128 myInt128 = 184467440737095516160;

That is called initialization.

>       Note that the foregoing literal would produce an error (correctly
> so) if I attempted to assign it to an uint64 because it is (2^64) *
> 10.
>
>       Is there any way I can set it up to deal with large literals?

Not portably.  Literals are defined by the standard/compiler and unless 
your compiler already provides a 128 bit signed or unsigned integer 
literal, you can't introduce it.  If 'uint128' is totally your own 
custom class, then your best bet is to introduce a constructor from a 
character string (uint128::uint128(const char*)) and surround those 
digits with double quotes:

    uint128 myInt128("184467440737095516160");

*Some* compilers can offer built-in integral types of that size and with 
those they usually offer a way to specify a literal by means of some 
suffix.  For instance, MS used to have i64 suffix before 64 bit targets 
became common:

    int64_t ms64bitInteger = 1234567890i64; // if memory serves

>      Or am I going to have to use strings and parse them?

Most likely.

T
-- 
I do not respond to top-posted replies, please don't ask

[toc] | [prev] | [next] | [standalone]

#4637

On 05/ 4/11 08:39 AM, cps wrote:
> I am considering writing an int128 class as a programming exercise.
> Before I start (and I have a fair idea of how to do it) I was
> wondering what I can do about literals and the assignment operator.  I
> would like to be able to use integer literals with the assignment and
> other operators with literals that are sized at 128 bits.
>
> E.g. I'd like to be able to do:
>
>       uint128 myInt128 = 184467440737095516160;
>
>
>       Note that the foregoing literal would produce an error (correctly
> so) if I attempted to assign it to an uint64 because it is (2^64) *
> 10.
>
>       Is there any way I can set it up to deal with large literals?
>
>      Or am I going to have to use strings and parse them?

The only way a system could have 128 bit literals would be if it had a 
128 bit type!

You'll have to settle for strings.

-- 
Ian Collins

[toc] | [prev] | [next] | [standalone]

#4639

On 2011-05-03 16:39:35 -0400, cps said:

> I am considering writing an int128 class as a programming exercise.
> Before I start (and I have a fair idea of how to do it) I was
> wondering what I can do about literals and the assignment operator.  I
> would like to be able to use integer literals with the assignment and
> other operators with literals that are sized at 128 bits.
> 
> E.g. I'd like to be able to do:
> 
>      uint128 myInt128 = 184467440737095516160;
> 
> 
>      Note that the foregoing literal would produce an error (correctly
> so) if I attempted to assign it to an uint64 because it is (2^64) *
> 10.
> 
>      Is there any way I can set it up to deal with large literals?
> 
>     Or am I going to have to use strings and parse them?

You can't do 128 bit literals if your compiler doesn't support them, 
but you can do something that's a little easier than parsing a string.

struct uint128 {
	uint128(unsigned long long word1, unsigned long long word0);
};

uint128 i0(0x21, 0x03);
uint128 i1(0x0, 3);



-- 
  Pete
Roundhouse Consulting, Ltd. (www.versatilecoding.com) Author of "The 
Standard C++ Library Extensions: a Tutorial and Reference 
(www.petebecker.com/tr1book)

[toc] | [prev] | [next] | [standalone]

#4661

On 3 mai, 22:39, cps <chrissus...@gmail.com> wrote:
> I am considering writing an int128 class as a programming exercise.
> Before I start (and I have a fair idea of how to do it) I was
> wondering what I can do about literals and the assignment operator.  I
> would like to be able to use integer literals with the assignment and
> other operators with literals that are sized at 128 bits.
>
> E.g. I'd like to be able to do:
>
>      uint128 myInt128 = 184467440737095516160;
>
>      Note that the foregoing literal would produce an error (correctly
> so) if I attempted to assign it to an uint64 because it is (2^64) *
> 10.
>
>      Is there any way I can set it up to deal with large literals?
>
>     Or am I going to have to use strings and parse them?

Parsing it from string is however a good idea.

