Help! I don't want a bignum

ruby-talk
1

Ok, I’m sure there is an easy way round this, but I can’t see it…

I want to do this…

syscall 88,-18751827,672274793,1126301404

Which is the kernel sys_reboot call to power-off.

However, ruby kindly converts the last number, 1126301404 (0x4321fedc in
hex which fits nicely in a long) to a Bignum, but syscall only accepts a
Fixnum

Any idea how to get around this???

Andrew Walrond

2

Nobody has a solution?? Surely I haven’t stumped you all :wink:

Andrew Walrond wrote:

···

Ok, I’m sure there is an easy way round this, but I can’t see it…

I want to do this…

syscall 88,-18751827,672274793,1126301404

Which is the kernel sys_reboot call to power-off.

However, ruby kindly converts the last number, 1126301404 (0x4321fedc in
hex which fits nicely in a long) to a Bignum, but syscall only accepts a
Fixnum

Any idea how to get around this???

Andrew Walrond

3

Hi,

···

At Tue, 27 May 2003 23:46:32 +0900, Andrew Walrond wrote:

However, ruby kindly converts the last number, 1126301404 (0x4321fedc
in hex which fits nicely in a long) to a Bignum, but syscall only
accepts a Fixnum

syscall should accept also Bignums.

Index: io.c

RCS file: /cvs/ruby/src/ruby/io.c,v
retrieving revision 1.209
diff -u -2 -p -r1.209 io.c
— io.c 21 May 2003 17:49:18 -0000 1.209
+++ io.c 27 May 2003 16:27:58 -0000
@@ -3450,4 +3450,7 @@ rb_f_syscall(argc, argv)
arg[i] = (unsigned long)NUM2LONG(*argv);
}

  • else if (!SPECIAL_CONST_P(*argv) && BUILTIN_TYPE(*argv) == T_BIGNUM) {
  •   arg[i] = rb_big2ulong(*argv);
  • }
    else {
    VALUE v = *argv;

–
Nobu Nakada

4

Hi,

···

In message “Help! I don’t want a bignum…” on 03/05/27, Andrew Walrond andrew@walrond.org writes:

syscall 88,-18751827,672274793,1126301404

Which is the kernel sys_reboot call to power-off.

However, ruby kindly converts the last number, 1126301404 (0x4321fedc in
hex which fits nicely in a long) to a Bignum, but syscall only accepts a
Fixnum

Any idea how to get around this???

It’s a bug. I will fix it soon. Ask me for a patch, if you’re in
hurry.

						matz.
5

“Andrew Walrond” andrew@walrond.org wrote in message

However, ruby kindly converts the last number, 1126301404 (0x4321fedc in
hex which fits nicely in a long) to a Bignum, but syscall only accepts a
Fixnum

Any idea how to get around this???

I don’t know, because on my (32-bit?) Win XP Pro:

···

C:>ruby -ve “puts((2**31).class)”
ruby 1.6.8 (2002-12-24) [i586-mswin32]
Bignum

C:>ruby -e “puts((2**30).class)”
Bignum

C:>ruby -e “puts((2**30 - 1).class)”
Fixnum

C:>ruby -e “puts(2**30 - 1)”
1073741823

C:>ri Fixnum

 class: Fixnum  < Integer

 A Fixnum holds Integer values that can be represented in a native
 machine word (minus 1 bit). If any operation on a Fixnum exceeds
 this range, the value is automatically converted to a Bignum.
 Fixnum objects have immediate value. This means that when they are
 assigned or passed as parameters, the actual object is passed,
 rather than a reference to that object. Assignment does not alias
 Fixnum objects. There is effectively only one Fixnum object
 instance for any given integer value, so, for example, you cannot
 add a singleton method to a Fixnum.

 <=>, Arithmeticoperations, Bitoperations, [], abs, div, divmod,
 id2name, modulo, quo, size, to_f, to_s, to_sym, zero?

