Re: [Jython-dev] int/long anomaly in jython 2.7.0

[Jeff A. <ja...@fa...>]

Now http://bugs.jython.org/issue2672.

Sorry your registration failed. (Spam filter somewhere in the chain? I'm 
afraid I've no way to investigate.)

Best wishes,

Jeff Allen

On 26/04/2018 14:05, Niemann, Hartmut wrote:
>
> Hello Jeff,
>
> thank you for confirming this behaviour and suggesting a workaround. 
> Using long() works perfectly in my case.
>
> Unfortunately i can not register at bugs.jython.org at the moment. At 
> least it hasn’t sent me the confirmation e-mail yet.
>
> Mit freundlichen Grüßen
> Dr. Hartmut Niemann
>
>
> *Von:*Jeff Allen [mailto:ja...@fa...]
> *Gesendet:* Donnerstag, 26. April 2018 08:58
> *An:* jyt...@li...; Niemann, Hartmut (MO RS LM EN 
> CCI SRD)
> *Betreff:* Re: [Jython-dev] int/long anomaly in jython 2.7.0
>
> Not AFAIK, thanks for finding it. It's still in the development tip:
>
> >>> x = int(-(1<<31))
> >>> y = x + 1
> >>> x, y
> (-2147483648, -2147483647)
> >>> x.__format__('d') # Oops
> '--2147483648'
> >>> y.__format__('d')
> '-2147483647'
>
> Almost certainly there is some thinking error in int.__format__(), 
> which % relies on, handling the positive of -2^31, which has no 
> representation as an int. A work-around is to make it a long:
>
> >>> '%d' % long(x)
> '-2147483648'
>
> The right place for this is http://bugs.jython.org/ 
> <http://bugs.jython.org/> . You could open an issue there (so you'll 
> get e-mail when we work on it), or I will do so eventually.
>
> Jeff Allen
>