With C++0x, you will be able to write:

uint128  operator "" ulll(const char * string_values, size_t
num_chars)
{
  return uint128::parse(string_values,num_chars);
}

uint128 myInt128 = 184467440737095516160ulll;

--
Michael

[toc] | [prev] | [next] | [standalone]

#4663

On 05/ 4/11 08:38 PM, Michael Doubez wrote:
> On 3 mai, 22:39, cps<chrissus...@gmail.com>  wrote:
>> I am considering writing an int128 class as a programming exercise.
>> Before I start (and I have a fair idea of how to do it) I was
>> wondering what I can do about literals and the assignment operator.  I
>> would like to be able to use integer literals with the assignment and
>> other operators with literals that are sized at 128 bits.
>>
>> E.g. I'd like to be able to do:
>>
>>       uint128 myInt128 = 184467440737095516160;
>>
>>       Note that the foregoing literal would produce an error (correctly
>> so) if I attempted to assign it to an uint64 because it is (2^64) *
>> 10.
>>
>>       Is there any way I can set it up to deal with large literals?
>>
>>      Or am I going to have to use strings and parse them?
>
> Parsing it from string is however a good idea.
>
> With C++0x, you will be able to write:
>
> uint128  operator "" ulll(const char * string_values, size_t
> num_chars)
> {
>    return uint128::parse(string_values,num_chars);
> }
>
> uint128 myInt128 = 184467440737095516160ulll;

Neat.  In which section of the standard is this described?

-- 
Ian Collins

[toc] | [prev] | [next] | [standalone]

#4764

On 4 mai, 11:15, Ian Collins <ian-n...@hotmail.com> wrote:
> On 05/ 4/11 08:38 PM, Michael Doubez wrote:
>
>
>
>
>
>
>
>
>
> > On 3 mai, 22:39, cps<chrissus...@gmail.com>  wrote:
> >> I am considering writing an int128 class as a programming exercise.
> >> Before I start (and I have a fair idea of how to do it) I was
> >> wondering what I can do about literals and the assignment operator.  I
> >> would like to be able to use integer literals with the assignment and
> >> other operators with literals that are sized at 128 bits.
>
> >> E.g. I'd like to be able to do:
>
> >>       uint128 myInt128 = 184467440737095516160;
>
> >>       Note that the foregoing literal would produce an error (correctly
> >> so) if I attempted to assign it to an uint64 because it is (2^64) *
> >> 10.
>
> >>       Is there any way I can set it up to deal with large literals?
>
> >>      Or am I going to have to use strings and parse them?
>
> > Parsing it from string is however a good idea.
>
> > With C++0x, you will be able to write:
>
> > uint128  operator "" ulll(const char * string_values, size_t
> > num_chars)
> > {
> >    return uint128::parse(string_values,num_chars);
> > }
>
> > uint128 myInt128 = 184467440737095516160ulll;
>
> Neat.  In which section of the standard is this described?

In the current draft (n3242)
2.14.8 User-defined literals [lex.ext]

--
Michael

[toc] | [prev] | [next] | [standalone]

#4771

On 2011-05-06 04:03:13 -0400, Michael Doubez said:

> 
> In the current draft (n3242)
> 

The current draft is N3291.

-- 
  Pete
Roundhouse Consulting, Ltd. (www.versatilecoding.com) Author of "The 
Standard C++ Library Extensions: a Tutorial and Reference 
(www.petebecker.com/tr1book)

[toc] | [prev] | [next] | [standalone]

#4775

On 5/6/2011 8:00 AM, Pete Becker wrote:
> On 2011-05-06 04:03:13 -0400, Michael Doubez said:
>
>>
>> In the current draft (n3242)
>>
>
> The current draft is N3291.
>

Huh...

One would expect it be linked to on the Committee's page

    http://www.open-std.org/jtc1/sc22/wg21/

, yet the link still points to n3242...  Just to let you know...  In 
case you somehow can influence what's put on that web page...

V
-- 
I do not respond to top-posted replies, please don't ask

[toc] | [prev] | [next] | [standalone]

#4779

On 6 mai, 14:00, Pete Becker <p...@versatilecoding.com> wrote:
> On 2011-05-06 04:03:13 -0400, Michael Doubez said:
>
> > In the current draft (n3242)
>
> The current draft is N3291.

Thanks. I just grabbed it from WG21 page.

Is there any hope that c++1x or C++2x will make it to N4242 ?

--
Michael

[toc] | [prev] | [standalone]

Back to top | Article view | comp.lang.c++

csiph-web