What gives ?
– shanko

6

Nobody has a solution?? Surely I haven’t stumped you all :wink:

Write your own C extension that accepts Bignums.
Of course there should be a better solution.

Andrew Walrond wrote:

Ok, I’m sure there is an easy way round this, but I can’t see it…

I want to do this…

syscall 88,-18751827,672274793,1126301404

Which is the kernel sys_reboot call to power-off.

However, ruby kindly converts the last number, 1126301404 (0x4321fedc in
hex which fits nicely in a long) to a Bignum, but syscall only accepts a
Fixnum

Any idea how to get around this???

Andrew Walrond

Regards,

Michael

···

On Wed, May 28, 2003 at 07:14:04AM +0900, Andrew Walrond wrote:

–
ruby -r complex -e ‘c,m,w,h=Complex(-0.75,0.136),50,150,100;puts “P6\n#{w} #{h}\n255”;(0…h).each{|j|(0…w).each{|i|
n,z=0,Complex(.9i/w,.9j/h);while n<=m&&(z-c).abs<=2;z=zz+c;n+=1 end;print [10+n15,0,rand99].pack("C")}}’|display

7

“Andrew Walrond” andrew@walrond.org schrieb im Newsbeitrag
news:3ED3D146.5030009@walrond.org…

Nobody has a solution?? Surely I haven’t stumped you all :wink:

No, we’re waiting for the first poster to shoot him down then. Ooops, that
might be me then… :slight_smile:

My guess would be, that you might have found an error in ‘syscall’ and / or
Fixnum and / or conversion between number literals and objects.

Maybe someone more intimate with the ruby source (matz?) could comment on
this.

Regards

robert
···

Andrew Walrond wrote:

Ok, I’m sure there is an easy way round this, but I can’t see it…

I want to do this…

syscall 88,-18751827,672274793,1126301404

Which is the kernel sys_reboot call to power-off.

However, ruby kindly converts the last number, 1126301404 (0x4321fedc in
hex which fits nicely in a long) to a Bignum, but syscall only accepts a
Fixnum

Any idea how to get around this???

Andrew Walrond

8

Yukihiro Matsumoto wrote:

It’s a bug. I will fix it soon. Ask me for a patch, if you’re in
hurry.

Thought so :wink: Although I don’t expect syscall gets used too often.

I need compatibility with current ruby versions, so I’m going with the
extension option, but it would be good to get it fixed. Nobu’s patch
looks good.

Andrew Walrond

9

Shashank Date wrote:

I don’t know, because on my (32-bit?) Win XP Pro:

C:>ruby -ve “puts((2**31).class)”
ruby 1.6.8 (2002-12-24) [i586-mswin32]
Bignum

C:>ruby -e “puts((2**30).class)”
Bignum

C:>ruby -e “puts((2**30 - 1).class)”
Fixnum

C:>ruby -e “puts(2**30 - 1)”
1073741823

What gives ?
– shanko

This is the correct behaviour. The valid range for a fixnum is -(230)
to (230)-1

For instance, if your machine had an 8bit word, then…

8bits considered as unsigned can hold 0 to 255 (0 to (28)-1)
8bits considered as signed can hold -128 to 127 (-(27) to (2**7)-1)

But ruby reserves one of those bits, so we only have 7 to play with,
giving a valid range of -(26) to (26)-1, where 6 is the number of
bits in our word minus 2

Now back to our realistic 32bit word; 6 becomes 32-2 = 30 and our valid
range is -(230) to (230)

Read up on twos complement if this makes no sense. Hope this helps :wink:

Andrew Walrond

10

Michael Neumann wrote:

Write your own C extension that accepts Bignums.
Of course there should be a better solution.

Already done it :wink: And as this was the first time I’d done this I was
amazed to have a working extension inside 10 minutes! I am more
impressed with this language every day.

11

“Andrew Walrond” andrew@walrond.org wrote in message

Read up on twos complement if this makes no sense. Hope this helps :wink:

This makes perfect sense … I was ignoring the negative range (doh!) and
hence was confused …

···

Andrew Walrond