Roku Developer Program

[greubel]

Has anyone come up with a way to convert a 4 byte or 8 byte binary value in a roByteArray to a real floating point number ?
And visa versa, float or double to a 4 or 8 byte binary value.

[RokuMarkn]

The 4 or 8 byte value in the ByteArray -- is this formatted as an IEEE floating point number? Or is it an integer?

--Mark

[greubel]

IEEE float as eight bytes, 3F F0 00 00 00 00 00 00 = 1.0

[EnTerr]

In a perfect world, roByteArray should get couple of pack/unpack methods:

function ifByteArray.unpack(fmt as String) as roArray
function ifByteArray.pack(fmt as String, data as roArray) as Boolean

where fmt = sequence of format characters:

fmt BrSc type  Size, octets
-   ---------  ------------
c	String     1
b	Integer    1
h	Integer    2
i	Integer    4
f	Float      4 in IEEE 754 
d	Double     4 in IEEE 754
s	String     zero-terminated utf-8
?	Boolean    1

I keep getting those urges to sketch a BSc function unpacking 4-byte float but not having a way to inject NaN, minus 0 and +/-Infinity values (luckily) keeps me from doing it.

[greubel]

I was hoping you would have a method !

[EnTerr]

Oh, "the things I'm gonna do for my country"*:

inf = 1e39
NaN = 0 * inf
negZero = -1 / inf

BrightScript Debugger> ? inf, NaN, negZero
 inf    nan            -0 
BrightScript Debugger> ? (negZero = 0), 0 * NaN, inf * NaN
true    nan             nan 

(*) And curiousity, and vanity.