Newbie question - how to write out specific bytes to an IO stream?

Hello – I’m trying to write out specific byte sequences over the
wire / to a file / etc. Let’s say for example that I wanted to write
out

99 111 108 108 105 110

as a simple, six byte sequence to a file. In Java, I might use the
byte primitive, but it seems that in Ruby you have Fixnum and then a
Float for fractions. If I simply create a Fixnum using the 99
literal, I think that that will create more than an 8-bit sequence
when I write out to the file.

Any ideas on how to do this?

On 11/6/07, Collin VanDyck [email protected] wrote:

when I write out to the file.

Any ideas on how to do this?

String objects are arrays of bytes, so you could simply do…

num_string = “99 111 108 108 105 110”
byte_string = num_string.split.map! {|word| word.to_i.chr}.join

Now, to illustrate…

byte_string.class

gives String

byte[0]

gives 99, the base 10 representation of 1100011

note that each byte can store numbers up to 255

puts byte_string

gives collin

If you want to store the 1’s and 0’s as a text string, you can use the
Integer#to_s(2) method.

hth,
Todd

On Nov 6, 6:42 pm, Collin VanDyck [email protected] wrote:

when I write out to the file.

Any ideas on how to do this?

I think you’ll want to use Array#pack. E.g., to write them out in
big-endian order:

a = [99, 111, 108, 108, 105, 110].pack ‘n*’
=> “\000c\000o\000l\000l\000i\000n”
file.write a

If you want little-endian, use ‘s*’ (or ‘S*’ for unsigned) instead of
‘n*’. See RDoc Documentation and check Array#pack for more
details on the directives.

On 11/6/07, yermej [email protected] wrote:

literal, I think that that will create more than an 8-bit sequence

If you want little-endian, use ‘s*’ (or ‘S*’ for unsigned) instead of
‘n*’. See RDoc Documentation and check Array#pack for more
details on the directives.

Hmm, I was thinking about #pack, but your example creates 12 bytes.

Thank you for your replies!

Both methods seem to work equally well, but I’m seeming to favor the
pack method because i can just push in my random bytes (0-255) into an
array. I decided to use pack “C*” though, because I think that the C*
directive will limit the packed members to 8 bits, if I’m not mistaken.

On 11/6/07, Collin VanDyck [email protected] wrote:

Thank you for your replies!

Both methods seem to work equally well, but I’m seeming to favor the
pack method because i can just push in my random bytes (0-255) into an
array. I decided to use pack “C*” though, because I think that the C*
directive will limit the packed members to 8 bits, if I’m not mistaken.

Good choice!

Todd

Todd B. wrote:

Float for fractions. If I simply create a Fixnum using the 99
If you want little-endian, use ‘s*’ (or ‘S*’ for unsigned) instead of
‘n*’. See RDoc Documentation and check Array#pack for more
details on the directives.

Hmm, I was thinking about #pack, but your example creates 12 bytes.

That’s because n isn’t the code for a byte - you want c or C, for signed
or unsigned char:

irb(main):004:0> a = [1, 2, 3, 4, 5, 6].pack ‘C*’
=> “\001\002\003\004\005\006”