Looking for the greatest negative float value

[Bengt Richter]

On 11 Jun 2003 20:36:38 GMT, bokr at oz.net (Bengt Richter) wrote:

>On Wed, 11 Jun 2003 18:52:17 +0200, "Gilles Lenfant" <glenfant at NOSPAM.bigfoot.com> wrote:
>
>>Hi,
>>
>>I need the same as...
>>
>>import sys
>>negmaxint = - sys.maxint -1
>>
>>... but for float data
>>
>>any hint ?
>>
>
>To get what you want (wyw) maybe you have to compute it. E.g.,
>
> >>> fmin = -1.0
> >>> while 1:
> ...    fnext = fmin*2.0-1.0
> ...    if fnext==fmin: break
> ...    wyw = fmin
> ...    fmin = fnext
> ...
> >>> `wyw`
> '-8.9884656743115795e+307'
>
Actually, that didn't do it right.

Better (depending on timbot's long-float conversion ;-):
(There's may be more bits temporarily internally, but the double
you get back loses them if so. The last n in float(n) has all ones
below the first zero after 53 bits, I believe, but that's not the
number you care about, I think.)

====< maxneg.py >================
def maxneg():
    n=-1L
    try:
        while 1:
            x = float(n*2L)
            n = n*2L
    except:
        pass
    bit = -n/2L
    while bit:
        try:
            float(n-bit)
            n -= bit
        except: pass
        bit = bit/2
    return float(n)

fn = maxneg()
print repr(fn)
print hex(long(fn))
print long(fn)
print hex(((-1L<<53)+1)<<(1024-53))
=================================

[15:45] C:\pywk\clp>maxneg.py
-1.7976931348623157e+308
-0xFFFFFFFFFFFFF80000000000000000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
0000000000000000000000000000000000000000000000000000000000000000000L
-17976931348623157081452742373170435679807056752584499659891747680315726078002853876058955863276
687817154045895351438246423432132688946418276846754670353751698604991057655128207624549009038932
894407586850845513394230458323690322294816580855933212334827479782620414472316873817718091929988
1250404026184124858368
-0xFFFFFFFFFFFFF80000000000000000000000000000000000000000000000000000000000000000000000000000000
000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
0000000000000000000000000000000000000000000000000000000000000000000L


>I'm not sure how portable this is, but the actual value
>can surely vary according to platform. I suppose you could set up a site-specific
>initialization to put such a value as sys.minfloat if you wanted to.
>
>The timbot will have the best info ;-)
>
Still applies ;-)

Regards,
Bengt Richter