On Tue, Apr 7, 2015 at 7:18 PM, Steven D'Aprano <steve+comp.lang.pyt...@pearwood.info> wrote: > On Wed, 8 Apr 2015 10:38 am, Steven D'Aprano wrote: > >> On Wed, 8 Apr 2015 03:44 am, Ian Kelly wrote: >> >>>>>> >> > to_base(2932903594368438384328325832983294832483258958495845849584958458435439543858588435856958650865490, >>>>>> 429496729) >>> [27626525, 286159541, 134919277, 305018215, 329341598, 48181777, >>> 79384857, 112868646, 221068759, 70871527, 416507001, 31] >> >> >> They're not exactly *digits* though, are they? > > Oh, I forgot... I think this is why Python long ints effectively uses a base > 256 internal storage. If memory serves me correctly, internally a long int > is stored as an array of bytes using digits: > > \x0 \x1 \x2 ... \xFF > > (in decimal, 0 to 255). Each digit takes a single byte, so it's nicely > compact, and Python includes a bunch of fast algorithms for doing > arithmetic on these.
According to the comments in longintrepr.h, Python uses either 15- or 30-bit digits, determined at configure time. -- https://mail.python.org/mailman/listinfo/python-list
