• Y.Tomino (25/25) Dec 02 2003 Hello.
• Walter (5/29) Dec 03 2003 to
• Y.Tomino (10/42) Dec 04 2003 Thanks!
• Sean L. Palmer (9/55) Dec 04 2003 I would not want it to be limited to 256 bits though. I would want to b...

"Y.Tomino" <demoonlit inter7.jp> writes:

Hello.

It was written on the document that "on Intel CPUs it would be rounded up to
the next 32 bit size. ".
bit[1] .. bit[32] should be 4 byte. Is that bug ?

int main()
{
 bit[8] a8;
 printf("%d\n", a8.size); //4 ok
 bit[16] a16;
 printf("%d\n", a16.size); //5
 bit[32] a32;
 printf("%d\n", a32.size); //7
 bit[256] a256;
 printf("%d\n", a256.size); //35 (256/8 = 32)
 return 0;
}

And, I want to use "&" "|" "^" "!" operators for bit array like Pascal's
Set.
bit[8] a = [3:1, 4:1];
bit[8] b = [4:1, 5:1];
a[] & b[] = [4:1]
a[] | b[] = [3:1, 4:1, 5:1]
a[] ^ b[] = [3:1, 5:1]
!a[] = [0:1, 1:1, 2:1, 5:1, 6:1, 7:1]

YT

"Walter" <walter digitalmars.com> writes:

"Y.Tomino" <demoonlit inter7.jp> wrote in message
news:bqjeh3$lmb$1 digitaldaemon.com...
 Hello.

 It was written on the document that "on Intel CPUs it would be rounded up

to
 the next 32 bit size. ".
 bit[1] .. bit[32] should be 4 byte. Is that bug ?

 int main()
 {
  bit[8] a8;
  printf("%d\n", a8.size); //4 ok
  bit[16] a16;
  printf("%d\n", a16.size); //5
  bit[32] a32;
  printf("%d\n", a32.size); //7
  bit[256] a256;
  printf("%d\n", a256.size); //35 (256/8 = 32)
  return 0;
 }

That looks like a bug. The results should be 4,4,4,32.

 And, I want to use "&" "|" "^" "!" operators for bit array like Pascal's
 Set.
 bit[8] a = [3:1, 4:1];
 bit[8] b = [4:1, 5:1];
 a[] & b[] = [4:1]
 a[] | b[] = [3:1, 4:1, 5:1]
 a[] ^ b[] = [3:1, 5:1]
 !a[] = [0:1, 1:1, 2:1, 5:1, 6:1, 7:1]

That's supposed to work, but I haven't implemented it yet.

"Y.Tomino" <demoonlit inter7.jp> writes:

Thanks!
I'm waiting for correcting these.

I'll use fixed-length bit array as set of flags like window styles (case
where unsigned int is used at C),
or bit[255] as cache of IsLeadBytes or ...
It will be powerful tool.

YT

"Walter" <walter digitalmars.com> wrote in message
news:bqlmp5$upi$1 digitaldaemon.com...
 "Y.Tomino" <demoonlit inter7.jp> wrote in message
 news:bqjeh3$lmb$1 digitaldaemon.com...
 Hello.

 It was written on the document that "on Intel CPUs it would be rounded


up
 to
 the next 32 bit size. ".
 bit[1] .. bit[32] should be 4 byte. Is that bug ?

 int main()
 {
  bit[8] a8;
  printf("%d\n", a8.size); //4 ok
  bit[16] a16;
  printf("%d\n", a16.size); //5
  bit[32] a32;
  printf("%d\n", a32.size); //7
  bit[256] a256;
  printf("%d\n", a256.size); //35 (256/8 = 32)
  return 0;
 }

 That looks like a bug. The results should be 4,4,4,32.

 And, I want to use "&" "|" "^" "!" operators for bit array like Pascal's
 Set.
 bit[8] a = [3:1, 4:1];
 bit[8] b = [4:1, 5:1];
 a[] & b[] = [4:1]
 a[] | b[] = [3:1, 4:1, 5:1]
 a[] ^ b[] = [3:1, 5:1]
 !a[] = [0:1, 1:1, 2:1, 5:1, 6:1, 7:1]

 That's supposed to work, but I haven't implemented it yet.

"Sean L. Palmer" <palmer.sean verizon.net> writes:

I would not want it to be limited to 256 bits though.  I would want to be
able to make associative arrays of bit.

And I *still* want to be able to convert bit array slices into integers, and
integers into bit array slices.  That would eliminate the need for bitfields
altogether, I think.  There's a whole thread on this somewhere on this NG.

Sean

"Y.Tomino" <demoonlit inter7.jp> wrote in message
news:bqn8c3$76r$1 digitaldaemon.com...
 Thanks!
 I'm waiting for correcting these.

 I'll use fixed-length bit array as set of flags like window styles (case
 where unsigned int is used at C),
 or bit[255] as cache of IsLeadBytes or ...
 It will be powerful tool.

 YT

 "Walter" <walter digitalmars.com> wrote in message
 news:bqlmp5$upi$1 digitaldaemon.com...
 "Y.Tomino" <demoonlit inter7.jp> wrote in message
 news:bqjeh3$lmb$1 digitaldaemon.com...
 Hello.

 It was written on the document that "on Intel CPUs it would be rounded


 up
 to
 the next 32 bit size. ".
 bit[1] .. bit[32] should be 4 byte. Is that bug ?

 int main()
 {
  bit[8] a8;
  printf("%d\n", a8.size); //4 ok
  bit[16] a16;
  printf("%d\n", a16.size); //5
  bit[32] a32;
  printf("%d\n", a32.size); //7
  bit[256] a256;
  printf("%d\n", a256.size); //35 (256/8 = 32)
  return 0;
 }

 That looks like a bug. The results should be 4,4,4,32.

 And, I want to use "&" "|" "^" "!" operators for bit array like



Pascal's
 Set.
 bit[8] a = [3:1, 4:1];
 bit[8] b = [4:1, 5:1];
 a[] & b[] = [4:1]
 a[] | b[] = [3:1, 4:1, 5:1]
 a[] ^ b[] = [3:1, 5:1]
 !a[] = [0:1, 1:1, 2:1, 5:1, 6:1, 7:1]

 That's supposed to work, but I haven't implemented it yet.