Mailing List Archive: 49091 messages

# How to obtain a 64-bit IEEE754 representation of a decimal number

### [1/4] from: lmecir::mbox::vol::cz at: 2-Feb-2004 13:59

convert: func [ x /local y ] [ y: make struct! compose/deep [x [(type?/word x)]] reduce [x] third y ] Usage: convert 0.2 ; == #{9A9999999999C93F} this is a big endian form of the number (in Windows). If you prefer a little endian form, which is more (human) readable, use: head reverse convert 0.2 ; == #{3FC999999999999A} -Ladislav

### [2/4] from: atruter:labyrinth:au at: 3-Feb-2004 0:32

Hi Ladislav,
> convert: func [ > x
<<quoted lines omitted: 3>>
> third y > ]
OK, this one has me stumpted. For us mere mortals, could you explain:
>> help y
Y is a struct of value: make struct! [x [decimal!]] [0.2]
>> first y
== [x [decimal!]]
>> second y
== [0.2]
>> third y
== #{9A9999999999C93F} how and why this works, in particular where "third y" comes from? Also, how does one reverse the conversion (i.e. get 0.2 from #{9A9999999999C93F} )? I assume this is another useful spinoff from the rounding project? ;) Regards, Ashley<

### [3/4] from: g:santilli:tiscalinet:it at: 2-Feb-2004 15:05

Hi Ashley, On Monday, February 2, 2004, 2:32:09 PM, you wrote: AT> how and why this works, in particular where "third y" comes from? I think it is documented in the struct! docs... AT> Also, AT> how does one reverse the conversion (i.e. get 0.2 from #{9A9999999999C93F} AT> )?
>> s: make struct! [val [decimal!]] [0] >> change third s #{9A9999999999C93F}
== #{}
>> s/val
== 0.2 Regards, Gabriele. -- Gabriele Santilli <[g--santilli--tiscalinet--it]> -- REBOL Programmer Amiga Group Italia sez. L'Aquila --- SOON: http://www.rebol.it/

### [4/4] from: lmecir:mbox:vol:cz at: 2-Feb-2004 15:33

Hi Ashley,
>>convert: func [ >> x
<<quoted lines omitted: 17>>
>how does one reverse the conversion (i.e. get 0.2 from #{9A9999999999C93F} >)?
This is a library interface, which is documented in: http://www.rebol.com/docs/library.html , where I found the informations needed to write the above function as well as: http://www.compkarori.com/vanilla/display/peek_and_poke.r reverse-conversion: func [ x [binary!] type [word!] ] [ y: make struct! compose/deep [x [(type)]] none change third y x y/x ] Usage: reverse-conversion #{9A9999999999C93F} 'decimal! ; == 0.2
>I assume this is another useful spinoff from the rounding project? ;) >
Not exactly, it is mostly independent.

Notes
• Quoted lines have been omitted from some messages.
View the message alone to see the lines that have been omitted