[prev in list] [next in list] [prev in thread] [next in thread] 

List:       koffice-devel
Subject:    Re: Problem with big integers
From:       Stefan Nikolaus <stefan.nikolaus () kdemail ! net>
Date:       2007-01-30 16:46:05
Message-ID: 200701301746.13378.stefan.nikolaus () kdemail ! net
[Download RAW message or body]

[Attachment #2 (multipart/signed)]


On Tuesday 30 January 2007 17:31:02 me wrote:
> On Monday 29 January 2007 11:45:01 Brad Hards wrote:
> > What I don't understand is why. Why not always just use a float, and
> > convert it in the asInteger() or asUnsignedLongLong() or whatever
> > functions?
>
> I was told, that long long is a C data type. Either way, I'd prefer to rely
> on Qt's platform independent qint64.
> But to your question: I see no good reason to store also the integer values
> in the double variable, whose mantisse is guaranteed to have 53 bits. The
> resulting range is -2^53 ... 2^53. Above (or below) it gets inaccurate. But
> that's at least more as long provides (at least): -2^31 .. 2^31-1.
>
> If nobody else objects and disabuses me, go ahead and modify Value to use
> the double variable also for the integers.

Thinking about it: Replacing the stored long variable by a qint64 one would 
not need more space as before. The size of double is also 64 bit. I think, it 
would: a) be easier/faster to change and b) prefered as we'll get a wider 
range for integers: -2^63 ... 2^63-1.

Opinions?

Regards,
Stefan

[Attachment #5 (application/pgp-signature)]

_______________________________________________
koffice-devel mailing list
koffice-devel@kde.org
https://mail.kde.org/mailman/listinfo/koffice-devel


[prev in list] [next in list] [prev in thread] [next in thread]