float32 to 4 bytes?

[Tai Trinh]

Hi.

How would i go about turning a float32 into 4 bytes?

And then turn 4 bytes back into a float32?

-Tai

[John Asmuth]

http://weekly.golang.org/pkg/encoding/binary/#Write

[Tai Trinh]

Ah, silly me so how obvious. Very neat :)

Thanks!

-Tai

On 23 Feb, 20:16, John Asmuth <jasm...@gmail.com> wrote:
> http://weekly.golang.org/pkg/encoding/binary/#Write

[Jan Mercl]

On Thursday, February 23, 2012 8:12:12 PM UTC+1, Tai Trinh wrote:

How would i go about turning a float32 into 4 bytes?

 

http://weekly.golang.org/pkg/math/#Float32bits 

And then turn 4 bytes back into a float32?

http://weekly.golang.org/pkg/math/#Float32frombits 

 

[Paul Borman]

That does not convert between a 32 bit native float and 4 bytes.  It converts to the specified byte order.  You can use unsafe to convert from 4 bytes in native byte order to a float 32.

    -Paul

[Aram Hăvărneanu]

Vaxen are little endian.

--
Aram Hăvărneanu

[Jan Mercl]

On Friday, February 24, 2012 8:42:55 PM UTC+1, Paul Borman wrote:

That does not convert between a 32 bit native float and 4 bytes.  It converts to the specified byte order.  You can use unsafe to convert from 4 bytes in native byte order to a float 32.

On Thu, Feb 23, 2012 at 12:52 PM, Jan Mercl <jan.m...@nic.cz> wrote:

On Thursday, February 23, 2012 8:12:12 PM UTC+1, Tai Trinh wrote:

How would i go about turning a float32 into 4 bytes?

 

http://weekly.golang.org/pkg/math/#Float32bits 

And then turn 4 bytes back into a float32?

http://weekly.golang.org/pkg/math/#Float32frombits 

I'm lost, that's IMO exactly what those function do: http://weekly.golang.org/src/pkg/math/unsafe.go

     9	// Float32bits returns the IEEE 754 binary representation of f.
    10	func Float32bits(f float32) uint32 { return *(*uint32)(unsafe.Pointer(&f)) }
    11	
    12	// Float32frombits returns the floating point number corresponding
    13	// to the IEEE 754 binary representation b.
    14	func Float32frombits(b uint32) float32 { return *(*float32)(unsafe.Pointer(&b)) }

 

[Hotei]

I think if you get the bytes from a float32 and convert back on the
same system it should be ok. Crossing between different endian
systems would of course be subject to the requirement to swap the
order appropriately.

[Paul Borman]

Oh, sorry.  I had not actually seen those.  I was generalizing from the endian package which.  My mistake.

    -Paul